Thursday, June 30, 2011

Mountain View streets: Pedestrian and Cycling Friendly?

The New York Times recently did a wonderful story about how European cities are not only directing resources towards improving access for cyclists and pedestrians, but are actively trying to retard the flow of motor vehicle traffic in order to promote cycling, walking, and public transit (which is often rail or priority-access bus lines). The text of the article was quite positive: streets for people, not for cars (which contain people, but are typically around 95% car, only 5% human), although the title was arguably directed at a motor-centric audience... a later opinion piece referred to this as "the windshield view".

In any case, what Europe realizes is a priority on infrastructure for individual-passenger motor vehicle and infrastructure for cyclists and pedestrians is to a large degree mutually exclusive. I have experienced this first hand since I began working in Mountain View, CA, from my previous job in Palo Alto just to the north.

There's a striking transition as one rides to the north along Central Expressway with its wide, bike-lane-like shoulders (not actually bike lanes but they certainly look and mostly act like bike lanes). As you head north you reach the San Antonio Road overpass next to the Caltrain station, when upon emerging from the other side: poof! The shoulder disappears and you're left to fight for roughly-paved right lane on what is now Alma with auto traffic which views the posted speed limit as a 3-sigma lower bound. This hardly seems to justify Palo Alto's self-proclaimed reputation as a hotspot of cycling friendliness. There's a recommended meandering detour for cyclists through a residential area, but it's obviously much slower than a direct road.

So point Mountain View, right?

Mountain View does have a lot of bike lanes, counting Central's shoulders. But it doesn't overcome the fundamental flaw in Mountain View's street grid. The roads are simply too wide. Four lanes, two in each direction, is narrow by Mountain View standards: there's superposed a network of "expressways" with three or in places even four lanes in each direction. Nobody wants to walk or bike on such a street, so driving is simply the way people get from point A to point B for destinations here, even if it's simply to cross the road or go a block away.

Middlefield in Mountain View

At intersections of such wide roads left turns are impossible without priority turn phases. So instead of two phases (a 50% chance to get a green light, on average) you now need six phases. When going straight, you now typically now have a 1/3 chance of making the light: you now wait at 1/6 of the intersections which otherwise you'd have been able to cruise through.

So you're not only waiting more, but each wait is longer. Intersections need to be timed sufficiently long to allow limited-mobility pedestrians to get across the street with a full green cycle. The wider the road, the longer the signal. The amount of time spent waiting at lights is directly proportional to the length of the light phases. Just because cars are expected to drive faster doesn't mean the old and infirm can.

The result: bike lanes or not, driving or cycling, you spend a whole lot more time waiting at red lights in Mountain View than in Palo Alto with its "human-sized" roads. Even on a bike it seems like you're stationary more than actually moving. Indeed riding or even driving through Mountain View on Middlefield or California, two roads which are consided bicycle-friendly, is slow going. Central northbound isn't much better. Cycling southbound on Central has the advantage of having the adjacent railway block cars from a good fraction of the intersections in that direction (you can go around the intersection rather than through it); ironically it is not listed as a recommended route on bike maps since it's an "expressway".

So what drivers do, of course, it take the freeway instead. If you can reach any of the three major highways which pass through the city, get on those. This increases travel distances, vehicle speed, pollution, and the temptation towards bigger, faster accelerating, and less fuel-efficient vehicles.

With all of this driving, people need to park wherever they go, even if it's just for coffee down the street. So if businesses want customers, and if landlords want commercial tenants, they need to surround themselves with parking lots. This makes walking even less attractive, as navigating parking areas with cars coming from every direction is discomforting or even dangerous. And parking lots add to the walking distance, and Americans typically have a very limited tolerance for walking.

So with the exception of the Castro Street oasis, a pedestrian and cycling friendly Mountain View is a lost cause. You simply cannot support high-speed high-volume car infrastructure and simultaneously provide an environment where walking and cycling is attractive. The only thing to do is follow Europe's example. Make driving less attractive and you automatically make walking and cycling more attractive. And more walkers, more cyclists, feeds the loop which encourages still more to these natural transportation modes. And walking enables public transportation, since public transit is rarely door-to-door. So there is then even less reason to drive, even more reason to practice human forms of transportation.

Matilda in Sunnyvale

The sad thing is Mountain View could be worse: it could be much worse. It could be like Sunnyvale, directly to the south. Yesterday I went for a run through along the beautiful Baylands trail. But when I emerged from the southern boundary of Moffit Field I was in Sunnyvale, near Matilda Ave. Matilda, a "local" street, makes the European autostrada look silly: 3-4 lanes in each direction, extending out to even 6 lanes with the addition of dedicated left-turn lanes. I was a hunted animal running along Matilda. There wasn't another pedestrian or even cyclist in sight. But I eventually made it, first to Maude (a road of only Mountain View proportions) and then, after criss-crossing the street due to disappearing sidewalks, back to work. Next time I run the Baylands trail, I'm turning back to return the way I came.

What a tragic, misguided waste of land and resources it all is.

Sunday, June 26, 2011

SB910 3-foot passing bill now fixed

SB910, the proposed 3-foot space requirement for motor vehicles passing cyclists, has now been amended. Now it's looking to me really good.

The earliest version of the bill was full of flaws, for example a requirement cars pass cyclists to he left (how does that work on Page in San Francisco, where the bike lane is on the left side of a one-way road?). Until the latest revision, there was an exemption allowed if the vehicle was no faster than 15 mph faster than the bike. The implication was speed differential matters. Speed is what matters, not differential.

A major problem with the speed differential is it would have taken the teeth from the bill. All you'd need to say to defend yourself against evidence you'd passed to closely to a cyclist whom you'd hit is that your speed was only 15 mph greater. It would have been very hard to prove this was not the case. Thus we'd be back to the status quo that it would need to be established the pass was unsafe. That's also difficult to prove because a driver can simply say the cyclist swerved. But with the 3-foot passing law, if you can establish the car trajectory failed to allow a 3-foot passing margin, the driver is de facto at fault whether the cyclist supposedly swerved or not. So getting rid of the 15 mph differential language was critical for this bill to be worth the bits it was stored in.

I was quote worried about this so wrote Dave Snyder of the California Bike Coalition to make sure it was going to be fixed. He assured me it was going to be, and his words proved true.

Even CABO is now on board. They were taking a hard line against SB910, refusing to buy the argument that the details were less important than the intent. And I'm glad they did.

There are two critical provisions to the bill. One is the 3-foot passing zone, the other an allowance for cars to cross the double yellow when passing cyclists. This latter part, receiving less attention, is arguably more important. It's also something close to my heart, since I'd proposed it to Joe Simitian as part of his "there outta be a law" contest several years ago. I ended up losing to a requirement headlights be on whenever windshield wipers are running. Sigh.

But here's that part:
(e) 21460. (e) (1) The driver of a motor vehicle in a substandard width lane on a two-lane highway may drive to the left of either of the markings specified in subdivision (a) or (b) to pass a person operating a bicycle proceeding in the same direction, if in compliance with Section 21751.
(2) For purposes of this subdivision, a "substandard width lane" means a lane that is too narrow for a bicycle and a vehicle to travel safely side by side within the lane.

The reference to Section 21751 is key: that says it is only legal to cross a centerline to pass when there is adequate visibility to do so safely without impeding oncoming traffic. So what SB910 does is acknowledge there are situations where fully crossing a double yellow to pass a fast-moving motor vehicle would be unsafe (which is why the line is double yellow) but it would be safe to partially cross the double yellow to pass a slower-moving cyclist, and that driver judgement is suited to determining when these situations are. And we already rely on driver judgement in passing in the case of dashed yellows. Obviously the presence of a dashed line is not sufficient reason to claim it is safe to pass.

Next is the 3-foot requirement itself:
21750.1. (a) The driver of a motor vehicle overtaking and passing a bicycle that is proceeding in the same direction on a highway shall pass in compliance with the provisions of this article applicable to overtaking and passing a vehicle, and shall do so at a safe distance that does not interfere with the safe operation of the overtaken bicycle, having due regard for the size and speed of the motor vehicle and the bicycle, traffic conditions, weather, and the surface and width of the highway. (b) A driver of a motor vehicle shall not overtake or pass a bicycle proceeding in the same direction on a highway at a distance of less than three feet between any part of the motor vehicle and any part of the bicycle or its operator, except that the driver may pass the overtaken bicycle with due care at a distance of less than three feet at a speed not greater than 15 miles per hour, if in compliance with subdivision (a).

There's three key aspects to this. First, it is absolutely explicit that the 3-foot requirement does not imply three feet is always sufficient: the bill even enumerates which factors can contribute to a greater margin being needed. Obviously if it's pouring rain and the road is covered with gravel and the passing vehicle is relatively large the passing margin will need to be larger than the three-foot minimum.

Second, it defines what is meant by a 3-foot margin. It's any part of the vehicle and any part of the bike or its rider. That means, for example, the tip of a passenger-side rear-view mirror to the rider's shoulder blade, not simply the edge of the car's fender to the frame of the bicycle.

Third, and perhaps most importantly, the 15 mph speed differential has become a 15 mph absolute speed. Cars pass me within 3 feet all the time... at traffic lights. They're stopped at the light, I pull up next to them, the light changes to green, and they pull away, repassing me. This situation shouldn't be illegal and isn't unsafe (unless the vehicle turns, of course). Yet 15 mph is slow enough that it's relatively implausible on the open road a driver will be able to claim he or she was driving slower than this, so it doesn't dilute enforceability much.

Okay -- here's what I really don't like about the bill:
(c) (1) A violation of subdivision (a) is an infraction punishable by a fine of thirty five thirty-five dollars ($35). (2) If a collision occurs between a motor vehicle and a bicycle causing bodily injury to the bicyclist, and the driver of the motor vehicle is found to be in violation of subdivision (a), a two hundred twenty dollar ($220) fine shall be imposed on that driver.

To quote a friend of mine: "somehow i would feel better if running me over would automatically mean your fine was more than throwing out a snicker wrapper,"... the fine for littering is "not less than
two hundred fifty dollars ($250)"
. Hopefully the fear of being exposed to damages, since violation of the law is an important factor in determining fault, will provide the deterrence.

So while I'd like the fine to be less insulting, I'm really pleased with the rest of the bill, and I hope it passes. I sent my state senator, Mark Leno, a note encouraging his support and I got an email back summarizing the previous version of the bill. I hope he comes up to speed soon and helps see this thing finally get passed.

Thursday, June 23, 2011

Strava Android App: head-to-head testing against Garmin Edge 500

This morning a head-to-head competition....

I had 8 minutes to catch Caltrain 206. Plenty of time, but no time to waste. I got my bike in the garage, turned on the Garmin, left the garage, shut the automatic door, then pulled my Android phone from my pocket, brought up the Strava app, and hit new ride.

Ten precious seconds later GPS signal was acquired and I quickly hit go. I clipped in and was off to the station.

Twice per block I checked the status of the Garmin. As I've noted, this is a GPS-challenged environment, and the Garmin doesn't like it. For the first three blocks the progress bar moved steadily to the right. Then it retreated a bit. Then it regained some of its previous progress... where it stalled.

I was doing well. Traffic was extremely light at 6:05 am, so my quest for scientific fairness wasn't placing too great a burden on my survival chances.

The train station is just south of 4th Street. When I reached 7th Street I missed the light. I stopped at the red. A few seconds later the Garmin switched to its main screen. Signal acquired! I hit start. There's not really a visual cue when start is hit, except in this case soon afterwards it said "timer paused". That meant, clearly, the timer had been successfully started. Since I wasn't moving it then went straight to auto-pause.

I continued the final three blocks to the station. I saw the Garmin clock switch from 6:06 to 6:07. As soon as I was off the bike I turned off the Garmin timer. I had plenty of time to put on my Speedplay cleat covers and walk to the platform inside the station before the 6:11 departure.

As I did so, I removed the phone and hit "stop", then "save". The ride was in the bank. Obviously the Garmin ride would need to wait until I arrived at work, since I don't have teathering required to connect my laptop, and can't connect the Garmin to the phone directly.

Results so far:
  1. speed:Android 10 seconds to acquire GPS, Garmin approximately 20 seconds more than one km. Android by a landslide
  2. Ease of use: The Android ride was safely in the bank by the time I stepped on the train. The Garmin was waiting until I get to work. As I type this I'm on the train passing Burlingame.
  3. Accuracy: From the phone display, the Android phone looks good. The phone display is simplified, however, using relatively long line segments which are clearly a numerical fit to the more detailed ride data. I need to wait until I can upload the ride to see how well it actually did. The Garmin data, or what data there is from the Garmin, needs to wait.

The phone does have the advantage of size over the Garmin. Bigger footprint = bigger antenna. GPS L5 band is at 1.17 GHz. Recall light travels at 30 cm per nanosecond, so that's a wavelength of 27 cm. I was never happy with antenna theory but from what I retained of it I'm pretty sure you'd like the antenna to be at least a quarter-wavelength in size, so if the antenna is less than 7 cm you're throwing away signal strength. My phone is 12 cm long by 6 cm wide while the Garmin is just 7 cm long by 4.5 cm wide. The phone is really too long and heavy for mounting on handlebars, even more so for a wrist (forearm straps are better but still cumbersome), so I think Garmin made a sound design decision here. The phone, on the other hand, has the advantage of a superior mounting point. But it definitely needs it: past experience shows the Garmin does not work when held in my hand.

Leaving Redwood City now... Menlo Park... Palo Alto.... Cal Ave.... Mt View. Time elapses.

Okay -- I looked at the data, and I'm impressed. First, most of the ride from the Android App:

Android full ride

Even though GPS is very challenging with thie buildings and hills of San Francisco, the Android app follows the roads without any obvious blips.

Here's a close-up of the end of the ride:

Android end of ride

You can see my track goes down the right side of Townsend, which is where I was riding, to the left of the parked cars, which is accurate... until I went onto the sidewalk and into the station.

Now here's what the Garmin reported. Recall due to the slow acquisition I got less than three blocks of data, so this is basically all I had with the Edge 500:

Garmin end of ride

Still not bad, although it has me passing now to the right of the parked cars, which was not my path. Is this statistically significant? Maybe, maybe not, but there's absolutely no evidence here the Android, despite being in my pocket rather than in the preferred mounting point of my bike stem (where the Garmin was), was in any way inferior.

When I got off the train in Mountain View I attempted a second experiment: Garmin Edge in one pocket, Android phone in the other. This was to try to reduce the confounding factor of placement. However, this experiment crashed and burned because when I moved the Edge 500 to my pocket (it was again slower on the acquisition, by the way) I must have accidently depressed the start-stop button, and the data were never recorded. With the Android phone this can't really happen, because I shut off the phone display before putting it into my pocket, and to turn off the data stream would require turning the display on, sliding my finger down the screen to unlock it, then hit the start-stop widget on the screen. This is all simple and quick if I want to do it, but is too complex for a mistake. On the other hand, it's easy to depress an external button.

At lunch I went for a run with the Android phone in my belt pouch. The run lasted around 80 minutes, so I was afraid the battery would be an issue, but it wasn't. And the results were great: here's the Strava record:

It was an out and back, much of the run on the Steven's Creek Trail. There's one minor glitch, but every other significant deviation of the outbound and inbound trajectory was real. Steven's Creek Trail has three dirt and gravel frontage trails along the segment I ran, and I can clearly see where I went onto and off these frontage trails. On the trail I tend to run from one side to the other depending on which appears to have better traction, so some of the variation in position was real. The only real downside was the inability to view the display when it was in my belt pouch and the lack of a "lap" button which I sometimes find useful. But it's clear that when running from work the Android phone is an excellent option, especially since I have such difficulty acquiring GPS signal with my Garmin near work.

So anyway, from these experiments all the evidence is, if anything, the Android phone is doing a better job than the Garmin Edge 500. The quick data acquisition is a great plus. The reduced battery life is an issue on longer events, the bigger weight is an issue in races, and the lack of a lap is an impediment for interval training, but otherwise it's working substantally better than I anticipated.

Wednesday, June 22, 2011

first impressions of Strava Android App

Recently I've been testing the new Android Strava app. My preferred method of recording ride or run GPS data has been the Garmin Edge 500. On the bike, I use the usual mount which I attach to my stem. For running I got a Forerunner wrist strap which works great: the Edge 500 and the Forerunner share the same strap.

But when the Strava app for Android app was released I was pretty excited. I'd read and heard hints and rumors of real-time features, perhaps for example instant notification of KOM rankings results while on the bike.

But, alas, nothing so spectacular. The app reports speed, distance, and time of exercise, then allows the user to upload the ride to Strava. Once uploaded, a limited resolution map of the activity can be viewed along with rankings on matched segments (once the servers have had adequate time to process the data, less than a minute).

So it seemed as if I'd have little use for it. But I decided to give it a try anyway.

And I was pleasantly surprised. First, my Android phone is substantially quicker to acquire GPS signal than the Edge 500 unit. For example, if I want to record a ride from the 22nd Street Caltrain station to home, I need to either turn the Garmin on well before the stop so it has a signal uninterrupted by the series of tunnels leading to the station, or I need to move out to to an area clear of obstructions near the station and wait there. The issue is the only such area I've found is the middle of the intersection between 22nd Street and Pennsylvania Ave. Staying here for sufficient time for the Garmin to pick up signal presents certain hazards. So if I don't turn it on by the time we pass the Bayshore station it's a challenge to get signal acquisition in time.

Contrast the Android phone. I can turn that on when I'm leaving the station and as soon as I leave the direct obstruction of the overhead freeway, I'm good to go. Very quick. And here's the result in this challenging GPS environment... even though the phone was in my front pants pocket, it even had the correct side of the street:

Riding home from 22nd St Caltrain

For a typical longer ride, if I'm meeting people I'm typically rushing to get out at the optimal time. If I've thought far enough ahead to turn on my Garmin before going downstairs to get my bike, by the time I get there the unit's typically auto-powered-down due to inactivity. More typically I turn it on as I get on my bike. I then ride down Potrero Hill, glancing at the display periodically to watch the progress of the satellite aquistion bar. This reminds me of Coach Sok in high school who used to count ten pushups finishing with "7", "8", "9", "9", "9", .... With short blocks and frequent intersections I of course need to watch where I'm going, but since I don't want to lose more ride than necessary I'm tempted to check the unit more often than necessary. Typically around a km from home it will declare "motion detected" indicated the signal is acquired, and I am asked to hit start. So I quickly hit the start button (there's four, so you need to remember which one this is) and try to make my appointment. Eventually I'll glance at it only to see I hadn't pushed the button with adequate firmness and it's still waiting for me to hit start. So rides recorded with the Garmin 500 often "begin" several km from home.

I simply don't see this mistake happening with the Android app. The satellite acquisition is only a few seconds: no problem waiting for that before I ride off.

Then there's upload. With the Garmin I need to get the micro-USB cable, plug the Edge 500 into a USB port on my laptop, log into Strava, select "upload activity", select "use local file", open the Garmin device once it auto-mounts, select the "Garmin" directory, the "Activities" subdirectory, wait for the directory to load, pick the last file in the long list, confirm upload. Not too bad, actually.

But with the Android app I click on "save ride" and it uploads. Done.

But it's not as one-sided as it seems. The issue is when I upload an activity, I need to enter certain data fields. For example, I need to select the activity type. Is it a bike or a run? If it's a bike, what bike did I use? What was the difficulty and the roughness of the terrain? Was it a group ride, a commute, or a stationary bike ride? And what was the name of the activity? Maybe I even want to post a ride report.

These things must be done through the Strava web page. On a laptop, it's all easy. On the phone, text entry is considerably more challenging, and selection of the widgets which allow entering many of the fields is difficult with human-sized fingers. And on the Android browser, not all aspects of the Strava site even work. For example, I am unable to select individual segments within a ride (although I can see the pages for the segments themselves).

This is especially an issue for runs. If I upload them via the phone, they'll receive the default designation of a bike ride. This leads to a corruption of the on-line data (for what that's worth). Usually, obviously, a segment completed by foot will be slower than one ridden by bike, so the matched segments from a run will appear towards the bottom of the standings. But this isn't always the case: I dare say I can run the Broderick sidewalk faster than I could ride it, assuming I could ride it. The Broderick sidewalk is a 35% grade...

I am sure this will be fixed. Obviously Strava wanted to get something robust and simple into the marketplace to get people who've not made the specific investment in a Garmin GPS unit able to use their service. Since smart phones with GPS are becoming virtually ubiquitous, and since Strava's social networking business model is invested in gaining a critical mass of users not just in the San Francisco Bay area but around the world, I understand their decision. There's time for more functionality later.

First thing is they need to add data entry directly to the phone app. I should be able to select the activity type and the other activty data fields. I would also like to be able to browse KOM lists.

Beyond that there's all sorts of things they can do, including the "real time" data I thought they might already have.

But as it is now, it's my option of choice for short commute rides. In the past, I didn't bother with these: to and from the train station, for example. But with the Android Ap, it's trivially to include these. I just select the ap when I leave, wait a few seconds for GPS acquistion, then hit "start". When I get to my destination I hit "stop" and "save". I annotate the rides next time I check the web page on-line.

Then there's the issue of accuracy. GPS accuracy is important to getting good matching and timing of segments. Here the phone has a big disadvantage to the Garmin, and that is location. I tend to keep the phone in a pocket: a jersey pocket if I have a cycling jersey or a pants pocket. This obviously reduces available signal strength since my body is an electrical conductor and blocks the signal from the satellites. However, while the Garmin Edge was terrible when I'd run with it in my pocket, my phone does surprisingly well. Maybe the phone uses more current; I'm not sure why. But it works a lot better than expected.

Another worry I had was battery life. My HTC Incredible Android phone has terrible endurance. Even moderate phone use the the phone won't last more than 12 hours without a recharge. but I've had no problem doing lunch rides, for example, with the Strava app running. I've not yet tried truly long rides: for those I prefer the Garmin.

Then there's weight. The Garmin Edge 500 checks in at 56 grams: almost exactly 2 ounces. My Android phone is mind-numbing 138 grams. Remember my threshold for saving weight on my race bike is $4/gram, so that 80 gram difference would cost more than $320 to counter, and for running, mass is even more important than for cycling. So I wouldn't use the phone if I was in a scenario where I placed a premium on lightness. But since I often have my phone with me on rides anyway, this typically isn't an issue.

So if you have an Android phone and use Strava but haven't checked out the Android app yet, I definitely recommend giving it a try. You may find you prefer it in many cases to the Garmin.

Monday, June 20, 2011

Zombie Runner / Coastal Trail Runs Half Marathon San Francisco

Zombie Runner Half-MarathonI felt confident going into the Zombie Runner Half-Marathon last Saturday. Under advice from a friend I'd pre-run the course the Sunday before, and I was immediately really glad I had. Navigation was far from trivial, and having worked these potential snafu hotspots out already, I could focus my attention on running and less on orienteering.

The only question was speed. I'd done very little approximating "speed work" this year, and with my run-specific training consisting of an average one run per week, I wasn't sure how I'd handle the increased pace that race day invariably inspires.

Conditions were essentially perfect at the start line adjacent to Crissy Field as Wendell gave the starting line instructions. The course was simple, he said: just stay on Coastal Trail. But I realized from last week's preview run it wasn't that simple. One point of clarification: I asked whether we were going through the low-headroom tunnel at the top of the stairs from Fort Point. He confirmed we were. But I worried about the other runners, as Coastal Trail has several scenic out-and-back spurs along the way. These are also marked "Coastal Trail", and runners might accidentlly et a few more miles than anticipated. At least they'd get to enjoy extra view.

Wendell at the start. Coastal Trail Runs photo.

The race opened with a short southern leg, totally flat and mostly paved, then a return northward on the same route. I settled into what felt like a fast but sustainable pace: sustainable for 10 km, certainly, but maybe not the 21 km I'd need to run today. But I didn't need to sustain it for 21 km... the many staircases along the way would enforce rest: walking up the stairs shifting the load to different muscles, while descending the stairs giving my cardiovascular system a break. So I was willing to invest a little more here than I might have otherwise.

And it payed off, I felt. Soon after the turn-around on Marina Green, I was passed by another runner. I glanced at my GPS and saw 15.2 kph: faster than my best guess at my sustainable 10 km pace (around 4.1 min per km, or 14.6 km/hr). I got onto his heel (if it was a bike race I'd say "wheel", so in a running race it must be "heel") and echeloned to the left, with the wind coming off the water to the right. I estimated this should save around 1% of my power and therefore increase my sustainable speed by close to the same fraction. That's around 25 seconds per 10 km: not to be squandered.

I hit 5 km 21 min after the start. Soon after the race would leave the lowlands by the water and begin to climb. I'd fallen back from my leader at this point, but then slotted in with a group of 3 others. We'd almost closed the gap on the other runner, so there was a decent group as we reached the base of the stairs; I was slightly gapped off from the rest.

But I closed that gap when the rest were slightly confused by the turn to the stairs, and so we took the stairs together. Nobody wanted to try them two at a time, so I ran as I had in my preview run, one at a time. It's not really running, actually, more of a power walk. The pace felt good, not too hard but decent in the context of the race distance. I was surprised, therefore, when I later looked at my data and saw I'd taken this section faster the week before. On that run I was being chased up the stairs by someone; this was more tactical.

At the top we turned right and ran toward the tunnel. It's dark with a low ceiling on the exit so I warned those around me. Nobody seemed to have an issue, though, so on we went.

We scrambled under the bridge, then over some paths, then down some stairs due to a section of trail being closed. I was slower then some of the other here and I lost the sense we were a group any longer.

Some more climbing, again where I tried to pace myself, and I was on the singletrack adjacent to Lincoln. A respite from navigational issues... just run, run run.

We passed Sand Ladder. I'd thought so... too bad because the sand ladder is fun.

A twisting descent to Baker Beach. I did my best to lengthen my stride here. The guy I'd been following during the early kilometers passed me here. I'd thought he'd been up the road. I have a short stride and it impedes me on downhills where it really helps to open up.

At the bottom, a short path, then the sand. I tried to float over the sand like Legolas running on snow, but to no avail: the sand feels a bit like running in place. Where I could I used plants infringing on the path for superior traction (I felt a bit guilty about this; poor plants): every tenth of a second counts.

Not That Lucas
Near Baker Beach aid station. Not that Lucas

I was getting parched, but just as I thought this might be a concern I saw the tent of the Baker Beach aid station ahead. I decided to place hydration above speed here, so instead of trying to grab and go like a road runner, I stopped and downed in succession three partially filled cups of sports drink. It maybe cost me three seconds but better that than to risk spilling water on the following stairs and falling behind on liquids 8 km into a 21 km run on this warm morning. I'd decided at the start to not bring a bottle with me, as the time lost from the extra weight would exceed a few seconds spent grabbing cups, some fraction of that time returned due to the recovery gained. This was probably the best choice and I saw very few of the runners around me carrying water.

Up the stairs, then I was on 25th Ave. Here I was very happy to have pre-run the course because I didn't hesitate in where to run, once I'd made the initial left. Signs proclaimed "runners in the road" so I took this as justification to use the full road to apex turns. There was virtually no traffic and line of sight was excellent so there was no issue with this, and the difference between optimized versus circumferential trajectories adds up.

Camino del Mar doesn't seem to be much of a hill but when I rode it the day after the race I was seeing 5% and 6% on the grade display. I felt as though I should be moving faster... but I was doing fine relative a runner ahead of me so I didn't worry about my pace.

The sign said "Lands End" marking the opening to the trail again. A volunteer was posted here to make sure runners didn't continue up the climb to the Legion of Honor. Here's where the real fun begins: the stairs.

The stairs are an interesting feature. Going up is just a matter of maintaining a sustainable cadence and staying in sync with the stairs. I never took them two at a time and only rarely had to take two steps per stair. So basically the stairs called the shots here.

I was more worried about going down. Here I tried to dance as lightly as possible on the stairs, using them to stop my acceleration but not my progress. I did better than expected relative to other runners I saw -- I'd expected to be massively slower and I was instead only slightly slower At one point I was able to run next to stairs on a dirt shoulder, which was faster.

Signs were posted all along the route warning about a "Trail Marathon" this morning. I was thus slightly impatient with trail walkers who let their dogs run free. But although I had to slow for maybe a half-second once, there were no other issues here.

The views are spectacular along this section, so when I was able to glimpse out to the water I was instantly rewarded with a wonderful, clear view. But my priority was on running so I didn't do this as much as I had the week before.

The trail to Sutro baths arrived more quickly than I remembered (probably because I was running faster) and I scampered down the steps as well as I could. I'd anticipated the aid station would be down at the level of the baths, so was slightly surprised when a runner ahead turned up the stairs going from the bath to the parking lot above. That was obviously where the aid station was.

I heard Cara's voice from above asking me what I wanted to drink. It was really nice having Cara here. Naturally I wished she could run herself, but that's not possible for her right now. "One Coke, two water!" I shouted, and they had Coke and water available when I reached the top of these stairs. I chugged these down, then a third drink (which turned out to be sports drink), and I was off, chasing the guy ahead.

More stairs, then we'd completed the Sutro Baths loop and were retracing our steps towards the start. This was fun because outbound runners were coming toward us, providing nearly constant opportunity to wish encouragement and receiving plenty of encouraging ourselves.

I caught the runner ahead... I was a bit faster going up, he a bit faster down, but we pretty much stayed together until we'd cleared the trails and returned to the roads. He ran on the sidewalk, but I went out onto the roadway to take advantage of the more direct route through turns. I think this is where I dropped him for good, although it's possible in the chaos approaching the finish he passed me again.

I stopped for two more cups (filled higher now) at the Baker Beach aid station. Then it was time again to wade through the sand (I forgot to use the plants this time) then begin the climb back to Lincoln. It was in this section I was caught... I thought initially it must have been the guy I'd passed but it turned out to be another runner. I wondered how he'd gotten behind me because he seemed strong. I tried to hang with him but no luck -- he pulled away.

Then the Lincoln singletrack, past the Sand Ladder trail, and then the section approaching the bridge.

Here's where things are very fuzzy. 10 km runners and at least a few half-marathoners somehow came together here in a small pack. I knew from my run the week before I needed to make the left to stay away from the parking lot, but in the twists and turns I was too focused on following. All of a sudden people slowed: we were at Merchant's Ave. "We did something wrong," someone muttered. Things seemed bleak.

But I knew the way, so led the others on the sidewalk adjacent to the road, then next to the cafe, and eventually down back to the trail.

Now the question is: was this wrong? We all ran together, and from the GPS data the distances are almost the same: within 50 meters which is the best I can resolve. None of us had seen a course marking, and when I later talked to people at the finish many others had also missed it. Someone said there was a marking, but it was a single ribbon along the left hand side of the course and runners at that point were running along the right. Another runner suggested there had been trail "vandalism": I'm not sure picking up what appears to be a discarded ribbon on the side of a popular hiking trail is exactly vandalism. Whatever the reason none of us had seen the marking and taking this route seemed the reasonable thing to do at the time. On the other hand, it spared us having to duck to run under the tunnel near the picnic area, and that's an obvious slow point, but on the other hand we spent some time trying to decide what to do, so I think it was a wash.

So we found ourselves back at the top of the stairs down to Fort Point. I ran these one at a time, as best I could, although more skilled descenders took them quicker. At the bottom, I returned the way we came, around some fencing ahead. But others turned right, taking a more direct path. Maybe they did the intended thing; I'm not sure. I didn't see any markings either way.

From here it was a flat sprint back to the finish. It was further than I expected, but it didn't matter, I was just in that zone of zeeing how far I could push my speed. Speed's about efficiency, about fluidity, and about relaxing. My form's not good enough to not be the limiting factor.

It was a bit surreal, as there were so many people on the path going in both directions: 10 km runners, a lot of non-participants just out for a run, and one half-marathoner I managed to catch.

At the finish I was handled a medal, which was nice. Results 14 deep were posted amazingly quickly (I'd barely caught my breath when Wendell taped them up) which meant I was last on the list. My time: 1:41:47 seemed impossibly good considering all of the impediments from km 5 to 19. I was really pleased with that. I got another medal for being second in my age group. Between men and women and with age groups every 10 years if you finished in the top 20 there was a really good chance of being top 3 in whatever age group you were. But that was nice, as well.

The finish was fun: just hang out and enjoy the wonderful day while I waited for Cara to return from her duties.

So what next? Dare I think about a 3:20 road marathon? Maybe possible. But I will need to pay more attention to my running in coming months.

Sunday, June 19, 2011

Strava, Garmin Edge, and the top of Old La Honda

The second worst offender in the little investigation I did last post was the "INTEGRATE Performance Fitness OLH Climb". Integrate is a fitness studio in Mountain View so I suspect they did a time trial there to test client fitness. But while the West Alpine climb which resulted in the 48 second error is a bit tricker to debug, this one is fairly clear.

Old La Honda starts with a bridge crossing and finishes 5.2 km later at a stop sign immediately before a ridge road which contains high speed car and motorcycle traffic. While a full stop honestly isn't critical here, at least some attention past the sign is necessary. And it's common for riders to stop past the sign on the dirt shoulder. With manual timing, you start your climb at the bridge (I use the trailing edge) and finish when your front tire crosses the perpendicular line passing through the stop sign. It's all very precise.

But nothing about commercial GPS is "very precise". As I've pointed out, there can easily be a 10 meter error, and in the compromised signal environment among the redwood trees near the top of Old La Honda Road, you're lucky if you do this well. So when defining a segment, it's good to include some error margin. Do not start or finish the segment through which a rider's GPS track, which deviates somewhat from his "true" track, might not pass. When defining segments, I usually move one sample point in at the start and finish of the segment. For one-block segments, for example steep climbs in San Francisco, obviously you can loose a lot of valuable pain this way, so I try to shave it relatively closer. But for a 5.2 km climb like Old La Honda, 10 meters here or there isn't going to change the rankings much. Sure, you will get timing different than you would have gotten with a stop watch, but as I've already described, you can already expect around a 4 second difference anyway.

So here's the finish of that INTEGRATE (TM) segment, taken from the Edge 800 data with the identified segment marked in blue:

and here's the same section of roadway from the Edge 500 data:

As suspected, the finish is flush right against the edge of Skyline Boulevard. This is bad. Here's the speeds recorded on the two data tracks:

The Edge 800 segment ends just as the speed is falling off. But with the Edge 500, the segment isn't terminated until the rider has been stationary for a considerable duration. Had the segment been terminated just 10 meters early, this problem wouldn't have happened. Fortunately there's other segments for Old La Honda: the INTEGRATE (TM) version defaults to "hidden". The difference in the Low-Key Hillclimb version, which is designed better, is only 8 seconds. That's still a lot, but obviously a lot better than the 25 second difference observed for the INTEGRATE (TM) segment.

As to why there was a difference in whether the inert portion of the ride was included or not, that's not easy to guess if you compare the trajectories reported by the two units:

Ouch. As I've noted, the top of Old La Honda road isn't very GPS friendly.

Thursday, June 16, 2011

testing Strava segment timing reliability

DC Rainmaker recently did an interesting study of the accuracy of GPS units. He mounted eight different GPS computers on the handle of surveyer's measurement wheel then walked, ran, or rode various courses, comparing the measured results (Part 1, Part 2). The GPS units tended to disagree on how far he'd gone, although usually the results were consistent with the claimed positional accuracy of the units.

I suggested he use the data to test Strava reproducibility in segment timing. No luck there, but I did find that a friend of mine has been in the habit of riding with a Garmin Edge 500 mounted alongside a Garmin Edge 800 on his rides. We have a mutual friend who works for Garmin, so he's doing this to compare the two.

I asked him for data from three of his rides. I then created two new accounts on Strava and uploaded the data from his Edge 500 into one, and from the Edge 800 into the other. The new accounts are necessary because Strava rejects what appear to be duplicate rides from the same account. I made these rides private to avoid contaminating the historical record, since he'd already uploaded data using his personal account. Then I created a spreadsheet with all of the segments for each of the three rides.

The Edge 800 data yielded 58 matched segments, while the Edge 500 yielded 56 matched segments. The two segments matched to the Edge 800 data but missed in the Edge 500 data were "West Alpine Road Start of Climb" and "West Alpine Road Portola State Park Road to Finish". Obviously there had been an issue with the Edge 500 data on Alpine Road. However, the Edge 500 data did trigger the "West Alpine Road Alpine Creek to Peak" segment. West Alpine Road is a relatively complex climb and has had an extraordinary number of segments defined for it: of the 56 total segments matched to both data sets, nine are on Alpine Road. And these are in addition to the two Alpine Road segments which were assigned only to the Edge 800 data.

For each segment I subtracted the claimed time for the segment derived for the Edge 500 data from that derived for the Edge 800 data.

Before I show the graph, I should add I expected some difference. Garmins sample at one or two second intervals. Strava interpolates on these data points, but interpolations can only do so well, so I'd expect an error of around 1 second on the start time and around 1 second on the finish time, so even if everything's perfect, an error of ±2 seconds is about the best I'd anticipate.

There's other errors, of course. The GPS signal is only good to around 10 meter accuracy. But these are units with essentially the same electronics and the same algorithms looking at a signal within 10 cm of each other. So while the general positional error of the GPS signal should add up to around 10 meters of uncertainty to the start and stop position, since this positional error should affect both computers close to the same. But since bikes move at around 5 meters per second up hill, a 10 meter error at either the top or the bottom along the direction of travel could create another two seconds or so of variation in the segment timing.

Then there's the problem that the segment was defined with data which was also subject to noise. You'd like to believe there's an imaginary line across the road defining the start and end of a segment but the reality is the virtual line, even if your GPS is perfect, is slanted. So if your position in the road varies, or if the GPS signal varies your trajectory to the left or right, that will affect at what point you intersect these virtual start and finish lines. This could be another two seconds or so, similar to the error from longitudinal position error, off the start and finish. But again this error should be relatively smaller because we're considering two GPS units on the same handlebars at the same time.

So worst case I have the following error estimates for ride-to-ride variation:
  1. 1 second at start due to sampling time
  2. 1 second at finish due to sampling time
  3. 2 second at start due to longitudinal position errors
  4. 2 seconds at finish due to longitudinal position errors
  5. 2 seconds at start due to transverse position errors
  6. 2 seconds at finish due to transverse position errors

I assume these errors are uncorrelated so I take the root-mean-squared-sum and get around 4 seconds typical variability for ride-to-ride variations, but less than that for two GPS units mounted on the same handlebars on the same ride... let's say 2 seconds.

So what's the data show? Here's the results:

segment time difference

If I look at the mid-range of the distribution, my estimate was spot-on: errors are typically between ‒2 and +2 seconds, without evident bias between the two data sets. However, the devil here is in the tails. A significant number of segment timings have far worse errors.

These segments, it turns out, are all either on West Alpine Road or on Old La Honda Road. The top of Old La Honda Road, in particular, is notorious for terrible GPS signal quality due to the trees and terrain creating confusion from signal reflection. But Strava's algorithm is relatively forgiving, and so assigns segment times anyway.

Here are the worst offenders where the Edge 800 reported shorter times:

climb                                          delta
INTEGRATE_Performance_Fitness_OLH_Climb        -25
OLH_Mile_3_to_End                              -12
OLH_(LowKey)                                   -8
Arastradero/Alpine_-_Portola_-_Wed_Valley_Ride -7
Old_La_Honda_(bridge_front_to_stop_sign)       -5
Old_La_Honda_(Bridge_to_Mailboxes)             -4
West_Old_La_Honda_Descent                      -4

Wow -- 25 seconds on that first Old La Honda segment!

And here's the culprit segments where the Edge 500 reported shorter times:

climb                                          delta
West_Alpine_-_full_length                       4
Old_La_Honda_Mile_3                             4
OLH_-_Mile_2_to_3                               7
West_Alpine_-_First_Half                        8
West_Alpine—Alpine_Creek_to_Portola_SP_Rd       18
W_Alpine_climb_-_Alpine_Creek_to_2nd_switchback 31
West_Alpine_-_Alpine_Creek_to_peak_(RR_gate2)   48

This one's even better -- the West Alpine segment has a whopping 48 second disagreement. It's as if the Edge 500 had dropped the 800 with enough of a gap to get out of sight on those final turns... And curiously Old La Honda data actually appears at both ends of this range, demonstrating what a problem child Old La Honda can be.

So it may be on most segments the Garmin-Strava link does fairly well: within a handful of seconds. But on problematic segments the error can be profound, enough to radically change rankings.

Perhaps Strava should tighten up the criteria by which it considers rides to be a match to segments. This would result in users complaining that they'd ridden a segment but not gotten credit. But on the other hand it would improve the integity of the KOM rankings for these difficult segments. An alternative would be to flag marginally matching data on the rankings, so it becomes clearer that the results are questionable.

Wednesday, June 15, 2011

CycleOps GPS Joule and the future of head units

Garmin released its Edge 500 head unit at around the same time as CycleOps released its Joule. Each of these was an ANT+ Sport compliant head unit which could read power data from compliant power meters like the Quarq and PowerTap. However, otherwise they almost couldn't be more different.

JouleThe Joule's focus was on functionality. Power users like to analyze their data, CycleOps reasoned, so we're going to give the user as many options as we can fit to crunch power numbers and even do history analysis. This sounded good, but the head unit is the wrong tool for this job. Riders would upload their data at the end of the ride anyway, for example with Golden Cheetah or WKO, then may be from there to a web site like Training Peaks. So the result was a heavy, bulky, and expensive head unit which sold like coldcakes.

And despite all the cost and complexity the unit lacked GPS. Why burden a head unit designed for a power meter with GPS? The Wattage List old-guarde eschewed GPS, and many still do. When I'm riding, all I care about are the hard numbers: speed, power, and altitude for context. I know where I rode, and frankly it doesn't matter much, they argued.

Garmin on the other hand scored huge with the 500. It focused on recording, not on display. And like all Garmin units it had GPS, of course. Sure, it told you what your present numbers were, but any sort of historical analysis was minimal and difficult to access. Even, for example, checking the power from previous intervals during an interval session required way more button clicks than anyone should attempt on-the-bike.

But in the San Francisco Bay area, the things have sold insanely well. I'll do group rides where 80% of the people have Garmin GPS units, most of them Edge 500. So why -- why do people need a GPS unit which barely even tells you where you are? Certainly my Edge has never helped me avoid getting lost or helped me find my way back from being lost. Although it does have limited navigational "follow a route" functionality, most users don't know how to use that. I've never seen noticed doing so. Frankly, most riders know where they're going; they don't need a GPS for that.

So if they don't need a GPS to tell them where they're going, why use a GPS? Around here, it seems to me the #1 reason is one that didn't exist at the time the Joule or Edge 500 were released: Strava. Strava is the paradigm shift in what Bike Snob calls "Cat 6 racing": competing with friends outside the domain of sanctioned events. It's also been called "asynchronous racing". Sure, this has always existed. People would obsess over their Old La Honda times, the records up that hill passed on in a complex oral tradition. But with Strava you finally have hard data accessible by all. And not just for Old La Honda, but for virtually every bump in the road. If you're not the Old La Honda KOM or QOM, there's surely some undiscovered hill somewhere where you can be KOM or QOM.

Power is hard to understand. Maximal power curves, smoothing functions, fourth-power norms, training stress scores... but speed is speed and everyone understands that point A to point B in the minimal time is what it's about.

Strava has been called the "social network for cyclists". It's been the killer app which has made a GPS unit essentially mandatory for the "competitive club riders" here, and it's catching on elsewhere as well.

So CycleOps responded with a the new Joule. Sleeker and lighter than the first Joule, this one has a GPS option. So clearly it's being recognized position determination on a bike is more than a passing fad. DC RainMaker has an initial review. So what's next?

I think the next thing is connectivity. Strava seems to be leading the way here with their phone apps: first for the iPhone, and upcoming for the Android. Since the phones are connected, data upload can be virtually real-time. The Android app promises to do one better: to provide near instant feedback on performance on segments, allowing a rider to know how they ranked on a just-completed climb even before they have time to catch their breath. There's obviously all sorts of possibilities once a computer is connected. For example, you can imagine being able to track other riders in your group.

But before getting carried away, I wish to return to the emphasis on relative simplicity in head units. Record + transmit. If I want to analyze data and run fancy apps, at least for now, leave that to the phones and other heavier, more power hungry, more general purpose devices which can be brought along if one wishes. Keep the cycling specific hardware light, robust, and simple. The key is to not keep the data sitting idle, and to not require users to perform cumbersome up-loads at the end of the ride. That should happen automatically, wherever you are, as long as you're within access to a wireless data network.

It'll be interesting to see more details on the Joule, but I'm even more interested in what Strava does wth its Android app.

Monday, June 13, 2011

Cunego's descending clinic

Not too much I can add to this... a spectacular demonstration of countersteering and trusting the traction of your tires by Damiano Cunego (and to a lesser extent Peter Sagan) in today's Tour de Suisse stage:

Oh yeah, and the scenery is gorgeous, as well.

It's definitely worth bringing this up in YouTube to increase the video size.

Sunday, June 12, 2011

Golden Gate Bridge: West versus East

Despite swearing off the Golden Gate Bridge last week after a harrowing experience crossing the bridge on the eastern "pedestrian" path, the western path shut down due to "seismic retrofits", I did it again today. Tired from my run yesterday, I needed motivation to ride, and the Mission Cycling ride is a local option to get me out the door.

There were only four of us at the cafe where the ride meets on Sundays at 10 am, and since the hills providing the preferred southern routes were fogged in on this chilly morning, the default took asserted its inertial self and we found our handlebars pointing the way to Marin. One reason group rides are so popular is they relieve one of the burden of thinking.

Arguably the chilly, foggy morning deterred some would-be-bridge crossers, and while the way out was very slow, it wasn't too bad. The way back, however, is where chaos bloomed: by the afternoon the sun had come out and that sent the tourists.

At one point I was following another rider on the right of the path, each of us going walking speed, when I had to slam on the brakes. A tourist had been pressed against the roadside railing shooting a photo of others in her group. Her mission accomplished, she darted back across the path, clearly without looking.

Anyway, I survived, and renewed my vow to avoid the bridge in coming weeks.

I decided to quantify the delay the present state of affairs is causing. Fortunately Strava maintains a database of my bridge crossing times. So Here's the numbers applicable to my rides over the past year that I've used Strava, at least those rides where I've used the GPS and uploaded the data (which is most).

22 rides on western path: avg = 446 seconds, σ = 41 seconds
2 rides on eastern path: avg = 577 seconds, σ = 64 seconds

22 rides on western path: avg = 425 seconds, σ = 36 seconds
2 rides on eastern path: avg = 679 seconds, σ = 13 seconds

So the eastern path has been around 131 seconds slower northbound, 254 seconds slower on the southbound. This is because both of my southbound east-path rides have been in the more crowded afternoon, while my northbound rides have been in the less crowded morning. This is an added delay of 385 seconds = 6 minutes and 25 seconds to the round trip.

It's claimed up to 6000 cyclists per day ride the bridge. If each of these "cyclists" is one-way across the bridge, and each is delayed on average as much as I have been, then that's up to 321 rider-hours per day of delay.

Now 6:25 doesn't seem like a deal-breaker in terms of a decision of where to ride, although since I was late in getting home today I wish I'd had that back. The real cost is in the nerve-racking unpredictability of the tourists: typically walking three abreast, suddenly jumping across the path for photos, even (in one case I saw today) riding a bike while holding a videocamera in one hand, handlebar in his other, two eyes on the view, none where he was going. It's fairly crazy out there.

Saturday, June 11, 2011

previewing Zombie Runner - Coastal Trail Runs San Francisco Half Marathon Course

Last week I registered for the Coastal Trail Runs San Francisco Half-Marathon, part of the Zombie Runner series. This race happens to follow virtually all of the Escape from Alcatraz Triathlon run course, plus more. The notable difference, it appears, is the Coastal Trail Runs race avoids the famous Baker Beach sand ladder.

Here's the official route map:

So after riding my bike to Sports Basement / Presidio, I locked my bike on their rack, changed to my running shoes, transferred my Garmin 500 to my wrist strap, stored my cycling shoes and jacket in my backpack, and checked the pack in the store. Sports Basement is an excellent hub for activities. So I was ready to set out.

The map had been unclear and I'd assumed the paved loop around Marina Green came first. In fact, when I later checked this course description, I discovered it had been last. My legs were still feeling the 30 km run I'd done a week ago: that had been a bit over my head. So I started nice an easy, at least until a woman passed me and I tried to at least match her pace to take advantage of the draft into the stiff headwind.

As I finished the loop I was feeling okay, and was ready for the run to the bridge. First this goes along the smooth, wide dirt Presidio Promenade until it passes the Warming Hut, then after navigating around some fenced-off construction hits the stairs leading to the coastal trail above.

I wrote about these stairs last time, when I described watching Alcatraz. I tested my claim they were better run one-at-a-time. I felt as though I was going well, but I heard someone breathing behind me, and it sounded as if he was gaining. I held him off, although I was tempted at the top to turn and ask him if he'd run the stairs one or two at a time. This is early in the run, though, so I prefer to save my legs a bit here so will stick to my one-stair-at-a-time approach. The goal isn't to KOM them, although a good effort can be made because there's some recovery after.

After the stairs there's a bit of gradual uphill, then a right (assuming we follow the Alcatraz course) and a dark tunnel with low headroom, especially on the exit. The floor of the tunnel is also irregular, so I usually slow down here.

The tunnel emerges at a small "picnic area" from which it's a short distance to the paved bike-pedestrian path to the bridge. This path is always crowded with both walkers and cyclists. Just before the pavement turns for the bridge, however, the hiking veers to the right.

This leads to a somewhat rugged route past old military buildings and eventually to a narrow path along busy Lincoln Boulevard. This is sufficiently early in race that there will be no returning race traffic here to create conflicts.

The path first climbs, then descends a bit before arriving at the turn-off for the Baker Beach sand ladder. We're not doing this, according to my reading of the map. Instead we continue further then take a series of switchback turns down to the beach.

The trail at the beach is through loose sand, what you'd probably expect for a beach. Here on my run I stopped at a public toilet, stopping my Garmin while I did so, then forgot to turn it back on. From my return route, I estimate I lost around 1.6 km of run data from this error.

At the end of the sand there's a short stairs which ejects runners into the cul de sac of 25th Ave. This is a super-rich neighborhood, and the houses are impressive, but the loss of the Coastal Trail here is unfortunate. While I ran here, I had fantasies of the owners of the houses having their land confiscated, the houses demolished, and the natural coastline restored with the continuation of the Coastal Trail, which claims to go all the way to Mexico. Not likely to happen.

25th leads in short order to Sea Cliff, then to Camino del Mar. Here's where I got a bit confused, but I think I was correct. I ran along Camino del Mar for awhile until the I saw the turn-off for the Coastal Trail and Lands End. So I turned.

It was here I realized any hope of a fast time on this half was lost. There's a considerable number of steps on this trail, also known as the Lands End trail. But having to spend a bit of extra time here is a gift -- the views are absolutely spectacular. Heading westward, the view of the open sea beyond the rocky coast below. Returning the Golden Gate Bridge lays beyond the coast. This is really going to be special.

It's not so much going up the stairs that's the issue, but down them. Sure, going up is slow, but at least I'll be working hard. Going down, however, is more a matter of care in foot placement. I expect I'll be passed here by runners with more experience on this sort of terrain.

After pausing a few times, confused over side paths labeled "Coastal Trail" which were actually out-and-back spurs to the water, I arrived sooner than expected at the Sutro Baths. Wow -- I've lived in San Francisco for most of ten years but while I've been to nearby Lands End several times I'd never before visited the ruins of the Sutro Baths. Really spectacular: I felt as if I was back in Greece. I was glad I had pre-run this so I had a chance to stop and admire the scene, rather than being tempted to do so during next weekend's race.

With the pause, I realized I was a bit tired, but once I started the climb of the steps which had led down to the Baths I started to feel good. Except for a small loop at the baths, the course is a pure out-and-back. Fortunately there's plenty of room on the Coastal Trail approaching the turn-around except for a few staircases on which people will be going slowly anyway.

Back at Baker Beach, I decided to go off-menu and try the Sand Ladder. This was fun: having not done a hard swim and ride beforehand, not to mention my relatively relaxed pace on this run, helped me avoid the struggle experienced by so many during the triathlon.

Returning to the bridge, I took a wrong turn, however. I tried to follow a sign for "hikers" and ended up on Merchant Road. The name is misleading: there's no merchants here, just way too many cars. Fortunately, I know the way well from cycling, so was able to get back on course.

I had a considerable tailwind which pushed me the rest of the way to Sports Basement. Remember during the race there's still that lap of Marina Green which I'd run at the start.

I'm really looking forward to next Saturday.

Wednesday, June 8, 2011

watching Escape from Alcatraz

The weather cooperated on Sunday morning as the feared thunderstorms never materialized. I thus put on my cycling clothes and rolled down Potrero Hill to begin my ride to the Marina to watch the triathlon.

The island. From EscapeFromAlcatraz2011

I arrived at 7:10 am, well before the 7:30 am start. Alcatraz island was visible through the slightly misty air. There was plenty of activity in the water, small craft to guide the swimmers through the strong current. It was definitely an exciting time: within a short time I knew that thousands of participants would bravely jump into the freezing Bay and begin swimming towards the shore more than a mile distant. I wondered if any had ever bailed out when crunch time came.

The bike: no problem. The run: no problem. The swim: I couldn't imagine swimming in such cold open water that far. I'd be lucky to be able to breathe again after the thermal shock of plunging into the icey waters. Wet suits can only do so much...

Escape from Alcatraz likes being punctual, and this year was no different, with participants getting wet right on schedule starting at 7:30. From my viewpoint there was the faint hint of activity in the water (including a claimed seal sighting) until, approaching 20 minutes after start, the swarm of splashing color became distinct. And it didn't look good: they were clearly overshooting the beach. Could they recover in time or would they be swept out through the Golden Gate to the pleasure of the waiting sharks? I couldn't tell.

The swarm. From EscapeFromAlcatraz2011

Yet then two deviant splashing bodies came into view. Dustin McLarty and Andy Potts were on a perfect line to the upstream corner of the beach. I thought back to 8th grade algebra: someone can swim at speed x and run at speed y, what's the optimal point to land on the shore to go from a point r0 offshore to a point r1 inland? These guys run around four times faster than they swim, so obviously you want to hit the relatively short beach as soon as possible. Dustin went through the timing arch in an official time of 22:29, with Steve right on his heals only two seconds down. Meanwhile the current-swept hoards were frantically trying to recover against the rushing waters of Golden Gate.

I later asked Dustin how he'd managed to take a better line than the others. "I guess I have more open water experience than most of them," he replied.

The main wave of swimmers lands on the beach. From EscapeFromAlcatraz2011

Fully 61 seconds after Steve followed Dustin through the arch, Andrew McCartey crossed in 23:32 for third place on the leg. Despite being first out of the water Dustin would fade by the end of the race to 14th. Andrew would finish eighth. But Andy Potts would exploit his excellent swim to win Escape from Alcatraz by 34 seconds over second place Bevan Docherty, who finished 9th in the swim, 76 seconds slower than Potts, in that second wave of finishers. So while in this swim-heavy triathlon the swim didn't dominate the results, a smarter line may well have been what gave Potts his win over Docherty. Or maybe most of that time simply came from Potts' superior water speed.

It was really incredible watching the swimmers energe from the freezing waters. There was a seeming endless wave of them. They varied in how far they overshot the beach, some actually needing to swim fully upstream against the strong current to make the landing zone, but more typically having to just hold off the influence of the flow and sustain an orthoganal trajectory to the coast. I asked a spectator, the wife of a first-wave competitor, whether the currents were weaker approaching the shore, thinking maybe there was a stagnation region. No, she responded, it's stronger there.

With swimmers still arriving in huge numbers, I gave up my viewing spot on the crowded transition region, and decided to watch the next "leg" of this race. This wasn't the bike, but rather the run from the shore to the bicycles, a considerable distance. Most runners did a double transition: running shoes just past the beach, the long run to the bikes, then change into cycling shoes for the bike leg. The best times I see among the top forty eventual finishers were 3:43 for Docherty (2nd overall) and Matt Chrabot (3rd overall) with Potts the only other to break four minutes with 3:55. It's curious the T1 podium seems to match the final podium ignorning order. I suspect each of these guys did the run barefoot, as did some of the runners I saw later. Good calluses could be said to be the key to winning here.

Most participants were considerably slower on T1 than these first three. The running pace generally seemed fairly relaxed. Most were likely just trying to recover from the numbing, tiring swim. Meanwhile in the opposite direction riders were beginning their bike leg.

The "cyclist" view of triathlon tends to be people with way too much disposable income riding their $10k P4's, Spiz, and Trinities with an inflated cost-to-fitness ratio and with body positions which squander the fractional advantage gained from their carbon über-frames. But the reality here is while most of the "pros" were justifiably on high-end rides, more pedestrian bikes were common even among riders who were well-placed. I was forced to rethink some harsh stereotypes.

I set out alongside the bike course to find a good viewing spot for the returning leaders. The bike leg is only 18 miles long, which I estimated at more then 40 kph should take the lead riders around 40 minutes. This turned out to be optimistic: Andy Potts turned the second-fastest bike leg on his slick Kestrel with 46:35. But as he passed me, he was absolutely flying. He looked super-smooth in his aero tuck as he extended the advantage he'd gained in the swim. Dustin McLarty, who'd been first out of the water, was 4:59 slower.

Approaching the finish of the bike leg. From EscapeFromAlcatraz2011

As riders would pass I'd shout their placing to them. It got hard to keep track once I'd gotten up to thirty, so I gave up and move on up the bike course to try and intercept the runners, who were running on a parallel trail closer to the water.

Lead woman on the bike. Note the conspicuous lack of dedicated time trial equipment other than clamp-on aero bars. From EscapeFromAlcatraz2011

I reached the stairs at Fort Point seconds after Potts had passed, so I missed him. But I was ready for the rest. I was pleased to see fellow Roaring Mouse Larry Rosa here with his camera. Larry was doing commercial work for the race and if he was here that meant it must be a prime viewing spot. He does really superb work.

Taking the stairs two at a time. From EscapeFromAlcatraz2011

I was a bit surprised to see the runners I watched take the stairs two at a time. I'm usually pretty good going uphill (not at the level of these guys!) and find it's better for me to "downshift" and take stairs like these one-by-one. I feel the increased load from bounding up them in pairs needs to be repaid with interest later. But I defer to their judgement. I'll be able to test my theory when I do the Zombie Runner Half-Marathon on the same steps in ten days. I'll try the steps one-by-one and see if I can keep up with people doing the double.

I then climbed up from Fort Point on the nearby road to find the famous "sand ladder", the most challenging part of the run. On the road above the stairs riders were going in each direction... only a few still heading out but many, many more returning to T2. The anti-drafting rule was completely unenforcable here, but I didn't see anyone taking advantage. Again I was impressed: my view of triathletes had been that most of them do the wink-wink, nudge-nudge on the anti-drafting rules. Yet I didn't see a single case of this.

I encouraged riders climbing the steep hill before the final descent, then parked my bike near the exit of the sand ladder. A saw a few of the runners near the front pass through here. I initially assumed it was off-limits to go down to see the ladder itself, but I was given permission to do so by the volunteer.

There were a few photographers there, but to my amazement nobody else. It was an extremely cool viewing spot. Everyone, without exception, looked beaten down to some degree. The climb is steep and the sand saps energy. Again this seems to me like a "shift down" moment... save it for what follows. Some of the competitors were pulling themselves by the cable, following the advice which had been offered at the previous day's briefing. All were walking, which had also been advised, although I'm not sure if that was always by choice.

The sighted member of the tandem team leads his site-impared teammate up the feared sand ladder. From EscapeFromAlcatraz2011

The guide cable along the sand ladder provided support for tired legs. From EscapeFromAlcatraz2011

After I'd seen enough of this suffering, I ran up the upper, shallow grade portion of the sand ladder I'd previously descended, got on my bike, and returned to the start-finish. A large number of participants were descending at this point.

The Stanford women's relay team shares the joy of crossing the finish line. From EscapeFromAlcatraz2011

At the finish I found a spot in the grandstand erected there to watch the Stanford women's relay team run together to the line. Very cool.

But the finish line didn't really interest me: the real action was out on the course. So I left the stands, took a final tour of the expo region, then went to catch the 10:30 AM scheduled meeting of the Mission Cycling ride at the Golden Gate Bridge.. along the way I stopped to cheer more of the runners.

Results are here.

Monday, June 6, 2011

Running into the Escape from Alcatraz Expo

After the storm passed through on Saturday, I decided to get a run in as it was 10 days since my last one. So I set off on my usual route the Embarcadero (after stopping to watch the Giants game for a bit), Fort Mason, and Marina Green. I'd planned on coming back on Filbert or maybe Lyon. But approaching the Green I passed first one then another rider on time trial frames with numbers markered on their calves. No doubt about it, there was a triathlon in town.

I came across the Expo in the Green itself. I had to stop to check it out. Trithletes were everywhere. I hoped nobody confused me for one. I'm a cyclist, not a triathlete. Sure, I've been running a lot lately, but there's a big difference between someone who mixes running with cyclist and a "multisport athlete". It goes way beyond the time period which expires between the activities; it's a deeply cultural difference.

Trek Speed Concept time trial frame: the head tube almost disappears from a head-on view, even with the front wheel slightly turned. From EscapeFromAlcatraz2011

The Scott Foil mass-start road frame has clean cable lines and a contoured head tube, but has a far more subdued aerodynamic character than that Trek time trial frame. From EscapeFromAlcatraz2011

Cervelo P4 at the Front of the Pack booth... this frame is always impressive, but I must admit the Trek has displaced it on the "coolness" scale. From EscapeFromAlcatraz2011

Highlights of the expo were first the Trek time trial frame -- the thing simply disappears from view head-on, and the new Scott Foil. I've written about that bike here but had never before seen one in the flesh. Clean cable lines, some contouring at the head tube, and sure enough the "Kamm tail" design which looks like a truncated parabola tube cross-section at the down tube. But assuming the test data is good, it hides its "aeroness" well. I approve: it seems to be a nice all-around bike.

So after stopping for a few samples (tasty Enervit chews, Cytomax which apparently switched from acesulfate-K to stevia) I was ready to go.

But then they announced an "athlete briefing". I wasn't in a rush to get back home so I decided to sit in. I stood off to the side. Again I didn't want anyone thinking I was actually in the race.

The briefing was interesting. Participants were warned about the 52-54F water and about the strong western current ("Current is to the west; swim south; magically you'll go southwest"). Then they were warned about a rough descent passing the Legion of honor on the bike route. And finally they were advised to walk the famous Sand Ladder at Baker Beach, using the cables for support. Wow -- what an interesting course, I decided. I'm coming back tomorrow...

And so I continued my run. But as I approached my planned turn-off, I couldn't resist following the triathlon course a bit further. Soon this was turning into quite a lengthy run. When I saw 25 km on my Garmin, I extended it a bit further still to take it over 30 km. I saw 30.3 km when I'd gotten back home. I was to pay for my enthusiasm a bit the next day when I was still a bit sore, but not too bad. I felt good I'd not suffered any real pain on the long run.

Sunday, June 5, 2011

Golden Gate Bridge PTSD

After watching Escape from Alcatraz triathlon (epic, epic, epic) I joined up with what portion of the Mission Cycling ride didn't flee south today and headed into Marin for a Paradise Loop. I wanted to work the kinks out of my legs after an over-enthusiastic 30 km road run yesterday.

But to get into and out of Marin, we had to ride over the Golden Gate Bridge. Little did we know what horrors awaited...

Starting May 31, the Golden Gate Transportation District shut down the western cycling path on the bridge for seismic retrofit work. As a result, cyclists must now share the eastern path with pedestrians for close to four months, the season of peak use of the bridge by both pedestrians and cyclists. The timing couldn't be any worse.

I remember the first time I ever crossed the western path. It was quite a shock, leaving me with a survivors rush. But over time I got to know the western path, became familiar with the order behind the seeming randomness, and it became an ordinary part of weekend rides ... even enjoyable with its spectacular views.

Here's what Bike Snob recently wrote about the cycling path:

If you're unfamiliar with the Golden Gate Bridge, it is a bridge that links the city of San Francisco and the profoundly smug county of Marin, and it also happens to remind me of a Samurai sword, albeit a really big one that spans a large body of water. Anyway, I heard about this proposal during my leave of absence last week, and I would agree that a 10mph speed limit is absurd. However, having ridden over the Golden Gate Bridge a number of times, I would also say that it's like someone took all the most annoying elements of the Brooklyn, Manhattan, Williamsburg, and George Washington Bridges and forced them onto a single path. Freds, tourists, wobbly triathletes... You haven't experienced fear until you've been approached head-on by an oblivious tourist on a rental bike, weaving as he simultaneously smokes a cigarette and attempts to take a photograph of one of the towers, while you brace yourself for a collision that could send you hurtling into the icy waters below.

But with the diversion of cyclists to the eastern path it's gotten worse. Much worse.

I'd hoped the increase in overt danger would yield a compensating increased awareness by tourists, both on foot and on their blazing saddles.... but there is no evidence this has occurred. I rediscovered, with interest, that feeling from my first ride. Survival? Technically yes, but senses overloaded...

And the number of rental bikes was substantially below normal today. The forecast of possible thunder showers, rare for June, had no doubt taken a toll. Next week there will no doubt be more... and more on foot as well.

The consensus among the Mission ride participants is it's time to take a break from Marin.

The disturbing thing about all of this is San Francisco has a "transit first" policy which at least nominally says that "travel ... by bicycle and on foot must be an attractive alternative to travel by private automobile." Combining cyclists with pedestrians on what is the only viable transportation link between San Francisco and Marin counties, while maintaining full auto capacity couldn't be a clearer violation. The problem is the Golden Gate Transit District isn't part of San Francisco: it has no "transit first" policy and it's actions speak, over and over, that its priority is drivers.

I absolutely guarantee recreational cyclists, most strongly affected by this closure since pedestrian traffic is heaviest on weekends, will either go elsewhere or will drive across the bridge to begin rides on the other side. This will result in increased auto traffic on the bridge and on surrounding roads, more strain on parking in the Headlands and in Sausalito, and more noise and pollution for everyone.

The solution? If the western path absolutely must be closed, and I don't accept that it must without seeing the plans, then shut down an auto lane and divert the cyclists to that. Then the pedestrians can keep their eastern path. "Can't be done", the Golden Gate Transit district reponds.... cyclists can't cross the expansion grates on the bridge. However, this issue has been addressed before:

Tour of California crosses Golden Gate Bridge
Tour of California crosses the Golden Gate Bridge in 2009

It's time San Francisco demand that the Golden Gate Transit district adopt the same "transit first" policy in the City Charter. At least then "plans" like this can be judged to the standard that bike access is fundamental to a well-working transportation network.