What is the speed advantage of a wetsuit?

Everyone knows wetsuits help keep you warm in cold water. Lesser known among the general public (but well-known among triathletes) is that wetsuits also make you swim faster! The buoyant neoprene in a wetsuit floats a swimmer higher in the water, decreasing drag and thus increasing swim speed.

But how much faster is a wetsuit? I’ve heard various rules of thumb: 10% speed increase; 4-6 seconds per 100m; 1 minute per kilometer. I’ve also heard various caveats: it depends on the swimmer’s skill (better swimmers benefit less); it depends on the swimmer’s body-type (naturally floaty people benefit less); it depends on the quality of the wetsuit (you get what you pay for); it depends on the fit of the wetsuit; and so forth.

So the answer is: It depends. Because I’m usually disinclined to let things go at “it depends,” I decided to conduct a field experiment. Reef & Run, which I’ve written about previously, provided the perfect laboratory. Almost every Thursday evening between June 21 and yesterday, August 23, I swam one mile in the ocean at East Beach in Santa Barbara. Two weeks were canceled because of shark sightings, and one week I was sick – leaving a sample of 7 swims.

The swim took place at the same time each Thursday: 6:30pm. The conditions were generally similar: low-mid 60s water temp; winds out of the W or SW, producing moderate surface chop and a W-to-E current (i.e., head current going out, tail current coming back). I would characterize them as “rough water conditions” – the view in the above photo is typical. The course was identical each week – a full mile (1609 meters) measured with GPS, and marked by permanently installed buoys.

Generally, I had done a full workout earlier in the day, plus one lap of the course as warm-up. So, for each of these races I was warmed-up but perhaps a touch fatigued. In any case – pretty close to an ideal setup for my field experiment.

My wetsuit is a cheap-o XTERRA Vortex sleeveless, which frankly doesn’t fit me very well. So – a conservative test of the wetsuit effect. Presumably, I would be even faster in a high-end, well-fitting, full-body wetsuit.

Me in the orange cap. Three-time Olympic water polo player Wolf Wigo at left. The two others in the photo were doing a different race. Photo by Mike Eliason, Santa Barbara News-Press

Of the seven races, I wore a wetsuit for four of them and went “naked” for three of them.

My wetsuit-assisted times were: 19:52*, 20:02, 20:14, and 20:14.

My “naked” times were: 21:36, 21:37, and 21:41.

* For the purpose of this analysis, I’m throwing out the 19:52 wetsuit-assisted time. That was the season opener, and it was different in several respects: gorgeous, flat conditions; bigger, more competitive field (thus more drafting opportunities). I’m not surprised I was substantially faster that week.

That leaves a sample of six times – three wetsuit-assisted and three “naked.” My average wetsuit time was 20:10, with a range of 12 seconds. My average “naked” time was 21:38, with a range of 5 seconds.

So, according to my field experiment, my personal “wetsuit effect” – even with an ill-fitting cheap-o sleeveless – was 1 minute, 28 seconds in an open-water mile. That converts to 5.5 seconds per 100m, or a speed effect of 7.3%.

Any other self-experimenters out there? Please leave your data in the comments!

Stroke count games

What’s the fewest number of strokes you can take for a single length of the pool? (No streamlining past the flags; no more than three kicks per stroke.)

I can get down to 8 strokes per 25 SCY, but it’s tough to sustain for more than one length. 9 strokes per length (SPL) I can do pretty much indefinitely – but it’s incredibly inefficient. The inefficiency is readily apparent: a huge dead spot in my momentum as I glide (glide, glide…) after each stroke. The Swim Smooth guys have a term for this: Overgliding.

I swim most efficiently between 13 and 15 SPL, depending on pace. 13 for channel/marathon pace; 14 for “threshold” pace (from the mile up to about 5K); 15 for 200/500 pace. For an all-out sprint, I’ll add one more stroke (16 SPL).

Experienced pool swimmers have an intuitive feel for this… but what if you don’t? Is there a formula to identify the most efficient stroke count for a given pace? This question led me to try the following set:

8×100, as fast as possible, with about a minute rest between each. But there’s a twist: Within each 100, hold a constant SPL. The first should be your lowest sustainable SPL. On each subsequent 100, add one SPL. So for me, #1 is 9 SPL, and #8 is 16 SPL. Record all your times. The set is best done short-course (it’s tougher to control SPL so tightly in a long-course pool).

Here are my results:

SPL time
9    1:20
10   1:14
11   1:10
12   1:07
13   1:05
14   1:02
15   1:00
16   1:01

Further evidence of the inefficiency of minimum stroke-count: At 9 SPL, the fastest I could go was 1:20! Merely by adding two strokes per length (8 total), I was able to go 10 seconds faster.

[Warning: Geeky content ahead.]

Does this remind you of anything? That’s right – it’s sort of like SWOLF! So, let’s sum the two columns above to produce SWOLF scores:

9    1:20  116  [(9*4) + 80]
10   1:14  114  [(10*4) + 74]
11   1:10  114  etc.
12   1:07  115
13   1:05  117
14   1:02  118
15   1:00  120
16   1:01  121

SWOLF doesn’t quite get it right. According to SWOLF, my efficiency peaks at 10/11 SPL (1:14 & 1:10 per 100y) – which I know to be false. Even at cool-down pace, 10 SPL is more taxing than 12-13 SPL, due to the constant stop/start motion of overgliding.

The key here is: I already know my most efficient SPL is 13-15, depending on pace. I have the result; so what’s the formula that produces it?

Given the above data, the problem with the original SWOLF formula (time in seconds + total number of strokes) seems to be that it overvalues stroke count (and by corollary, undervalues speed). So I just tried the simplest thing I could think of: dividing stroke-count by two (thus reducing its importance in the final equation).

(Stroke Count / 2) + Time = modified SWOLF

SPL time SWOLF(mod)
9    1:20   98  [(9*4/2) + 80]
10   1:14   94  [(10*4/2) + 74]
11   1:10   92
12   1:07   91
13   1:05   91
14   1:02   90
15   1:00   90
16   1:01   93


What’s funny about my “discovery” is that there’s another term for “stroke count divided by 2” — stroke cycles. Which incidentally is exactly how the Swimsense calculates SWOLF. So perhaps our friends at FINIS were on to something?

If any readers out there want to try this set and report back, I’d be grateful for additional data. Does modified SWOLF find your most efficient stroke count(s)?

Fat vs. Fast

There’s an old saying about cold-water marathon swimming:

Either be fat, or be fast.

Is it oversimplified? Probably. Crass? Definitely. But there’s a kernel of truth worth examining. Thin swimmers have made it across the English Channel, but they’re usually fast. Slow swimmers have made it across the Channel, but they’re usually… carrying a healthy layer of bioprene.

The common factor: Core temperature must be preserved. Either generate heat, or retain it. Fast swimmers are good at generating heat. Fat swimmers are good at retaining it.

In the English Channel (from what I gather), it’s considered prudent for non-overweight swimmers to put on some weight, even if they’re “fast.” A Channel attempt is expensive and, unless your name is Petar Stoychev, just getting across is the main priority. Bioprene increases the probability of success.

But at what cost? How much does the extra weight slow you down? Swimming is a gravity-less activity, so obviously it matters less than in running or uphill cycling. Further, the flotational benefits of fat may improve your body position in the water.

In running, the rule of thumb is 2 seconds (faster) per mile per pound (lost). Is there a similar rule of thumb for swimming?

Out of curiosity, I asked Coach AB to estimate the benefit of losing 10 pounds of body fat on threshold pace per 100m (assuming stable fitness & muscle mass). He said 2-3 seconds per 100m. Some quick conversions: 32-48 seconds per mile, 10-16 minutes per 20-mile channel swim. Or, for an apples-to-apples comparison with running: 3.2-4.8 seconds per mile, per pound.

And actually… that accords fairly well with my own experience. I do a lot of threshold (a.k.a. CSS) training – so I’m intimately familiar with my basic pace per 100m. Also, my weight has fluctuated a bit in the past couple years – giving me some data to draw on.

Can we do better than a rule of thumb? Scientists being scientists, it turns out someone has actually studied this question. In a paper published in the Journal of Strength and Conditioning Research, Ilka Lowensteyn and two colleagues artificially varied the body fat of competitive swimmers by fitting them with weighted latex pads under a spandex triathlon suit. The swimmers were timed at 50-yard sprints at various weights.

Lowensteyn et al. estimated the swimmers were slowed by 0.2 seconds per 50 yards, per pound. That’s 4 seconds per 100, per 10 pounds – not far off Coach AB’s estimate. And it makes sense there would be a larger effect in a sprint (compared to threshold pace), because in water, drag increases exponentially with speed.

Bottom line: Let’s say you gain 20 pounds for your English Channel attempt. You might be looking at about an extra half-hour in the water. Given the thermal-protective benefits of those 20 pounds, though, it seems like a small price to pay.

Water temperature in the Catalina Channel

There are 14 years of publicly available data on the surface water temperature in the Catalina (a.k.a. San Pedro) Channel – via NOAA and CDIP. Unfortunately, that’s all it is – data. No summary statistics, no long-term charts – nothing particularly useful if you’re just looking for a simple, big-picture view of trends and cycles in sea temperature (perhaps to inform your upcoming swim across the channel).

So I decided to make one myself:

Catalina Channel water temperature, 1998-2012

NOAA buoys take readings every 30 minutes. Over 14 years, that works out to almost 239,000 observations. Don’t try this on an old computer! For a smoother line, I calculated a weekly average. Same data – just prettier.

If you really need more detail, I also made an interactive chart with daily-level resolution (5,044 observations). Keep in mind, Javascript is required to view the chart, and it probably won’t look good on mobile devices. If you’ve ever used Google Finance to view stock prices, the chart format will look familiar.

Summary Statistics by Day of Year

Sea temperature varies by season, but there are also year-to-year variations. In 2010, for example, the Catalina Channel was unusually cool (even in summer). In 2006 it was unusually warm. Perhaps you’ve wondered: What is the typical water temperature on a given day of the year? If your swim is scheduled for August 15, what is the average water temperature on August 15, averaged across all years?

To answer that question, I made this chart (click to enlarge):

Most Catalina swims take place in summer and early fall (not winter or spring), so here’s a zoomed version of the same data, for the swim season only:

So, water temps in the Catalina Channel tend to peak around August 1, and remain more or less steady through the first week of September. But even in early June and late October, the water is still “warm” by English Channel standards.

Note: It’s important to remember that surface water temps in Southern California tend to drop a few degrees as one approaches the coast, due to upwelling from the steeply sloping ocean bottom. My understanding is that this tends to happen about 3 miles from shore. So, if the buoy reading (6.5 miles offshore) is 63 degrees, the actual surface temp might actually be sub-60 during the last part of your swim.

Venus, Mars, and Catalina

Previously, we’ve looked at some general stats on Catalina Channel finishing times, and the growth in participation since George Young’s pioneering swim in 1927. What about gender differences? (Taking a page from Katie’s playbook…)

From 1927-2004, there were 90 successful swims by men and 44 successful swims by women (a ratio of 2.05 to 1). From 2005-2011, there were 80 successful swims by men and 49 successful swims by women (a ratio of 1.63 to 1). So, the gap is narrowing…a bit.

Here, again, it would interesting to see the data on failed swims. Is the ratio of men to women the same for failed swims as for successful swims?

Side note: I decided to split the data-set at 2005 because it offered similarly-sized groupings, and because this was the year when there was a surge in popularity of Catalina Channel swimming (possibly due to the advent of the “triple crown”).

And here are the average & median finish times for each group (C-M one-way crossings only):

Average Median
Men 1927-2004 13:14 12:14
Women 1927-2004 12:17 11:03
Men 2005-2011 11:23 10:51
Women 2005-2011 11:00 10:39

In both eras, women are faster – despite lower levels of participation. Perhaps we shouldn’t be surprised, given that women have the overall records in both directions – Karen Burton from Catalina (7:43) and Penny Lee Dean from the mainland (7:15). Interestingly, in my analysis of MIMS times I also found women were almost uniformly faster.

This raises an obvious question without an obvious answer: Why? (See the comments section for a couple theories.)

The third in a series of posts taking a statistical look at the history of Catalina Channel swimming (see parts 1 and 2). These analyses have not been validated or endorsed by the Catalina Channel Swimming Federation and should be considered “unofficial.” 2011 swims are included, but are unofficial until the ratification banquet on November 5.

CCSF’s official list of successful swims is available here. Penny Lee Dean’s authoritative history is here.

Catalina Channel stats: An epidemiological view

The second in a series of posts taking a statistical look at the history of Catalina Channel swimming. These analyses have not been validated or endorsed by the Catalina Channel Swimming Federation and should be considered “unofficial.” 2011 swims are included, but are unofficial until the ratification banquet on November 5.

CCSF’s official list of successful swims is available here. Penny Lee Dean’s authoritative history is here.

On January 15, 1927, George Young was the only one of 102 participants to finish the Wrigley Ocean Marathon, and in so doing, became the first person to swim across the Catalina Channel. For his achievement Young earned a $25,000 prize – approximately $325,000 in 2011 dollars, and richer (even in nominal dollars) than any current cash prize in professional marathon swimming.

Seven of the DNF’s in the Wrigley Ocean Marathon – four men and three women – returned later that year to try again; four finished. But Catalina Channel swimming didn’t catch on after this rousing first year. Over the next 25 years only two more swimmers added their names to the list. Despite a brief resurgence in the late 1970’s (including double-crossings by Penny Lee Dean, Cindy Cleveland, Dan Slosberg, and John York), the typical number of calendar-year crossings was still 5 or fewer into the mid-2000’s.

Catalina Channel solo crossings per year, 1927-2011

Then it took off. In 2005, 12 swimmers crossed the Channel. Followed in subsequent years by 13, 8, 25, 16, and 29 crossings. So far in 2011, there have been 22. What happened? My guess would be the marketing of the “Triple Crown.”

Catalina Channel solo crossings - Cumulative


Catalina Channel: A history in numbers

The first in a series of posts taking a statistical look at the history of Catalina Channel swimming. These analyses have not been validated or endorsed by the Catalina Channel Swimming Federation and should be considered “unofficial.” 2011 swims are included, but are unofficial until the ratification banquet on November 5. CCSF’s official list of successful swims is available here.

I should note that Penny Lee Dean did some similar statistical work in her authoritative History of the Catalina Channel Swims Since 1927. However, the book has not been updated in 1996, and in any case, the stats chapter seems to have been removed from the online version.

The Catalina Channel was first conquered in 1927 by George Young of Canada, in 15 hours, 44 minutes, 30 seconds. Since then (through September 2011) there have been 259 successful solo crossings by 220 individuals, including 7 double-crossings.

The short list of double-crossers includes some of the greatest marathon swimmers in history.

From the mainland (M-C-M):

  • John York – 16:42 in 1978
  • Dan Slosberg – 19:32 in 1978
  • Tina Neill – 22:02 in 2008
  • Cindy Cleveland – 24:30 in 1977

From Catalina (C-M-C):

  • Penny Lee Dean – 20:03 in 1977
  • Forrest Nelson – 23:01 in 2010
  • Greta Anderson – 26:53 in 1958

Of the 252 one-way crossings, only 19 went from the mainland to Catalina (M-C). Penny Lee Dean still holds the overall record for this direction: 7:15 in 1976. With the exception of the Swim 22 relay last year, there hasn’t been a one-way M-C crossing since 1977. The most recent M-C crossing was achieved by Suzie Dods in 2010.

The remaining 233 one-way crossings started at Catalina and finished on the mainland (C-M). The 10 fastest C-M crossings are as follows:

  • Karen Burton – 7:43 in 1994
  • Todd Robinson – 8:05 in 2009
  • Hank Wise – 8:07 in 2010
  • Chad Hundeby – 8:14 in 1993
  • Blair Cannon – 8:18 in 2011
  • Gemma Jensen – 8:20 in 2006
  • Jim McConica – 8:27 in 1983
  • Rendy Lynn Opdycke – 8:28 in 2008
  • John York – 8:32 in 2000
  • Penny Lee Dean – 8:33 in 1977 (first leg of a C-M-C double)

My 8:55:59 ranks as the 24th-fastest swim, a mere 11 seconds ahead of David Blanke, Elizabeth Fry, and Marcia Cleveland’s tandem crossing in 2005.

The slowest C-M crossing was achieved by Paul Chotteau of France in 1936 – a herculean 33 hours, 50 minutes! The median C-M crossing is 11 hours, 10 minutes.

Of the 240 C-M crossings (including legs en route to a double):

  • One was faster than 8 hours (Karen Burton);
  • 26 were between 8 and 9 hours;
  • 38 were between 9 and 10 hours;
  • 51 were between 10 and 11 hours;
  • 32 were between 11 and 12 hours;
  • 26 were between 12 and 13 hours;
  • 23 were between 13 and 14 hours;
  • 8 were between 14 and 15 hours;
  • 17 were between 15 and 16 hours;
  • 10 were between 16 and 20 hours;
  • and 8 were longer than 20 hours – the most recent being Jamshid Khajavi of Iran in 1995 (20:47).

Four swimmers have crossed the channel using a stroke other than what we now call “freestyle”:

  • Henry Sullivan – 22:45 breaststroke in 1927
  • Vicki Keith – 14:53 butterfly in 1989
  • Tina Neill – 10:37 backstroke in 2008
  • Jason Lassen – 15:59 breaststroke in 2010

To be continued…