Wednesday, March 14, 2007

Selling Short

A month or two ago, in response to a post about an interview with UCF head coach Leeann Crain in which the lights were not mentioned, a couple of UCF rowers showed commendable loyalty and team spirit by commenting that they understood why the heavies were getting the spotlight and that they deserved it. One noted that "we did get 5th out of 12 teams and the heavies were ranked 20th out of well over 100 teams." While I certainly agree that the UCF heavies had a commendable season, I think this UCF lightweight is selling herself short. Let me tell you why.

First, given the genesis of this discussion, let me define a "competitive" crew as one with a credible chance to make the grand final at IRAs or NCAAs. In addition, at NCAAs, I am only talking about the DI grand final (more on that later). I am also going to make the assumption that boat speed among both lightweights and heavyweights is normally distributed. I haven't done the analysis, but I think this is a reasonable assumption. In any given year, particularly among lightweights, speed may not be normally distributed, but as a working assumption, I'm going with it. (Even if I'm wrong about this, it doesn't necessarily harm the theory because this assumption is used for illustrative, not probative, purposes.)

The NCAA Web site lists 86 DI schools with women's crew, 15 DII schools, and 43 DIII schools, for 144 total programs. Of those 144, 16 eights make it to NCAAs (DI eights; the assumption is that the DI eights are the fastest boats), and 6 make it to the final. Nonetheless, let's say that 20 (with a nod to UCF) heavyweight boats are "competitive." Twenty out of 144 is about 14%. That means that those boats are all more than one standard deviation from the mean heavyweight boat speed (in a normal distribution, 15.8% of boats would be one standard deviation or more faster than the mean). The rest of the heavyweight universe, which in a given year is slower than one standard deviation better than the mean, is, to use an honest but crass term, cannon fodder for the fast DI boats. I have more experience with (in) cannon fodder crews than I would like, and I can tell you that victories are just as glorious and defeats are just as painful, but you spend your season trying to win local duals, state championships, and make the finals at regional championships. You compare yourself with peers who, for one reason or another, are not those crews that are nationally competitive. You plug along, looking for that marginal improvement while hoping for a leap forward that finally makes you competitive. For many it comes, and there is movement into and out of cannon fodder territory.

The chart below illustrates this distribution.
(Recognizing the cannon fodder problem, the NCAA introduced DII and DIII championships, moving those programs onto their own curves, even as it killed serious DIII lightweight rowing. Because lightweights don't have separate divisions, however, I'm ignoring the distinction among heavyweights as well.)

I also assume that lightweights are normally distributed and that there are cannon fodder lightweight boats as well. As you move back from one standard deviation, however, something odd happens. (OK, now we're into theory, and I have no proof, but hear me out.) As the season progresses, the cannon fodder boats disappear. Sure, the 12th place boat at IRAs would get hammered by the 1st place boat, but the same is true of the heavyweights at NCAAs. The real fodder among the lightweights simply vanishes by the end of the season. Because I believe in the law of conservation of mass, I know those athletes have to go somewhere, and where they go, is into heavyweight crews. That's right, by IRAs most lightweight crews have self-selected themselves out of competition and the athletes are rowing in heavyweight boats. Most programs don't even try a lightweight eight because they know they can't be competitive with the Wisconsins of the world, and there is no other lightweight level. As the top lightweight crews plow through early season lightweight hopefuls, those defeated boats realize they're not competitive and fall back into the least restrictive category, and therefore the category that will always be available - heavyweight.

Without lower level championships, lightweight rowing is all or nothing. Yes, there's Dad Vail, but the Dad Vail champ can be pretty competitive at IRAs, meaning that it's a really fast boat. A lightweight boat of moderate speed doesn't have the option of satisfying itself with a DIII championship because there is no DIII championship. The coaches of these boats realize that it would be a waste of time and money to race them as lightweights all season and watch them get their brains beat in. The better option is to take the best rowers out of those boats, put them in the heavyweight varsity or 2V, and make those boats faster. The heavyweights MUST exist, but the lightweights are optional. If the optional boat is slow, you get rid of it and use all of your resources to make the mandatory boat less slow.

The chart below illustrates this, with a line drawn in that I'll call (because I'm an optimist) the Line of Hope. To the left of that line are those lightweight boats that have abandoned hope and moved their rowers into heavyweight boats. (We can argue about the placement of this line, but it exists.) In between the competitive line and the Line of Hope are the hopefuls, those crews hoping they can pull off the upset of upsets.

If the boats at IRAs are really the cream of the lightweight crop, rather than the only available boats, the time spreads from 1st to last should be similar to those spreads among the heavyweight cream of the crop. In 2005 and 2004 things didn't look so good for my theory. The spread between the winning boat and the sixth place boat was 34 and 37 seconds, while among heavies it was 6 and 8 seconds. In 2006 (the year that prompted this post), however, things changed. First to 6th among lights was 14.5 seconds and 1st to 6th among heavies was 14.4 seconds. From 1st to 12th among lights the spread was 37 seconds and from 1st to 6th among heavies the spread was 26 seconds. A dominant Princeton heavyweight boat may have skewed this a bit, but I think we can feel confident that the 2006 lightweight national championship was as competitive as the heavyweight championship.

You see, lightweights, this notion that the lightweight champ is the champion of a very small field is a canard. When you reach the mountaintop first, you don't look around and feel diminished because only 12 people came up behind you. You also count the scores of bodies lying on the slopes below. In lightweight rowing, the last breaths of the defeated are spent changing into heavyweight unis.

Is there room for improvement? Of course. There are still plenty of outstanding lightweight rowers who choose to row in good heavyweight programs. With a stronger lightweight league, perhaps those athletes would row as lightweights. Quite honestly, the lightweight spread should be tighter than the heavyweight spread and if we could get more programs to focus on lightweights it would be. But don't let heavyweights, heavyweight coaches, and the mainstream rowing media dictate your thoughts on this. If heavyweights didn't make fun of you, and everything else, they wouldn't be worth the label "athlete." Likewise, you have every reason to look down your nose at them. This is healthy. This is believing in yourself. This is knowing that what you do is special. This is knowing that you're pulling a 230 pound boat and a 120 pound coxswain down the river while weighing 130 pounds instead of 180 or 200 pounds. Lightweights and heavyweights should fight just like brothers and sisters because in the end, we're all one family. Just don't ever start believing everything your big sisters tell you. As for the rowing media, well, I think we've already seen that they're thought followers, not thought leaders.

Stop selling yourself short. The lightweight national championship is every bit as precious as the heavyweights' and you're about to get another chance to win it. The time is now - get off the erg, get on the water, and kick some ass.


Mahalo said...

I will give you credit that "Top End" lightweight women are fast. Let' say top 3-4 at IRA's. The fastest lightweight womens crew would be 15 sec behind the last place crew at NCAA's. If you want proof look at the Califorina Cup at San Diego. Wisco women could not make the finals. The cal cup is a 2nd tier race for varsity women. So the UCF women are right to put there fastest women in there fastest boats. I would compare lightweight womens rowing to southern cali. high school rowing a couple of fast teams but making the finals is really no big deal.

JW Burk said...

Well, although I might argue with you about the "15 sec behind the last place crew at NCAA's," this post made no claim about how fast lightweights are compared to heavyweights. We all accept as fact that there is a speed difference between heavyweights and lightweights, otherwise why would there be two different categories? I'm not sure, though, how it follows from this fact that a program should put its fastest women in the heavy boats. If, for example, you could have a national championship lightweight eight, but a 20th place heavyweight eight, why do you choose 20th place?

The point of my review of the time spread at the 2006 IRAs and NCAAs was to show that they were just about the same. Therefore, if it's no big deal to make the finals at IRAs, it also must be no big deal to make the finals at NCAAs.

Mahalo said...

I am sorry I completely disagree with your argument that the IRA championship for women’s lightweight rowing is as competitive as the NCAA DIV 1. I will give you that the IRA is as competitive as NCAA DIV 2. Your standard is time difference between 1 – 6 and first to last place. You said that your theory was challenged in 04 and 05 but in 06 it held up. Maybe it’s an anomaly or maybe it’s because the dam was open at IRA’s. I really think if your theory is to hold up you need more than only 3 samples. Also, there are no barriers of entry for the lightweight women. The NCAA you have to qualify for. You have to be judged by your peers. All you need for the IRA is a school, boat and some athletes. If a field was more competitive you would also see a greater distribution of schools. Let’s look at the top three finishers in the NCAA and IRA.
NCAA in 03’ Rad, Mich, Stan
04’ Brn, Yale, Mich
05’ Cal, Prn, Rad
06’ Pri, Cal, Brn

03’ Pri, Rad, Wisco
04’ Wisco, Pri, Rad
05’ Wisco, Pri, Rad
06’ Wisco, Geo, Rad

If you really want to see a greater distribution just look at your 4-6 NCAA WSU, OSU, VIR, MSU, USC, and Wash

At the IRA you have your Stan, UCF, Purdue, URI and Bucknell. The distribution is close but you really don’t know year end and out if they are even going to fill out a team. I would have to say that making it to the finals at NCAA is much more difficult than making the finals at IRA.

JW Burk said...

You make several points, some with which I agree and some with which I disagree. I completely agree that the medal distribution should be better among lightweights, it's a point that's been discussed often on FITD, and really underlies most of what I right about the growth of the category. However, if we turn to parity, there is another way to look at it. In the last ten years, four different schools have won the men’s heavyweight championship, five different schools have won the women’s heavyweight championship, and four different schools have won the lightweight women’s national championship. (That fifth school for the heavy women just won last year.) Yes, there is a lot more froth in the upper levels of the heavies (a good thing), but only a few schools have actually won championships.

Now of course, I like 2006 because it supports my theory. I would agree that the sample size is way to small to draw any conclusions. But, here is my counter argument: When a child learns to walk across the room, you don't say, "Yes, but last year she had to crawl." Women's lightweight rowing is in a growth stage. We're not at a stable point where we can look back over a period of time and find consistent results. Last year was a big year for lightweights for all the reasons I mentioned.

With the barriers to entry argument, I think you confuse quantity with quality. The very point of my (too long) post was to explain why that is a mistake. The lightweight version of the NCAA qualification process you note is handled by self selection. There need be no qualification because slow boats drop out and mix in with heavyweights. Again, those heavy weight boats that don't make the NCAAs - they're mere fodder for the ultimate champion. I would say it's a safe bet that Wisconsin was more worried about Bucknell in the IRA final than Princeton was about Michigan State in the NCAA final. Neither boat ever gave a thought to the boats in the petites.

If you want to talk about competitive, how do you get more competitive than having your first and second place boats come out of the reps? They were each beaten by two other crews in the course of the same regatta! At NCAAs the top boats cruised the whole way. Then we come back to the time distribution being the same. The notion that in 2006 the NCAA championship was more competitive than the IRA lightweight championship cannot be supported by the facts. In fact, now that I think about it (the reps) it seems clear that IRAs was more competitive!

Look, we're both talking about a theoretical concept without an objective definition (competitiveness). Women's lightweight rowing has a long way to go to be as robust as the heavies (again, this idea underpins nearly everything I write here), but in 2006, the stars aligned, and we had a championship that was as exciting and competitive as the heavies. And my real point is that simply because there are only 12 or 15 lightweight eights left standing at the end of the season doesn't mean there weren't a lot more that tried to make it that far.

Anonymous said...

If we discount the finals at IRAs which were highly suspect due to the open dam...

Look at the quality of the racing to make the final for the following categories included are the advancing boats plus the one who just missed it

LWT Women IRA (Heat 2 adv)
1 Princeton 6:39.79
2 Cent. Fla. 6:40.83
3 Wisconsin 6:40.88
1 Radcliffe 6:41.60 (dragging a stick as mentioned in a previous post)
2 Bucknell 6:41.9
3 Georgetown 6:46.35

LWT Men IRA (Heat 3 adv)
1 Cornell 5:53.53
2 Columbia 5:55.25
3 Harvard 5:58.20
4 Penn 6:01.23
1 Navy 5:53.25
2 Princeton 5:53.25
3 Yale 5:56.57
4 Dartmouth 5:56.60

HVY Men (IRA Semi to Grand 3 adv)
1 Yale 5:45.88
2 California 5:46.4
3 Harvard 5:49.52
4 Northeastern 5:49.84
1 Princeton 5:43.1
2 Washington 5:44.18
3 Brown 5:45.99
4 Wisconsin 5:47.14

HVY Women (NCAA semi to grand 3 adv)
1. Princeton 06:20.85
2. Brown 06:26.35
3. Washington State 06:27.41
4. Minnesota 06:28.81
1. California 06:30.47
2. Michigan State 06:32.13
3. Ohio State 06:32.57
4. Washington 06:33.99

According to my math...
If you want to make the finals.
You must be within X% of the fastest qualifier:

Hvy women - 1.7% if you are in the fast heat... 3.1% regardless of heat placement

Hvy men - 0.8% if you are in the fast heat... 1.9% regardless of heat placement

Lt Men - 0.9% if you are in the fast heat... 1.4% regardless of heat placement

Lt Women - 0.3% if you are in the fast heat... 0.5% regardless of heat placement

Which event category was most competitive last year?