Competitive Intelligence

Our least conventional entry at Nationals is the Men's Masters B 2- mostly because Craig has not really rowed sweep oared shells in about 10 years and I have done it twice since November of 2019. Neither one of us is comfortable with toe steering, and we are both predominantly starboard rowers. Considerations around jumping on sweep line ups for scullers that have a solid collegiate sweep basis but don't really sweep any more will be the topic of another post, not now we are using the pair as a conduit for competitive intelligence.

David Garcia, one of our competitors in the 2- from Texas Rowing Center, has a public Instagram profile with a number of screen shots from ergometric tests at various distances that are free for us to evaluate. Importantly, the development of "Paul's formula", allows us to homogenize distances and obtain a 2km test equivalent mean 500m split for any testing distance. The formula is based on the premise that the 500m split of a given athlete worsens by 5s for every doubling of the distance. This of course results in a log chart with a slope of 5/ln[2] and an intercept that varies depending on the physiology of the athlete. For example, an athlete that can complete 2'000m in 6:00 (a 1:30 = 90s, 500m split) will show the following intercept:

b = 90s - (5/ln[2]) x ln(2000m) = 35.2 s.

The theoretical interpretation of this intercept is that the peak power of that athlete would be 35.2s / 500m or 8'000W. This is about 11HP which has never been recorded on a rowing machine, but we know that the equation collapses at very short distances and that it is not possible to truly measure peak power at a sufficiently small discrete stroke interval.

Either way, the logarithmic equation is now sufficient to predict maximum 500m splits at any testing distance. For the hypothetical 6min, 2km rower it would look like this:

Distance	Predicted 500m split
250m	1:15.0
500m	1:20.0
1000m	1:25.0
1500m	1:27.9
2000m	1:30.0
5000m	1:36.6
6000m	1:37.9
10000m	1:41.6
42195m	1:52.0

For our competitor David, the data looks as follows:

Date	Distance, m	Mean 500m split	Extrapolated 500m split at equivalent 2km test
9/9/2015	500	1:21.8	1:31.8
1/10/2016	5000	1:44.0	1:37.4
2/16/2019	1000	1:30.4	1:35.4
1/21/2020	5000	1:42.9	1:36.3
1/26/2020	100	1:15.5	1:37.1
5/18/2020	100	1:10.5	1:32.1
8/6/2020	500	1:21.3	1:31.3
9/14/2020	21097	1:56.3	1:39.3
11/5/2020	1000	1:30.0	1:35.0
1/1/2022	500	1:19.7	1:29.7

The extrapolated split data shows a very heavy tilt towards shorter distances ,though seemingly optimized for the 500m and nothing shorter. The subject is faster than me for distances of 1'000m and less, and slower for 5'000m and more. It is not possible to know where our curves cross but I suspect it is right under the 2'000m mark. Our problem of course is that the pair event is a direct 1'000m final. Public Instagram profile does not include mass information but inspection suggests a height of around 1.85cm and mass in the 95-100kg range, similar to our boat average.

Corollaries:

1. Since we find our selves in a slight physiological disadvantage at the tested distance, technical superiority becomes sine qua non.

2. The race strategy suggests using technical advantage to maintain overlap in the first 250m, establish a punishing pace in the middle 500m to challenge the endurance deficiency shown in the data, and maintain advantage against a potential brutal assault in the last 250m by all means necessary.

I will post the video after the event to see how the analysis held up.