Courtesy of SwimSwam contributor Chris O’Linger, an assistant coach at the University of the Incarnate Word.
Anyone who swims, or has swum, in my program knows that I choose to focus on research-driven methods pertaining to my training program and performance tactics. One of the most intriguing topics I have run across in my research has been freestyle stroke pulling indices (tempo, stroke length, cycle count), and its ability to predict the average pace and closing speed of eilte 800 and 1500 meter swimmers. Unfortunately, I have not been able to find such a consistent formula for short course swimming just yet, as there will be a much higher concern for control of VO2 max levels, and hypoxic exhaustion when flipturns and underwaters play at least twice the role in the events.
For the purpose of this study, I am using a formula to determine stroke cycles for each swimmer encompassing the amount of cycles taken with a few control factors concerning tempo and distance from the wall during the final stroke extension. At an elite level, these variables can be controlled at a much better rate due to the sheer fact that these swimmers have rehearsed these indices for numerous years, and, for the most part, have settled into a constant rhythm at this point in their career. I believe that stroke cycle, with these minor adjustments, can account for all indices to a degree of confidence high enough (a = <.02).
I analyzed the top 16 male and female miles swam at the 2013 FINA World Championship meet in Barcelona, and came up with an descriptive statistic model which would allow me to find correlational trends and regressional predictive factors for a swimmer’s closing speed based off of their projected stroke cycle counts. Below are the average statistics for each variable to get a feel for what each individual’s data would comparable with:
| CS | AP | D | SC |
| N = 32 | N = 32 | N = 32 | N = 32 |
| M = 62.2819 | M = 63.0316 | M = 18.5122 | M = 1.2491 |
| SD = 2.433 | SD = 3.450 | SD = 2.399 | SD = 1.331 |
| Min = 58.68 | Min = 54.48 | Min = 13.79 | Min = -1.16 |
| Max = 66.34 | Max = 66.83 | Max = 21.83 | Max = 4.20 |
*AP = Average 100m pace
CS = Closing 100m split
D = Difference between average 100m pace and closing 100m split
SC = Projected stroke cycle count
After finding many significant correlations and coefficients high enough to run regressional analyses to determine prediction rates, I found that swimmers with a lower cycle count throughout distance races results in a faster closing speed than swimmers with higher cycle counts. The effect sizes and R-squared values determined that this was a strong enough correlation to predict an elite swimmers closing speed based purely off of their stroke indices.
The regressional line-fit analysis is shown below:
Simply put, at an elite level where training and technique have been practiced with great emphasis, we are noticing that the ultimately technical swimmers are the ones who are still prevailing. Remember, this data is for the top swimmers in the world, and not anyone can take 15 cycles per lap and swim at this speed, but the trends do show that drill work and awareness of the stroke can lead to vast improvements across all events.
Chris O’ Linger, assistant coach, Incarnate Word swimming & diving. (Image courtesy of UIW)
Swimming fast and trying hard is not enough. Technical integration needs to be established in a foundational manner before the workload is unmanageable at proper form.
O’Linger is an assistant coach for the Incarnate Word swimming and diving program. He swam collegiately at both the University of Florida and University of Tampa. He earned a degree in social psychology from Tampa. He is studying kinesiology.
The relation between finishing speed and stroke count while clear in the data presented, and theoretically reasonable doesn’t seem to get at the heart of the matter.
Namely, if your goal is to have your fastest overall time, isn’t the correlation that matters one with the Y variable equal to overall time or average pace/speed?
Catherine, nothing is ‘pretty clear’ unless we define our terms. Does ‘decrease’ mean decreasing the number of cycles, which is the equivalent of increasing stroke length (SL or DPS), or does it mean decreasing the frequency (SR or ‘minute rate’)? If its the first and you want ‘improvement’, then the SR has to remain constant. If it slows down due to the increase in SL then the best you can hope for is the same time. If it’s the second, then the SL has to increase in order to stay at the same speed.
Either way, it is spurious to consider SR or SL in isolation. They are two sides of the same coin and one CANNOT exist without the… Read more »
It seems pretty clear that *decreasing* stroke cycles counts as an improvement, but Clive raises a good point. How is that actually done? Grow bigger hands? Seriously, I’d like some advice. When I try to decrease my stroke count I end up doing a catchup type of stroke. Is that what I’m supposed to be doing?
Oh dear, we do like to complicate things in swimming, don’t we? I’m struggling with the title of the piece:how to improve stroke cycles. How do you ‘improve’ stroke cycles? Do you make them more cyclical? More ‘stroky’? Or does the author mean increase cycle frequency? And how do you predict closing speed when they’re ‘improved’? The article doesn’t tell us.
Averaging 100m splits to one, ten-thousandth of a second seems like an over-abundance of pseudo-accuracy to me. Which brings me to the most obscurating factor: averages. The graph shows a regression line drawn through a population of n=32. 30 of the data points (30 out of 32 swimmers) do not fall on the line. They do not conform to… Read more »
Catherine,
Yea, those should certainly be points looked into in the future. The data set up there is merely a sample to see how the info was gathered. There are more complex variables and tests run, but theism age displayed is the overall category, mainly used to further validate the claims investigated per sex. The results were significant for both genders, but concerning regressional fit analysis shows that the closing difference and stroke cycle count are much more predictive for women. A number of things can be drawn from the results to investigate further, including the variance of other factors (tempo, underwaters, positioning of vertical pull, etc.), and I agree that stroke count to total time would be another… Read more »
Catherine, I forgot to answer your earlier question concerning Yang. Yes, he certainly posted the largest difference between average pace and closing split. He also hones the lowest cycle count by a decent margin.
Leo, I’m sorry your not a fan of my model or the trends shown. I’ve said it on several occasions, skepticism is the driving motivation of research! Maybe another work in the future will spark your interest!
And thank you, Braden.
Chris, thanks for the replies. However, Leo’s comments suggest yet another question. Are the points with SC > 18 mainly women, and SC < 18 mainly men? If that's the case maybe you could just plot SC vs. total time.
Russell Mark from USA swimming does a lot of work like this. His tweets with data are interesting and the USA swimming stats twitter is also very helpful.
This is a crappy model, for the cluster of 14 of the 18 males with sc of 16-17 closing speed is all over the map. For 10+ of the women in the 20-21 sc range closing speed is also not predicted by the model.