For many triathletes, the swimming portion of a triathlon race is often the most challenging and scary of the three disciplines. Many athletes seek the advice of coaches with a background in swimming to learn the basics and improve performance. Many of these coaches place emphasis on the metric of distance-per-stroke as a means of evaluating swimming efficiency. A minority of coaches and swimmers have alternative metrics. As an athlete, how do you sort out what is right for you?
As a former Division-1 swimmer in College, I can tell you what my metrics for efficiency are. But why would you care how I feel about metrics? You will just as easily find other coaches with the same passion for different metrics, namely distance-per-stroke. The best way to settle the issue is to do an experiment to see if distance-per-stroke really makes a difference in swimming efficiency. The results of this experiment are contained in this article. As most of you know, an experiment is only as good as its definitions and assumptions. I have listed each below.
Assumptions in the Experiment to Relate Distance-per-Stroke to Swimming Efficiency
- Efficiency in this experiment is defined as Heart Rate per specific time and specific distance. The specific distance used in this experiment is 50 meters, taken in a 50 meter pool. The specific target times tested include 30, 33, 36 and 39 seconds over the 50 meter distance. Heart rate was selected because it is economical to measure (does not require expensive equipment) AND is a reasonable indicator of exertion.
- Freestyle was the swimming stroke utilized throughout this experiment.
- The distance-per-stroke technique (hereafter called the ‘downhill’ technique) is defined as follows. Angle of arm entry was approximately zero. An attempt was made to stretch the arm forward immediately after entry and before the pull was initiated. The kicking method employed was four-beat, each beat defined as an up or down motion of the foot over a complete stroke cycle of the left and right arm. The finish of the stroke was below the suit line. The pulling motion created propulsion throughout the stroke. Although the specific number of strokes per length was not controlled, an attempt was made to minimize the total number of strokes while meeting the target times
- The alternative technique (hereafter called the ‘float’ technique) is defined as follows. Angle of arm entry was approximately 20 degrees from horizontal. No arm stretch was done after entry of the hand. The kicking method employed was six-beat, each beat also defined as an up or down motion of the foot over a complete stroke cycle of the left and right arm. The finish of the stroke was also below the suit line. The pulling motion also created propulsion throughout the stroke. For the float technique, no attempt was made to regulate stroke count while meeting the target times.
- A single athlete (Duane Dobko) was used as the test subject for the study. The assumption made in this test is that body type and swimming ability are independent of evaluating swimming efficiency between different techniques. Since different swimmers were not evaluated, this assumption could not be confirmed. The benefit to using an experienced swimmer was assurance that the swimming motion would create movement throughout the stroke regardless of technique, resulting in a fair comparison of the two swimming methods. Changing techniques while still generating optimal movement is more difficult for inexperienced swimmers.
- The two techniques were evaluated for all four target times for three repetitions apiece. Twenty-four test runs were completed in order to meet this requirement. The test runs were all randomized in order to prevent biasing of results due to fatigue or test conditions. Each repeat was swum in the same lane and direction of the pool in order to minimize any effects of wind or current.
- The test was completed in a single session in order to eliminate effects of athletic conditioning over time.
- Immediately after each repeat, heart rate was measured by counting beats at the neck over a 10 second period. The heart rate and time was recorded if the total time (measured by stopwatch to 0.01 second resolution) was within plus/minus 0.5 seconds of the target. If the time requirement was not met, then the test run was repeated until the time requirement was met. Failed attempts were not recorded. For successful attempts, the stroke count was recorded. See attachment A for raw data.
- Between repeats, adequate time was allowed for complete rest and recovery. Such time was at the discretion of the athlete.
- An in-the-water ‘backstroke’ push-off start was employed for each start. Four dolphin kicks were taken off of each start, after which the athlete surfaced and began swimming. The total stroke count includes the stroke taken at the finish. As the test distance was 50 meters in a 50 meter pool, there were no flip turns that could have biased results.
- The athlete wore the same equipment (brief style swimsuit) throughout the test. Wetsuits or full-body suits were not worn at any time.
Results: Heart Rates versus time for the two swim techniques are shown in Figure 1.
Click on the images to view larger images
On examination of Figure 1, it is apparent that the Heart Rate vs. Swim Time looks the same regardless of technique! A statistical analysis was performed to verify this result numerically.
For this result, a differential between each trial and target time was made. An example: the difference between trial-1 Float-30 second target and trial-1 Downhill-30 second target. All of these differentials were calculated for each trial, and the average differential between these was analyzed by paired comparison. The average differential, 90% confidence limits and p-value of the differential is shown in Figure 02.
Figure 02 – results
The results and conclusions that can be made from Figure 02 are described below
Time Differential: As the experiment evaluated Heart Rate at a set goal time, the Float-Downhill must be as close to zero as possible in order for the experiment to be valid. Figure 02 shows that Float vs. Downhill times were performed to a precision better than plus/minus 0.2 seconds. A significant difference in time could not be detected between the two methods. The conclusion is that this experiment successfully tested time targets that are equivalent between the Float and Downhill swim techniques
Stroke Count Differential: As the differences between Float and Downhill techniques are quantified in this experiment by stroke count, it is imperative that a statistically significant difference of stroke count is proven and quantified. The p-value of zero proves that a difference was detected in stroke count of the two methods. To 90% confidence, the Float technique involved anywhere between 6.0 and 7.2 strokes more than the Downhill technique, with an average of 6.6 strokes over a 50 meter distance. This is a very consistent result and enough to show that this experiment is a fair comparison of the two swimming methods.
Heart Rate Differential: Heart Rate was the response variable in the experiment. As the p-value was greater than 0.05, the conclusion of this experiment was that a difference between Float and Downhill techniques could not be established. The confidence interval of this result is 10.8 beats per minute. It can be concluded that any difference between Downhill and Float methods is less than 6.4 beats per minute.
Analysis of Results – Differences in Stroke Count
The plot of stroke count vs. recorded time is shown in Figure 03. It shows clearly the two populations of stroke counts (high and low) for each target time. It also shows that as target time decreased, so did stroke count for both the Float and Downhill techniques. Furthermore, the differential between the two techniques appears constant for a given target time.
Figure 03 – Stroke count (over 50 meters) vs. 50 meter swim time (seconds)
The data gathered from the Float and Downhill techniques were analyzed. This was done to ensure that there were no odd data points or “fliers” which impacted the experimental result. The Float technique is analyzed in Figure 04 and the Downhill Technique is analyzed in Figure 05. From these figures, it is conclusive that there are no ‘flier’ data points which impacted results. The charts shown in Figures 4 and 5 are explained below.
Top Left – Histogram of residuals – The residuals (actual value – expected) for all data points should be roughly centered at zero with no single points unusually far from zero (compared to all other data). Both Float and Downhill data appear acceptable
Top Right – Residuals vs. order of data – This plot should appear random. If it appears like a straight line up or down or has some other pattern, it is likely that the order of test runs had an impact on the result (like athlete fatigue for example). There are no patterns in the Float and Downhill data
Bottom Left – Normal Probability Plot – This plot should be approximately a straight line. If it is not, then the data set might contain “fliers”. The line appears straight for Float and Downhill data
Bottom Right – Residuals vs. Fitted Values – If the residuals are different for a given fitted value (for example, residuals all negative for low heart rates, then all positive for high heart rates), then the results may have been impacted by unforeseen factors. The residuals are random for all fitted values for both Float and Downhill swim techniques.
Figure 04 – Float Swim Technique data analysis
Figure 05 – Downhill Swim Technique data analysis
The analysis of data shows that this experiment is a fair comparison of the Float and Downhill swim techniques. Specifically, no ‘flier’ data points were noted in the results. The difference between Float and Downhill techniques were successfully quantified as the stroke count in a 50-meter pool. In this experiment, the stroke count differential was approximately 6.5 strokes at any given speed.
This experiment result shows that the difference in Heart Rate at set speed between two swim techniques, Float and Downhill, is inconclusive within plus/minus 6.4 heart beats per minute. This result is evidence that distance-per-stroke is an inappropriate metric for evaluating swimming efficiency as defined by heart rate for a given speed. As this experiment specifically quantified distance-per-stroke, it does not identify the true metric for swimming efficiency. Either an alternative swim metric, or an alternative efficiency metric (like blood lactate levels, for example) would have to be tested in similar fashion to this experiment in order to confirm and quantify what really defines swimming efficiency.
Does this mean that coaches should let their swimmers thrash in the pool with a high stroke count? Absolutely not! Any swim coach will tell you that deliberate swimming form is far more efficient than random and unbalanced swimming form. Many coaches communicate good technique to athletes by telling them to slow their turnover down and lengthen the stroke. But it is important that both coach and athlete know that any gain in swim efficiency is NOT going to come from the distance covered per stroke. Efficiency must come from something else the athlete is doing in the process of attempting proper form. Altering an athlete’s distance per stroke (up or down) is a coaching tool to teach this alternative metric, whatever that metric might be.
Attachment A – Raw Data