Predictive value of Ergometer Performance based on 1993 Data
News
Elite Ergometer Performance Analysis
Gender, Bodyweight, and On-water Performance Comparisons
Introduction
In this article I will try to provide data that answers four distinct questions:
- What are the Ergometer Performances of Elite Rowers?
- How do Heavyweight and Lightweight Performers Compare?
- How Do Male and Female Performers Compare?
- Can Ergometer Performances be used to accurately predict On-Water Performance?
The ergometer data I have used for this analysis comes from actual U.S. National Team test results submitted by team candidates in Fall, 1994. Data is broken down by gender and weight class. From these listings, average data for the top 10 performers in each group were used for comparisons. These data can be considered representative of world class performances based on the multiple medals earned by the U.S teams in 1993, 1994 and 1995 World Rowing Championship events. Three ergometer performance times are used by the U.S. National team coaches: the 500 meter, the 2000 meter, and the 6000 meter time trial. All are performed on Concept II ergometers.
TABLE 1: Physical Characteristics of the Rowers
| AGE | HT (in) | HT (m) | WT (lbs) | WT (kg) | ||
| Heavy Men | 24.9 | 76.4 | 1.94 | 209.1 | 95 | |
| Light Men | 26.6 | 72.3 | 1.84 | 162 | 73.64 | |
| Heavy Women | 26.4 | 70.4 | 1.8 | 164.7 | 74.9 | |
| Light Women | 29 | 66.8 | 1.7 | 130.7 | 59.41 |
It is worth pointing out that the light weight men and the heavy weight women are almost exactly the same weight. This makes a gender-only comparison most reasonable for these two groups.
TABLE 2: Performance Times from U.S National Team Testing, Fall, 1994
| 500M | pace/500 | 2000M | pace | 6000M | pace | ||
| HW Men | 1:20 | 1:20 | 5:58.3 | 1:29.5 | 19:05 | 1:35 | |
| LW Men | 1:26.8 | 1:26.8 | 6:19.9 | 1:35 | 20:25 | 1:42 | |
| HW Women | 1:34 | 1:34 | 6:52.5 | 1:43 | 21:51 | 1:49 | |
| LW Women | 1:41.9 | 1:41.9 | 7:21.6 | 1:50 | 23.44 | 1:58.7 |
From the data above, I have generated three relative perfomance matrices based on time percentage.
TABLE 3: 500 meter trial: Relative Performance Matrix (percent)
| HW MEN | LW MEN | HW WOMEN | LW WOMEN | ||
| HW MEN | 100 | ||||
| LW MEN | 92.16 | 100 | |||
| HW WOMEN | 85.1 | 92.3 | 100 | ||
| LW WOMEN | 78.5 | 85.2 | 92.2 | 100 |
TABLE 4: 2000 meter trial: Relative Performance Matrix (percent)
| HW MEN | LW MEN | HW WOMEN | LW WOMEN | ||
| HW MEN | 100 | ||||
| LW MEN | 94.31 | 100 | |||
| HW WOMEN | 86.9 | 92.1 | 100 | ||
| LW WOMEN | 81.0 | 86 | 93.4 | 100 |
TABLE 5: 6000 meter trial: Relative Performance Matrix (percent)
| HW MEN | LW MEN | HW WOMEN | LW WOMEN | ||
| HW MEN | 100 | ||||
| LW MEN | 93.5 | 100 | |||
| HW WOMEN | 87.4 | 93.5 | 100 | ||
| LW WOMEN | 80.4 | 84.8 | 92. | 100 |
In the next serise of tables I have taken into account the non-linear relationship between velocity and power that is in effect on the ergometer (and on the water). I have converted performance times to average power output in watts maintained over the duration of the test. The conversion was made using a specific formula supplied by Concept II Inc. for their ergometer. This formula closely approximates the expected cubic relationship between power and velocity, based on known drag force-velocity relationships.
TABLE 6: Performance Power (watts) derived from Performance Times in Table 2
| 500m | 2000m | 6000m | ||
| HW MEN | 694.8 | 494.3 | 409 | |
| LW MEN | 543.9 | 415.4 | 335.3 | |
| HW WOMEN | 432.3 | 324.1 | 272.8 | |
| LW WOMEN | 338.3 | 264.1 | 212.7 |
Next, I have created three more relative performance matrices, this time based on absolute power differences.
TABLE 7: 500 Meter trial: Relative Power (watts) Differences (percent)
| HW MEN | LW MEN | HW WOMEN | LW WOMEN | ||
| HW MEN | 100 | ||||
| LW MEN | 78.2 | 100 | |||
| HW WOMEN | 62.2 | 79.5 | 100 | ||
| LW WOMEN | 48.7 | 62.2 | 78.3 | 100 |
TABLE 8: 2000 Meter trial: Relative Power (watts) Differences (percent)
| HW MEN | LW MEN | HW WOMEN | LW WOMEN | ||
| HW MEN | 100 | ||||
| LW MEN | 84 | 100 | |||
| HW WOMEN | 64.9 | 78.02 | 100 | ||
| LW WOMEN | 53.4 | 63.6 | 81.5 | 100 |
TABLE 9: 6000 Meter trial: Relative Power (watts) Differences (percent)
| HW MEN | LW MEN | HW WOMEN | LW WOMEN | ||
| HW MEN | 100 | ||||
| LW MEN | 81.9 | 100 | |||
| HW WOMEN | 66.7 | 81.4 | 100 | ||
| LW WOMEN | 52 | 63.4 | 78 | 100 |
By now you can see that the differences in power output among the groups are much greater than the actual difference in performance time. This is due to the exponential relationship between ergometer velocity and the power necessary to achieve it! In order to account for the influence of bodyweight on performance differences, I have next generated weight factored performance tables. The specific weight factor used is an important consideration. Obviously, the simplest way would be just to divide power by bodyweight. For anaerobic capacity, this approach is appropriate, based on the assumption that there is a linear relationship between absolute muscle mass and absolute anaerobic capacity. So, the 500 meter test was scaled in this way. However, the longer trials are presumed to be dependent primarily (2000) to exclusively (6000) on aerobic capacity. Maximal oxygen consumption does not vary linearly with bodyweight. Heart cross sectional diameter will scale with the square of height. However, body mass will scale with approximately the cube of height. This gives a predicted relationship between VO2 and bodyweight of V02 proportional to Mass ^2/3. So, I have chosen to use this “allometric scaling” method for the 2000 and 6000 meter relative power matrices.
TABLE 10: Performance Power Scaled for Bodyweight
| 500m/kg | 2000m/kg^2/3 | 6000m/kg^2/3 | ||
| HW MEN | 7.32 | 23.71 | 19.62 | |
| LW MEN | 7.4 | 23.61 | 19.06 | |
| HW WOMEN | 5.77 | 18.22 | 17.32 | |
| LW WOMEN | 5.69 | 17.32 | 13.95 |
Now I will again save you some calculator time by converting the table above to three relative performance matrices.
TABLE 11: 500 meter: Relative Power Differences (watts/kg)
| HW MEN | LW MEN | HW WOMEN | LW WOMEN | ||
| HW MEN | 100 | ||||
| LW MEN | 101 | 100 | |||
| HW WOMEN | 78.8 | 77 | 100 | ||
| LW WOMEN | 77.7 | 76.9 | 98.6 | 100 |
TABLE 12: 2000m Relative Power Differences (watts/kg^2/3)
| HW MEN | LW MEN | HW WOMEN | LW WOMEN | ||
| HW MEN | 100 | ||||
| LW MEN | 99.6 | 100 | |||
| HW WOMEN | 76.8 | 77.2 | 100 | ||
| LW WOMEN | 73 | 73.36 | 95 | 100 |
TABLE 13: 6000 meter: Relative Power Differences (watts/kg^2/3)
| HW MEN | LW MEN | HW WOMEN | LW WOMEN | ||
| HW MEN | 100 | ||||
| LW MEN | 97.1 | 100 | |||
| HW WOMEN | 78.13 | 80.4 | 100 | ||
| LW WOMEN | 71.1 | 73.2 | 91 | 100 |
At this point some observations are in order. At face value, it seems that the scaling choices made are appropriate. For the 500 meter anaerobic capacity test, power/kg is almost identical between light and heavy men, and light and heavy women. However, lightweight men and heavyweight women are almost identical in bodyweight, yet 20+ percent different in scaled power output. Why? The bodyweight I have used is absolute bodyweight, which does not account for differences in body composition, which are significant between these two groups. Lightweight men have more muscle mass at the same bodyweight. Whether this completely accounts for the difference would require actual body composition data for the subjects, which I do not have. However, elite male lightweights are at about 5% bodyfat, while elite heavyweight females are at about 15% bodyfat. This 10% differences translates to 7.4 kg in this group. 7.4 kg greater muscle mass in the lightweight men would translate to about 20% of total muscle mass. This would account for the difference in relative anaerobic capacity observed in these similary trained groups.Second, scaling the longer tests based on allometric assumptions also seems appropriate. Performance power differences between the two male groups are only 1 to 3%. The difference between the women is 5-10%. This larger difference is consistent with the relatively lower performance level in the U.S lightweight women on the water, but is still reasonable.
Finally, after allometric scaling, a 20+% difference in 2k and 6k power persists between men and women . Can this difference be explained by differences in aerobic capacity between these two groups? I have VO2 max data for the 1992 US Olympic team (From Dr. Frd Hagerman, Ohio University). If VO2 max is scaled to the 2/3 power, the men had a value of 320 ml/kg^2/3, compared to 245 for the women. This is a difference of 23.5%. Thus, it appears that Maximal oxygen consumption differences can adaquately explain scaled power differences.
Finally, I can take the data above and attempt to predict on-water performance differences among the groups. I will use actual differences in 2000 meter ergometer performance power scaled allometrically for bodyweight differences. Then, I will predict the ON-WATER boat velocity differences for the given power differences based again on the power-velocity relationship described above. This cubic relationship holds for air and water drag. Only the drag coefficient changes.
TABLE 14: Predicted On-Water Performance Differences (% of Heavyweight Males) for 2000m
| HW MEN | LW MEN | HW WOMEN | LW WOMEN | ||
| HW MEN | 100 | ||||
| LW MEN | 99.9 | 100 | |||
| HW WOMEN | 91.6 | 91.7 | 100 | ||
| LW WOMEN | 90.1 | 90.2 | 98.4 | 100 |
My physiology-based predictions, coupled with known physics relationships, predicted on-water performance differences that range from 0 to 10% among the groups. Is this accurate? Well, to answer that question, I have found actual on-water results for the U.S. from the 1993 World Championships. I didn’t use 1994 results due to the varying wind conditions during the regatta. Using 1993 is reasonable, since the turnover of athletes during the time span has been small. I calculated on-water velocity by averaging the performance velocities of the 8+, 4 X, 4 -, and 2X boats. These boats contained many of the specific athletes represented in the erg data. For all comparisons involving light weight women, I had to make the cross-comparisons by using only the 2X and 4- for each group, since I did not have actual times for their bigger boats (They did not make the finals).
TABLE 15: Actual On-Water Performance as a percentage of Heavyweight men, 1993 World Championships
U.S. Boat Results
| HW MEN | LW MEN | HW WOMEN | LW WOMEN | ||
| HW MEN | 100 | ||||
| LW MEN | 99.27 | 100 | |||
| HW WOMEN | 90.48 | 91.2 | 100 | ||
| LW WOMEN | 91.02 | 90.7 | 97.1 | 100 |
Wow, predicted and actual performance differences are pretty darned close! It seems that in similarly trained groups, we can make very accurate predictions of on-water performance based on dry-land ergometer performances, using 2000 meter or 6000 meter erg tests and the scaling methods chosen. And, it seems that anaerobic capacity (500 meter performance) would bear little additional value from a predictive standpoint. Rowing people are already aware that among the best teams, the differences between lightweights and heavyweights are very small, despite big dryland erg performance differences. This is explained by the increased boat wetted surface area and drag that comes with each kilogram of mass inside the boat. Combing available power data with the 2/3 power scaling method seems to give accurate estimations of the effect of a mass and power change (new lineup) on boat velocity, if technique is held constant. In very short races (less than 1000 meters), we would expect a slight advantage to persist for the heavier rower, due to an increase in anaerobic capacity that is slightly greater relative to the corresponding increase in boat drag.It has been suggested, that if the mass of boats rowed by lightweights was in the same proportion to their bodyweight as that for heavyweight crews, performance time differences would disappear. These data support that suggestion.