Developing an anaerobic power model for determining cross-country ski racing performance.
By Bob FitzPatrick


            I have selected this particular topic due to observations I have made during my last few years of cross-country ski coaching. There is a strong relationship between upper body power and ski performance for double poling to exhaustion on a ski ergonometer. At the high school level I feel that a shorter test along these same lines may also be a good indicator of ski performance. I would like to develop a model that high school ski coaches could use for both testing power improvement as well as measuring an association with ski performance. Since a ski ergonometer is not easily accessable I will use an uphill double pole test on roller skis for 300 meter distance. I will compare these times to times for a 5 km roller ski time trial, skate technique.

The purpose will be to determine if a high correlation (r = >0.80)exists between the 300 yard anaerobic test results and the 5km skate time trial results. I will attempt to utilize a subject base of between 6 and 10 skiers. Both of these tests will be maximal efforts over the stated distances. I will collect the data from both tests and determine if a high correlation exists between both tests.


         Our sample group (N) will be made up of 6-10 subjects all with prior roller ski experience as well as current race experience. Their ages range between 17 and 55 with the mean age of 41.42years(plus/minus SD of 15.94) and mean roller ski experience of 7.71 years (plus/minus SD of 7.11 years).

All subjects are males. These athletes are highly motivated junior, senior, and master’s level racers. Testing time will take place at 9:30am in the morning for the 5km time trial and approximately 10:30am or when the subjects feel recovered for the anaerobic test. Arranging a date and place for this test took most of the work. Some of the subjects had to drive over one hour to the test site (Hopkinton State park). All subjects would be considered to be in the upper 10% fitness bracket of the general population. Most are involved with other cross training pursuits (running, swimming, rowing, cycling, and kayaking).


Asphalt road surface with 4% grade for uphill DP test (slope measured with 2 foot level manufactured by Stanley products; 25 foot tape measure manufactured by OHIO Gauge).

Asphalt road surface 5km in distance with 1/3 uphill, 1/3 downhill, and 1/3 flat terrain.

Course distance measure by Lemand, Triple crank bicycle utilizing a Flight Deck cycle Computer manufactured by Shimano (measures to hundredths of Kilometers).

Seiko Computer timer model Cal.S129 with recorder and trigger button.

2 Seiko hand held stop watches (model-30 lap split memory) for back-up.

Thermometer manufactured by Swix Corporation of Chelmsford MA

Freestyle Chronograph for time of day recordings.

V2 Jenex 850 Roller skis (provided by myself and some of the subjects-same speed)

V2 Jenex Aero Roller Skis.

Classic length roller poles with sharp ferrules-provided by subjects.

Skate length roller poles with sharp ferrules- provided by subjects.

V2 Jenex Diamond hone sharpener for roller tips.

Skate technique ski boots provided by subjects.

Helmets, roller gloves also provided by subjects.

10 orange traffic cones (8inch by Passon’s of Philadelphia, PA) marking course and start/finish areas.

1” painted white line for start mark/finish mark for DP test.

1” painted white line for start/finish mark for TT test.

Clipboard and pencils for recording.

Compaq computer Pentium III utilizing MS Excel spreadsheet for computations.

SPSS software provided by Jim Drew (Worcester Polytechnic Institute).

Text “Measurement and Evaluation in Human Performance” (6).

Brief Review of literature

            When a review of the literature on this subject and related subjects turned up more studies then I expected I was shocked! Earlier studies on the effect of strength training on ski performance focused on maximal strength issues and not power. Other studies focused on poling forces involved with cross country skiing (5) and the effects of grade on technique and rolling resistance (4). These two studies although not directly related to my theory, provided much insight into some issues I would deal with in my tests. Pole deflection is certainly an issue on DP on steep uphills and thus the necessity to use stiff poles to maintain efficient power transmission. Also roller ski speed (4)is an issue if we are to develop standardized testing sites in the future and on absolute placing of skiers in my test. Pavo Kami, the Finnish Biomechanist, found that the energy output required for uphill cross-country skiing far exceeded the highest VO2 max values. Thus, short periods of anaerobic energy play a major role in Cross country ski performance. Ultimately, it was a study conducted by Steve Gaskill (2) that provided me with the information most related to my proposed study “The development of upper Body Power in Junior Cross Country Skiers”. In another study Ken Rundell has been developing norms at the OTC Lake Placid, NY for elite junior cross country skiers performing a DP test on roller skis, on a ski ergonometer, to exhaustion (usually around 5 minutes). Also Gaskill, Serfass, and Rundell studied the Upper Body comparison’s between runners and xc skiers (3) and the need for upper body power development as part of the training program (many xc skiers are recruited from the running community).

            A call to Steve at the University of Montana rekindled an old friendship of 20 years ago and he was excited about the test I was proposing. He mailed me all of his research which was invaluable in my study but more importantly turned me onto a study that the US Olympic Committee funded, which he co-authored,

and will be released shortly after the 2002 Winter Olympic games. This study is in draft phase and is not for distribution but it focused on a power development model for cross-country skiers using roller boards (very similar to uphill roller skiing) and the need for specificity in training for the upper body power component. In this test 65 Minnesota high school skiers were tested for the effects of various training regimens on ski performance and the one significant factor was the use of roller boards for upper body power development.

            I asked myself “ How can I bring all of this down to a level that can be useful for high school or master’s racers, who do not have testing facilities at hand (ski ergonometers)?” Developing a field test skiers and coaches alike could use both to test upper body power improvement as well as establishing norms to compare to race performance (speed), was the answer.

If the two tests I have selected prove to have a high correlation then further study could be initiated that would enroll high school ski racers, as subjects to start to develop norms along these lines.


            Subjects used for this test are highly motivated junior (junior Olympic participation), senior, and master’s level cross-country ski racers. Subjects were recruited for two tests that would take place on the same day a 300 meter uphill double pole-roller ski test (DP) and a 5km skate roller ski test(TT). Subjects were instructed to work on their technique for DP skiing to prepare for the test. All subjects were given a release form to sign with instructions about the test. Background data were collected that reflected age, sex, height, weight, years of ski experience, last race experience, and amount of time roller skiing per week Two test courses were set up at Hopkinton State Park (Boston marathon). Both courses were on well maintained asphalt roadways free of motor vehicle traffic (gated). Both courses were accurately measured to 1/100 of a kilometer. Both courses were clearly marked and start/finish lines placed accordingly. The uphill DP course was measured to be 300 meters and on a mean slope of 4%. The 5km roller ski skate loop was also set within the park beginning on a parking area by the lower beach area that was flat for approximately 800meters and then gained in elevation consistently until the half way mark where upon the skiers turned around traffic cones and descended back to the lower beach area and around the parking area to the finish line. All skiers were encouraged to take as much time as necessary to warm-up . At approximately 9:30am all skiers were gathered in the start area for the 5km skate TT. All poles were sharpened and skiers were given ample water etc. Skiers used Jenex V2 roller skis, model # 850 and aero models (same speed). Poles were provided by the participants and were measured to the upper lip. Skiers were started at 30 second intervals using a random draw for start order. All skiers were instructed that this was to be a time trial effort (maximum effort). All skier’s finish times were adjusted by the interval at which they started. Each skier was assigned a score based on his/her elapsed time. All times were timed by a Seiko handheld computer timer with recorder and trigger switch. Times were all computed as total seconds and rounded to the nearest whole second. Two seiko timers with 30 lap/split memory were used as backup timers. The TT started exactly at 9;44am EST. Temperature at race start was +5 degrees C.

            Skiers were given fluid replacement in the form of water or Gatorade. All were given 30 minutes to ski easy at level 1 intensity for complete recovery. Skiers were then given practice time on the uphill DP course. When all were ready we started based on the order of start from the skate race. One skier was started and run to completion before the next skier was started. Each skier was given a countdown 30 seconds to start; 10 seconds; 5,4,3,2,1 go. This procedure was followed for the 5 km time trial as well but with 30 second start intervals. This allowed about the same rest time for all skiers. Skiers were told that this is a maximal effort (DP) and their time would be taken as their toes crossed the finish line. DP technique was the only allowable technique for the test. Times were recorded to the nearest second. Any time that was part of a second was rounded to the next highest second.

            All of the data were recorded on prepared tally sheets which were then converted over into excel spread sheet for evaluation and computation. Dr. Jim Drew (one of our subjects also provided SPSS software so I could do a regression formula for the DP race as a predictor of ski performance for a 5km TT).


All results were computed on excel spreadsheet and SPSS software but I converted over to MS Word tables to clean up the presentation a little. All times were taken from seconds and computed over to speed (seconds/km). Means and Standard Deviation scores were computed for all variables listed below. Total subjects actually tested (N) were 7. Below I used a scatter plot to see how the scores would compare to a linear relationship. With a correlation of r = 0.937382 I was surprised to see the compactness of the subjects performances on the scatter plot as well as the linear relationship of the two tests. The two of the seven subjects farther out on the plot were more a function of roller ski speed (George) and experience (Eric). Controlling these variable for further studies will be important yet the correlation still holds due to the fact that the same roller skis had to be used for both tests. I believe the statistics bear this out. The r squared value also remained in the high range reflecting the shared variance between the two variables (DP and TT). Even though not included in the raw data many of the subjects chose to do another DP test about 15 minutes after their first DP test. Most were proportionately slower (fatigue) but the relationship would still have been linear and lends more support to the reliability of the test. Unfortunately I did not have time to run this test a few times due to our canceling one session due to site construction (beyond our control). Also I would like to have had a larger sample size but getting 7 together was a chore in itself! Master’s athletes are much more difficult to work with as far as personal schedules are concerned.

Descriptive Statistics and DP and TT Test Data








































































































SD scores:











Correlation between DP test and TT test: r = 0.9374

Regression equation coefficients: TT speed = 9.081 + 0.779 x DP Speed

r sq = 0.878

Below is a scatterplot with the scores from the TT test and DP test.


I believe the results we saw support the premise that there is much more of an anaerobic power component involved with cross-country ski racing then originally thought. Most of the studies I used to base my research on point more toward the power component then studies of ten years ago. I have felt for the last few years that improvements in power development would support faster race times, especially for the skate technique. As skiers ascend even steeper hills it is the upper body involvement that allows them to ski faster whereas the function of the lower body is to keep the skis gliding.

            There are technique implications that could have positively influenced testing as well as practice effect of double poling uphill. Those who have spent more time using this training would fare better on the test. Yet the test data seemed to support the fact that practice on DP would have made for a faster TT as demonstrated by the correlation coefficient. Aaron is a good example of this. He had the best technique (subjective) and he has had more uphill training then most of the others. Kurt probably has a better engine than Aaron but technique might have held him back from skiing more efficiently. Roller ski speed was not tested due to time constraints and I went on a subjective evaluation and manufacturer’s knowledge (Aero’s and 850’s being around the same speed). For future studies I would use the same skis for everyone for both tests.

            There is no question as to the part aerobic base work has in developing good cross-country ski racer’s but the time is upon us to beef up our programs via power development training for the upper body through uphill roller skiing in shorter duration 1-3 minute anaerobic bursts and repeats. By measuring improvements in this area we should be making time improvements in our racer’s 5km TT .

            I have received 15 emails or so awaiting the posting of my test results. The NENSA (New England Nordic Ski Association) Webmaster at UNH has asked for an article about my research and it’s implications to HS racing and its part relative to what Gaskill and Rundell have been doing. I will post these on the NENSA website and the NENSA newsletter. I hope to continue this testing with larger groups and test more of the variables involved. Ultimately I would like to give the High School coaches a model they could use for quick and easy application relative to power improvement relative to the upper body.

            This has been a very exciting project for me both personally and professionally! I have always wanted to get into the testing area and this has helped me get my feet wet. I also want to thank Jim Drew from WPI for helping me to get reacquainted with test and measurements and take things a step further with the utilization of the SPSS software.

I never realized how much work was involved with setting up a small study like this. Logistics can be a nightmare.


1.     Bilodeau, B., B. Roy, and M.R. Boulay. Upper body testing of cross-country skiers. Med. Sci. Sports Exerc. 27:1557-1567, 1995.

2.     Gaskill, S.E., R. C. Serfass, D.W. Bacharach, and J.M. Kelly. Response to training in cross-country skiers. Med. Sci. Sports Exerc. 31: 1211-17, 1999.

3.     Gaskill, S.E., R.C. Serfass, and K.W. Rundell. Upper Body Power Comparison Between Groups of Cross-Country Skiers and Runners. Int. J. Sports Med. 20:290-94, 1999.

4.     Hoffman,M.D., P.S. Clifford, A.C. Snyder. Physiological effects of technique and rolling resistance in uphill roller skiing. Med. Sci. Sports Exerc. 30:311-17, 1998.

5.     Millet, G.Y., M. D. Hoffman, and R. B. Candau, Poling forces during roller skiing: effects of grade. Med. Sci. Sports Exerc. 30:1637-44, 1998.

6.     Morrow, J.R., A.W. Jackson, J.G. Disch, and D.P. Mood. Measurement and Evaluation in Human Performance. Champaign, Il: Human Kinetics, 1995.