Thirty-Eight Years of Training Distribution in Olympic Speed Skaters

Abstract

Unlabelled:
During the last decade discussion about training-intensity distribution has been an important issue in sports science. Training-intensity distribution has not been adequately investigated in speed skating, a unique activity requiring both high power and high endurance.

Purpose:
To quantify the training-intensity distribution and training hours of successful Olympic speed skaters over 10 Olympiads.

Methods:
Olympic-medal-winning trainers/coaches and speed skaters were interviewed and their training programs were analyzed. Each program was qualified and quantified: workout type (specific and nonspecific) and training zones (zone 1 2 mMol/L lactate, zone 2 2-4 mMol/L lactate, zone 3 lactate >4 mMol/L). Net training times were calculated.

Results:
The relation between total training hours and time (successive Olympiads) was not progressive (r = .51, P > .5). A strong positive linear relation (r = .96, P < .01) was found between training distribution in zone 1 and time. Zones 2 and 3 both showed a strong negative linear relation to time (r = -.94, P < .01; r = -.97, P < .01). No significant relation was found between speed skating hours and time (r = -.11, P > .05). This was also the case for inline skating and time (r = -.86, P > .05).

Conclusions:
These data indicate that in speed skating there was a shift toward polarized training over the last 38 y. This shift seems to be the most important factor in the development of Olympic speed skaters. Surprisingly there was no relation found between training hours, skating hours, and time.

Full text

93
www.IJSPP-Journal.com
ORIGINAL INVESTIGATION
International Journal of Sports Physiology and Performance, 2014, 9, 93 -99
http://dx.doi.org/10.1123/IJSPP.2013-0427
© 2014 Human Kinetics, Inc.
Thirty-Eight Years of Training Distribution
in Olympic Speed Skaters
Jac Orie, Nico Hofman, Jos J. de Koning, and Carl Foster
During the last decade discussion about training-intensity distribution has been an important issue in sports
science. Training-intensity distribution has not been adequately investigated in speed skating, a unique activity
requiring both high power and high endurance. Purpose: To quantify the training-intensity distribution and
training hours of successful Olympic speed skaters over 10 Olympiads. Methods: Olympic-medal-winning
trainers/coaches and speed skaters were interviewed and their training programs were analyzed. Each program
was qualied and quantied: workout type (specic and nonspecic) and training zones (zone 1 ≤2 mMol/L
lactate, zone 2 2–4 mMol/L lactate, zone 3 lactate >4 mMol/L). Net training times were calculated. Results:
The relation between total training hours and time (successive Olympiads) was not progressive (r = .51, P >
.5). A strong positive linear relation (r = .96, P < .01) was found between training distribution in zone 1 and
time. Zones 2 and 3 both showed a strong negative linear relation to time (r = –.94, P < .01; r = –.97, P < .01).
No signicant relation was found between speed skating hours and time (r = –.11, P > .05). This was also the
case for inline skating and time (r = –.86, P > .05). Conclusions: These data indicate that in speed skating
there was a shift toward polarized training over the last 38 y. This shift seems to be the most important factor
in the development of Olympic speed skaters. Surprisingly there was no relation found between training hours,
skating hours, and time.
Keywords: polarized training, threshold training, progression
Orie, Hofman, and de Koning are with the MOVE Research
Inst, VU University Amsterdam, Amsterdam, The Netherlands.
Foster is with the Dept of Exercise and Sport Science, University
of Wisconsin–La Crosse, La Crosse, WI.
During the last half-century the quality of speed skat-
ing performance during international competitions such
as the Olympic Games and various world championships
has continued to improve. The simple fact that world
records continue to improve is evidence that sports
performance is progressing. Almost 50% of this improve-
ment can be explained by technological improvements
(indoor ovals, klapskates, high altitude, aerodynamic
suits, excellent ice preparation), and the other 50%, by
athletic improvement.1
With the introduction of refrigerated and covered
skating rinks, athletic improvement was expected because
there are more training facilities, training locations, and
training times available, so athletes are training more
hours under better conditions. During the last 25 years,
the availability of indoor rinks has allowed skaters to do
specic practice virtually year-round, whereas before
1987, the possibility of on-ice training was limited to ~4
mo/y. Furthermore, during the last 20 years the emer-
gence of professional speed skating teams has allowed
athletes to have longer careers and potentially enhanced
performance development.
If the possibility for increased specic training hours
is improved, another question is, has better access to
specic training resulted in a change in training pattern
and practice? One major element of the training pattern
is the training-intensity distribution. Training-intensity
distribution is generally recognized as a major compo-
nent of the development of elite athletes.2–6 Studies have
shown that athletes have 2 primary patterns of training-
intensity distribution. The 2 main training models are
the threshold and polarized training patterns.3 During the
last decade, the polarized-training model has appeared to
become more common in endurance athletes. However,
the threshold-training model is still an accepted type of
training. A number of studies have been published sug-
gesting the benet of polarized training on endurance
performance.2–6 Despite the fact that during competition
the intensity distribution is dominantly at higher intensi-
ties,4–6 endurance athletes appear to train surprisingly
little in the intensity range between the lactate threshold
and the intensity of the maximal lactate steady state.
One reason for this could be the demand on glycogen
as an energetic substrate during this type of training and
the restricted training time associated with the limited
glycogen stores.7 Speed skating is a unique sport, in that
competition is dominantly at high power but also requires
signicant endurance.8,9 Furthermore, the inherent physi-
ologic response during on-ice skating is very often above
the maximal lactate steady state.

94 Orie et al
Training-intensity-distribution studies have not been
widely done in ice speed skating.8 Speed skating is differ-
ent from other endurance sports. In particular, the small
angles in knee and hip in combination with a static body
position and a long duty cycle of the skating stroke (~55%
vs ~33% during cycling and ~10% during running)10,11
results in intermittent blood-ow restriction during parts of
the skating movement. Consequently, speed skating has the
tendency to a more anaerobic character, as demonstrated in
several studies based on blood lactate measurements10–12
and muscle O2 saturation.10–13 The question arises as to if
(and how) speed skating training has evolved, given the
physiological constraints that are inherent to speed skating.
To qualify and quantify the training programs used
during the last 10 Olympiads we needed an intensity-
quantifying distribution method to evaluate these
programs. Published studies reporting the training
characteristics of endurance athletes have employed
several methods of quantifying intensity distribution.2,3,5
We choose to follow the method used by Seiler and Kjer-
land,3 because this method makes it possible to qualify
and quantify the training programs (and lactate tests) we
retrieved based on interviews with trainers, athletes, and
coaches and their stored documentation. Accordingly, the
aim of this study was to investigate the total training hours
and pattern of training-intensity distribution in elite speed
skaters. Our hypotheses were to nd an increase in train-
ing hours over the last 38 years, with a more polarized
training-intensity distribution, and given the increased
possibility of on-ice training during the last 25 years, to
nd increases in on-ice training hours.
Materials and Methods
We interviewed trainers and coaches of Dutch Olympic
medal winners in speed skating, as well as the Olympic
medal winners, in long-track middle- and long-distance
events (1500-m, 5000-m, and 10,000-m, approximate
competitive duration 2–15 min). Our analysis was limited
to male athletes. We retrieved complete training programs
of 4 Olympic Seasons (1988, 1998, 2006, and 2010) and
2 almost complete training programs (1972, missing
maximally 2 wk in total, not more than 1 wk in a row,
and 1992, missing maximally 4 wk in total, not more than
2 wk in a row). The missing weeks were discussed with
the trainers/coaches and athletes, who suggested that the
missing weeks were almost equal to the former and later
training weeks, because they were in the same period of
the training year. The athletes who trained on the analyzed
training programs were members of the National Dutch
team or of one of the commercial speed skating teams.
In total, 19 of these athletes qualied for the Olympics
(1971, 4; 1988, 3; 1992, 4; 1998, 4; 2006, 1; and 2010, 3)
and won 8 gold, 5 silver, and 4 bronze Olympic medals.
The information from the interviews allowed us to
quantify the workout intensity of each training session.
There were also testing data available from several train-
ing forms in each year. We discussed the test result for
each type of workout with the trainers/coaches to have a
better understanding of the intensity. If there were doubts
about the training intensity of a specic workout, the
workout was mimicked in a contemporary group (6 men
and 5 women) of compatibly elite speed skaters. Halfway
through these workouts and 3 minutes afterward, comple-
tion lactate concentration was obtained to allow assign-
ment of the workout to a certain zone. Measurements
were made under eld conditions (–8°C to 20°C) using
the Lactate Pro LT-1710t (ArkRay Inc, Kyoto, Japan).
The simulated workouts consisted of circuit training,
extensive and intensive endurance training, and extensive
and intensive interval training.
For each year we qualied and/or quantied each
training session on the basis of
• Training workout types: speed skating, inline skat-
ing, running, cycling, jumps, weight training, slide
board.
• Training-intensity zones: zone 1, lactate ≤ 2 mMol/L;
zone 2, lactate 2–4 mMol/L; zone 3, lactate > 4
mMol/L.4 It is difcult to distribute weight training
over these 3 training zones, so weight training was
excluded from the training distribution, but it counted
for the total training time.3
• Net training minutes: To calculate net training min-
utes we counted warm-up (5 min) + cooldown (5
min) + active training minutes (rest times within the
training were excluded). See example in Table 1.
Races were qualied as zone 3 activity, while race
preparation was rated in the zones described. For each year,
the analysis of the training program started with the rst
available training week (generally early May) and ended
on the day of the Olympic race (mid-February). The yearly
total hours were divided by the number of weeks of the
training season (40–43) to obtain training hours per week.
Table 1 Example of the Calculation of Gross and Net Training Time of 4 Different Workouts
Workout Gross Net (gross – rest)
2 h cycling 120′120′
Warm-up 5′; interval 6 × (3′—3′ rest); 5′ cooldown 46′28′
Warm-up 5′; rest 3′; interval 5 × (6′–3′ rest); 53′35′
Warm-up 10″; interval 4 × (15″–2′ rest); rest 10″; interval 6 × (10″–2′ rest) 42′12′

Training Distribution in Speed Skating 95
Results
Relationship Between Training Hours
per Week and Time
There was not a progressive increase in total training
hours (expressed as net training hours per week) across
the period of analysis (Figure 1). By comparison, during
this period the world records for the men’s 1500-, 5000-,
and 10,000-m improved on average by 18%. Probably this
does not look dramatic, but expressed in power needed
to skate at these higher speeds the increase averages
an impressive 57%. Half of this improvement can be
explained by technological improvements, and the other
half, by athletic improvement.1 The relationship between
total training hours and time (successive Olympiads) was
not signicant (r = .51, P > .5).
Relationship Between Training
Distribution and Time
Figure 2 shows the distribution of the training hours
over the 3 zones. It is clearly visible that the contribu-
tion of zone 1 has increased at the expense of zones 2
and 3. The gure shows a signicant linear increase
for the contribution of zone 1 and a signicant linear
Figure 1 — Relationship between total net training hours per week and time. The analysis is done over a time period of 38 years.
Net training time is the total training minus all rest components of the training.
Figure 2 — Relationship between training intensity distribution and time. Training intensity is divided into 3 zones: zone 1 £ 2
mMol/L lactate, zone 2 2–4 mMol/L lactate, zone 3 lactate > 4 mMol/L.

96 Orie et al
decreasing contribution for zones 2 and 3 (r of, respec-
tively, .96, P < .01; –.94, P < 0.01; and –.97, P < .01).
The training-intensity distribution in 1972 was essentially
representative of a classic threshold pattern, whereas
the training-intensity-distribution pattern after that has
become increasingly polarized in character.
Relationship Between Total Hours
Skating per Season and Time
Further analysis of specic components of training shows
that there is no systematic trend in the total hours of
on-ice speed skating across time (r = –.11, P > .05; Figure
3). Specic summer training for speed skating is inline
skating. The hours of inline skating show a decreasing
trend over the years (r = –.86, P > .05 (Figure 4).
Discussion
In the current study, we analyzed the training of 6
Olympic seasons over the last 38 years. For a proper
comparison of the different training years we calculated
the net training minutes, excluding the signicant recov-
ery time between repetitions during training sessions
that are often strongly interval in character. This was
done because in the rst decades of our analysis it was
precisely recorded which activities were done during
Figure 4 — Relationship between total net training hours inline skating per year and time.
Figure 3 — Relationship between total net training hours on-ice speed skating per year and time.

Training Distribution in Speed Skating 97
the training sessions, but no precise information was
retrievable about recovery times in training (example in
Table 1). For the later years we were able to retrieve this
information, and, for instance, the training logs of 2010
showed that net (actual) training time was about 60% of
gross (total) training time. We are aware of the fact that
training adaptation depends partly on proper recovery
time during a workout.14,15 However, it was not possible
to trace all the recovery times of the older training pro-
grams. These limitations are inherent to retrospective
data such as these, so we chose to report only the actual
time of effective training (net training). For comparison
of our calculated net training times with the literature,
this fraction (gross training time = ~1.67 × net training
time) needs to be taken into account. Still, the average
amount of training hours was small compared with other
endurance sports such as cycling, cross-country skiing,
swimming, and distance running. There is a strong varia-
tion in training hours within a speed skating season.16
In the winter, the speed skaters’ traveling program is
extensive. World Cup competitions and European and
World championships are organized around the world,
which is partly the reason for the low total seasonal train-
ing volume. Traveling, acclimatization to time zones and
altitude, and tapering are associated with a strong decline
in average training hours, which is partly the reason for
the low total training hours.
To explain the performance development over the
years, we rst hypothesized that the total amount of train-
ing hours would have been increased over the years. Our
analyses showed that this was not the case, although it
is well know that training volume is an important para-
meter for a training effect.17,18 One reason for the lack
of a signicant progression of training hours across time
could be the fact that specic speed skating workouts,
in crouched position as used during races, restrict blood
ow in the legs.9–13 Combined with the minimal speed
necessary for maintaining a proper technique, this could
be of such high intrinsic intensity that only a limited
volume can be tolerated, both in terms of momentary
metabolite burden and in terms of availability of glycogen
as an energetic substrate. In favor of specic adaptation,
workouts involving movements with small knee- and hip-
joint angles are important. There is evidence that muscle
load during weight-bearing activities with these small
joint angles causes blood-ow restriction. This restricted
blood ow causes an oxygen decit in the working muscle
bers, which results in a more anaerobic energy contri-
bution. This anaerobic character may explain why many
specic training hours are qualied in moderate or high
intensities (zones 2 and 3). Too much anaerobic training
might hinder endurance-capacity development.19 This is
in line with our current nding, namely, a lack of increase
in speed skating hours, a declining tendency of inline
training hours, and no increase in total training hours.
According to trainers and coaches, on-ice speed skating
training is an important factor in enhancing speed skat-
ing efciency. However, there was no increase in speed
skating hours, although there was a distinct change in
terms of spreading of the skating hours over the season.
Since the existence of covered speed skating ovals, speed
skating is possible during part of the summer. In the past,
speed skating was more concentrated during the winter
months. The spreading over the training season of the
specic training hours, which tend to be inherently high
intensity, might be an important factor in terms of allow-
ing the training distribution to evolve to a more polarized
character. Given the change in performance, this wider
distribution of skating hours appears to be benecial for
the development of athletes. Additional research is needed
to explore this process.
A remarkable result of our analysis was the obser-
vation of a dramatically changed distribution of training
hours over the different zones. There was an obvious
increase in the percentage of training hours spent in zone
1 (lactate ≤2 mMol/L) and a decrease in the percentage
of training hours in zones 2 and 3 over the years. The
observed increase in the percentage of low-intensity
training is also demonstrated in other studies of endur-
ance athletes in a variety of sports.2–6,8,20 In contrast to
our study, the study of Fiskerstrand and Seiler2 used a
2-intensity-zones model (low or high intensity). Further-
more, Seiler and Kjerland,3 Seiler,4 and Esteve-Lanao et
al5,6 (3-zone quantication model) and Yu et al8 showed,
as well, that a larger amount of low-intensity training
is effective in stimulating physiological adaptations.
Although different experimental research designs were
used (cross-sectional in Seiler and Kjerland3 and Esteve-
Lanao et al,5,6 quasi-experimental in Yu et al8), their
results demonstrate that a larger percentage of training
in zone 1 was benecial. Geijsel21 reported in 1979 that
marathon skaters who had formerly done a lot of skat-
ing simulations (high intensity) became better as they
added more cycling (lower intensity) to their training
program. In addition, Seiler and Kjerland3 and Yu et al8
reported that training at low intensities in combination
with a much smaller amount of training hours at moderate
intensity (zone 2) compared with high intensity (Zone 3)
is even more effective. This combination of low and high
intensities is called polarized training. Our data showed
that both moderate and high training zones declined over
time but the percentage of moderate training intensities
is still higher than the percentage at high training intensi-
ties. This is reasonably attributable to the inherently high
intensity of skating training, such that even the so-called
endurance sets cannot be accomplished in zone 1 and
became zone 2 activities.22
A potential limitation of the 3-intensity-zone quanti-
cation model used in the current study is the resolution
of the intensity zones. The division of the high-intensity
zone in only 1 zone with lactate >4 mMol/L is especially
open to discussion. In this relatively large zone many
different training intensities specic for the different
competitive race distances can be used. Stepto et al23
identied 5 intensity zones at 80%, 85%, 90%, 100%,
and 175% of the subjects’ maximal power output within
our zone 3. They showed a U-shaped relation between
physiological adaptation and training intensity. Because

98 Orie et al
of this U-shaped relation, the conclusion can be made
that zone 3 is polarized by itself. In that case it could
be more appropriate to speak about a 3-wave model. To
avoid the early mentioned limitation and to get a better
understanding of training-intensity distribution, in future
investigations a method with a better differentiation in
this zone could be benecial.
Conclusion
Our data indicate that for successful middle- and long-
distance speed skaters there was a shift toward polarized
training over the last 38 years. Surprisingly, there was no
increase in net training hours and hours of on-ice skating
over these years, while performance increased consider-
ably. The current ndings clearly show the importance
of training-intensity distribution.
Practical Application
Our ndings could be an important guide for trainers/
coaches to balance their training programs and to transfer
results from research in other sports to speed skating. On
the basis of the presented ndings it can be concluded
that changes in training-intensity distribution were an
important factor in the development of speed skating
during the last 38 years.
Acknowledgments
We would like to thank coaches, trainers, and athletes for their
time and willingness to cooperate in this research project.
References
1. de Koning JJ. World records: how much athlete? how much
technology? Int J Sports Physiol Perform. 2010;5:262–
267. PubMed
2. Fiskerstrand A, Seiler KS. Training and performance
characteristics among Norwegian international rowers
1970–2001. Scand J Med Sci Sports. 2004;14:303–310.
PubMed doi:10.1046/j.1600-0838.2003.370.x
3. Seiler KS, Kjerland GO. Quantifying training intensity
distribution in elite endurance athletes: is there evi-
dence for an “optimal” distribution? Scand J Med Sci
Sports. 2006;16:49–56. PubMed doi:10.1111/j.1600-
0838.2004.00418.x
4. Seiler S. What is best practice for training intensity and
duration distribution in endurance athletes? Int J Sports
Physiol Perform. 2010;5:276–291. PubMed
5. Esteve-Lanao J, San Juan AF, Earnest CP, et al. How do
endurance runners actually train?: relationship with com-
petition performance. Med Sci Sports Exerc. 2005;37:496–
504. PubMed doi:10.1249/01.MSS.0000155393.78744.86
6. Esteve-Lanao J, Foster C, Seiler S, et al. Impact of train-
ing intensity distribution on performance in endurance
athletes. J Strength Cond Res. 2007;21:943–949. PubMed
7. Beneke R, Leithäuser RM, Ochentel O. Blood lactate
diagnostics in exercise testing and training. Int J Sports
Physiol Perform. 2011;6:8–24. PubMed
8. Yu H, Chen X, Zhu W, Cao C. A quasi-experimental study
of Chinese top-level speedskaters’ training load: threshold
versus polarized model. Int J Sports Physiol Perform.
2012;7:103–112. PubMed
9. Rundell KW. Compromised oxygen uptake in speed
skaters during treadmill in-line skating. Med Sci Sports
Exerc. 1996;28:120–127. PubMed doi:10.1097/00005768-
199601000-00023
10. Rundell KW, Nioka S, Chance B. Hemoglobin/myoglobin
desaturation during speed skating. Med Sci Sports Exerc.
1997;29:248–258. PubMed doi:10.1097/00005768-
199702000-00014
11. Foster C, Rundell KW, Snyder AC, et al. Evidence
for restricted muscle blood ow during speed skating.
Med Sci Sports Exerc. 1999;31:1433–1440. PubMed
doi:10.1097/00005768-199910000-00012
12. Hesford CM, Laing S, Cardinale M, Cooper CE. Effect of
race distance on muscle oxygenation in short-track speed
skating. Med Sci Sports Exerc. 2013;45:83–92. PubMed
doi:10.1249/MSS.0b013e31826c58dd
13. Hesford C, Cardinale M, Laing S, Cooper CE. NIRS mea-
surements with elite speed skaters: comparison between the
ice rink and the laboratory. Adv Exp Med Biol. 2013;765:81–
86. PubMed doi:10.1007/978-1-4614-4989-8_12
14. Rønnestad BR, Hansen J. Optimizing interval training at
power output associated with peak oxygen uptake in well-
trained cyclist [published online ahead of print August 12,
2013]. J Strength Cond Res. PubMed
15. Millet GP, Candau R, Fattori P, Bignet F, Varray A. VO2
responses to different intermittent runs at velocity associ-
ated with VO2max. Can J Appl Physiol. 2003;28:410–423.
PubMed doi:10.1139/h03-030
16. van Ingen Schenau GJ, Bakker FC, de Groot G, de Koning
JJ. Supramaximal cycling tests do not detect seasonal
progression in performance in groups of elite speed skat-
ers. Eur J Appl Physiol. 1992;64:292–297. doi:10.1007/
BF00636214
17. Volkov N. Human Bioenergetics in Strenuous Muscu-
lar Activity and Pathways of Improved Performance in
Sportsmen. Moscow, Russia: Anokin Research Institute
of Normal Physiology; 1990.
18. Hawley JA, Stepto NK. Adaptations to training in endur-
ance cyclists: implications for performance. Sports Med.
2001;31:511–520. PubMed doi:10.2165/00007256-
200131070-00006
19. Furrer R, de Haan A, Bravenboer N, Kos D, Lips P,
Jaspers RJ. Effects of concurrent training on oxidative
capacity in rat gastrocnemius muscle. Med Sci Sports
Exerc. 2013;45:1674–1683. PubMed doi:10.1249/
MSS.0b013e31828fc65f
20. Guellich A, Seiler S, Emrich E. Training methods and
intensity distribution of young world-class rowers. Int J
Sports Physiol Perform. 2009;4:448–460. PubMed
21. Geijsel JS. Training and testing in marathon speed skating.
J Sports Med Phys Fitness. 1979;19:277–284. PubMed

Training Distribution in Speed Skating 99
22. Beneke R, von Duvillard SP. Determination of maximal
lactate steady state response in selected sports events.
Med Sci Sports Exerc. 1996;28:241–246. PubMed
doi:10.1097/00005768-199602000-00013
23. Stepto NK, Hawley JA, Dennis SC, et al. Effects of dif-
ferent interval-training programs on cycling time-trial
performance. Med Sci Sports Exerc. 1999;31:736–741.
PubMed doi:10.1097/00005768-199905000-00018