Content uploaded by Stephen Seiler
Author content
All content in this area was uploaded by Stephen Seiler on Jul 16, 2014
Content may be subject to copyright.
Available via license: CC BY 4.0
Content may be subject to copyright.
Content uploaded by Espen Tønnessen
Author content
All content in this area was uploaded by Espen Tønnessen on Jun 15, 2014
Content may be subject to copyright.
The Road to Gold: Training and Peaking Characteristics
in the Year Prior to a Gold Medal Endurance Performance
Espen Tønnessen
1
*, Øystein Sylta
2
, Thomas A. Haugen
1
, Erlend Hem
1
, Ida S. Svendsen
3
, Stephen Seiler
2
1The Norwegian Olympic Federation, Oslo, Norway, 2Faculty of Health and Sport Sciences, University of Agder, Kristiansand, Norway, 3School of Sport, Exercise and
Health Sciences, Loughborough University, Leicestershire, United Kingdom
Abstract
Purpose:
To describe training variations across the annual cycle in Olympic and World Champion endurance athletes, and
determine whether these athletes used tapering strategies in line with recommendations in the literature.
Methods:
Eleven elite XC skiers and biathletes (4 male; 2861 yr, 8565 mL. min
21
.kg
21
_
VVO
2max, 7 female, 2564 yr,
7363 mL. min
21
.kg
21
_
VVO
2max) reported one year of day-to-day training leading up to the most successful competition of
their career. Training data were divided into periodization and peaking phases and distributed into training forms, intensity
zones and endurance activity forms.
Results:
Athletes trained ,800 h/500 sessions.year
21
, including ,500 h. year
21
of sport-specific training. Ninety-four
percent of all training was executed as aerobic endurance training. Of this, ,90% was low intensity training (LIT, below the
first lactate threshold) and 10% high intensity training (HIT, above the first lactate threshold) by time. Categorically, 23% of
training sessions were characterized as HIT with primary portions executed at or above the first lactate turn point. Training
volume and specificity distribution conformed to a traditional periodization model, but absolute volume of HIT remained
stable across phases. However, HIT training patterns tended to become more polarized in the competition phase. Training
volume, frequency and intensity remained unchanged from pre-peaking to peaking period, but there was a 32615% (P,
.01) volume reduction from the preparation period to peaking phase.
Conclusions:
The annual training data for these Olympic and World champion XC skiers and biathletes conforms to
previously reported training patterns of elite endurance athletes. During the competition phase, training became more
sport-specific, with 92% performed as XC skiing. However, they did not follow suggested tapering practice derived from
short-term experimental studies. Only three out of 11 athletes took a rest day during the final 5 days prior to their most
successful competition.
Citation: Tønnessen E, Sylta Ø, Haugen TA, Hem E, Svendsen IS, et al. (2014) The Road to Gold: Training and Peaking Characteristics in the Year Prior to a Gold
Medal Endurance Performance. PLoS ONE 9(7): e101796. doi:10.1371/journal.pone.0101796
Editor: Matjaz Perc, University of Maribor, Slovenia
Received April 29, 2014; Accepted June 10, 2014; Published July 14, 2014
Copyright: ß2014 Tønnessen et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The authors confirm that all data underlying the findings are fully available without restriction. All relevant data are within the paper and its
Supporting Information files.
Funding: The authors received no specific funding for this work.
Competing Interests: The authors have declared that no competing interests exist.
* Email: espen.tonnessen@olympiatoppen.no
Introduction
Winning a gold medal in a major international championship
requires not only outstanding athletic ability and long-term
training progression, but also that the athlete achieves peak
performance at the right time. In recent years, increased attention
has been given to quantifying the training characteristics of elite
endurance athletes [1–3] and this information has provided a
fruitful foundation for hypothesis testing regarding training load
and physiological adaptation. At the same time, a strong
knowledge base has developed regarding best practice for the
tapering and peaking process, based largely on experimental
interventions [4–6]. However, studies linking the characteristics of
the long-term training process to those of the short term pre-
peaking and peaking process are lacking.
Recently, a number of descriptive studies, both retrospective
and prospective, have been published on the training character-
istics of athletes from endurance sports such as running [7–12],
cycling [13–14], XC skiing [15–17], swimming [18–19], rowing
[20–21], triathlon [22–23], speed skating [24–25] and kayaking
[26]. Training load variables such as volume, frequency and
intensity distribution appear to play an interactive role in
maximizing physical capacity and performance [27]. Depending
on the specific muscular loading characteristics of the sport,
athletes typically train 500 h (distance running) [7,8,11,12,28,29]
to well in excess of 1000 h per year (rowing, swimming, cycling,
triathlon) [13–14,18–23] performed during 400–800 annual
training sessions [11–12,15–17,23], in order to reach an interna-
tionally elite level. When examining the intensity distribution of
this large training volume, a number of studies across a broad
range of sports converge on the finding that 75–90% of all
endurance training time is performed as low intensity training
(LIT, below the first lactate turn point) for athletes training .
PLOS ONE | www.plosone.org 1 July 2014 | Volume 9 | Issue 7 | e101796
500 h per year. The remaining 10–25% is comprised of high
intensity training (HIT) performed above the first lactate turn
point [7–8,14–15,22–23]. This approximate ‘‘80–20’’ distribution
between low and high intensity training among high-level
performers is a robust finding [3], even if mechanistic explanations
for its ubiquity remain speculative. Furthermore, the ‘‘best
practice’’ magnitude and distribution of HIT remains unclear.
In addition, the training-dose response relationship has a
significant individual variation component, which further compli-
cates the generalizability of the picture. While the influx of
descriptive data on the training of elite endurance athletes has
informed training practice, and stimulated new experimental
studies, methodological compromises are inherent to the challenge
of measuring behavior reliably and precisely in highly selected
groups. Available descriptive studies typically only present data
over a shorter time frame [7–8,12], at a sub-elite level [9,14–
15,17,21–22] or as single case studies [11,23,28–29]. Accuracy in
training monitoring is also unclear, due to weaknesses in methods
such as questionnaires [10] or compilation and analysis of data
that are in part or completely based on training plans instead of
strict quantification of actual training performed [7–8,20]. Limited
data currently exist on the long-term training of highly trained and
elite athletes, based on accurate training monitoring [30].
Short-term training manipulations to achieve peaking for
optimal sports performance have been investigated for more than
20 years. A synthesis of studies on well-trained athletes from a
variety of different sports has shown a performance improvement
of up to 3%, providing the final 4–28 days of training are executed
correctly [5–6,31–33]. All other things being equal, a well-
executed peaking phase can therefore dramatically increase the
odds of winning a gold medal at a championship event for an
individual athlete with finalist potential. In the research literature,
the peaking process is typically divided into two phases: a pre-
tapering phase, and the final taper period culminating with the
intended competition. The aim of the pre-tapering phase is to
stimulate a controlled ‘‘over-reaching’’ state and elicit a super-
compensatory adaptive response in the following taper. Experi-
mentally, the optimal duration of the actual taper depends on the
training executed in the pre-taper [6,33–34]. The taper is initiated
approximately 14 days prior to the desired peak performance, and
the aim of this phase is to facilitate regeneration and reduce
fatigue, while maintaining or increasing fitness and technical/
psychological readiness in order to mobilize maximal performance
in competition [4–6,35]. The optimal training volume, frequency
and intensity in each phase is debated [4–6,31,35], but a reduction
in training load of between 41–60% from pre-taper to taper has
been recommended. This reduction appears to be best achieved
via reduced training duration per session, while maintaining
session frequency [5]. Maintaining training intensity is considered
to be a key factor in a successful peaking regime, and it is therefore
recommended that the frequency of HIT sessions be maintained
during the taper [36–39]. Despite 20 years of research on tapering
and peaking for athletic performance, we are unaware of any study
that has successfully quantified the ‘‘real life’’ peaking strategies of
athletes achieving ultimate success in Olympic Games or World
Championship events. It is therefore unclear how the models
developed from controlled experimental studies are translated to
the actual peaking practices of elite endurance athletes. Further-
more, few studies havelinked peaking strategies to annual training
characteristics and the competitive season of elite athletes [30].
The current study therefore aimed to: 1) present highly accurate
day-to-day annual training data from a cohort of endurance
athletes that all won Olympic or World Championship gold
medals and, 2) quantify and examine relationships between annual
training and peaking characteristics in these athletes.
Methods
Subjects
Four male and seven female former and current Norwegian elite
XC skiers and biathletes were included in the study (Table 1). All
athletes had won a least one individual Olympic or World
Championship senior gold medal during their career. In total,
included males had won 41 (5–26) and females 25 (1–9) gold
medals (includes both individual and relays from 1985 through
2011). In addition, included athletes had systematically and
accurately recorded their day-to-day training in detail from junior
through to senior level. In the current study, we have analyzed and
reported the year specifically leading up to their most successful
competition at senior level. The regional ethics committee of
Southern Norway reviewed the study and concluded that, due to
the nature of the investigation, it did not require their approval.
The study was therefore submitted to and approved by the
Norwegian Social Science Data Services (NSD), and all athletes
gave their oral and written informed consent prior to study
participation.
Physiological testing
All athletes underwent regular physiological testing during their
career. The test values presented in Table 1 represent the highest
result achieved during the analyzed year. There were no
physiological tests performed during the competition period, and
the presented results therefore represent tests from October or
November, while Olympic or World Championship events were
typically held in February-March. All physiological testing was
conducted at the Norwegian Olympic training center. _
VVO
2max
testing was performed as running at 10.5% inclination on a
motorized treadmill (Woodway Gmbh, Weil am Rhein, Germany)
calibrated for speed and incline. The procedure started with an
extensive warm-up sequence, followed by a stepwise increase in
running velocity every minute thereafter until volitional exhaus-
tion, normally occurring after 4–6 minutes. Starting velocity for all
athletes corresponded to 85–90% of _
VVO
2max. The increase was
1 km.h
21
.min
21
, and the last velocity step was held for at least
1 min. The test was terminated before voluntary exhaustion if the
_
VVO
2values leveled off or decreased despite increasing workload
and ventilation, in addition to respiratory exchange ratio (RER) .
1.10. _
VVO
2max was defined as the highest average of two
consecutive 30 s measurements. Oxygen uptake was measured
using EOS Sprint (Jaeger-Toennis, Wurtzburg, Germany) until
2002, after which an Oxycon Pro (Jaeger-Toennis, Wurtzburg,
Germany) metabolic test system was used. An internal comparison
between the two analyzers was conducted during the transition in
2002 and showed identical regression lines for the treadmill
running velocity – _
VVO
2relationship with both systems. Primarily
two exercise physiologists supervised all testing during the entire
period.
Training monitoring
Athletes included in the study recorded their day-to-day training
during their most successful year in paper diaries designed by the
Norwegian Ski Association [40–41], the Norwegian Biathlon
Association [42] or, since ,2005, in the digital version developed
by the Norwegian Olympic Federation (OLT). The training
recorded for each session included total training time distributed
across training form (strength, endurance, sprint), activity form
Training and Peaking of Gold Medallists
PLOS ONE | www.plosone.org 2 July 2014 | Volume 9 | Issue 7 | e101796
(skiing, roller-skiing, running, cycling etc.), and intensity zone, as
well as specific comments regarding session details. All paper
training diaries were transferred session by session to digital format
by persons from the current research group. Total training time
and frequency of sessions was distributed in line with the structure
in Figure 1. Digitized diary data was rigorously cross-checked for
internal consistency among different training distribution break-
downs at the individual level. Internal consistency of digitized
training records from all included athletes was $99%.
All the athletes included in the study used a 5 intensity zone
model, where zones 1–2 are classified as LIT and zones 3–5 as
HIT. The intensity scale presented in Table 2 represents average
self-reported zone-cut offs from 29 elite XC-skiers from a previous
study [43]. In the results section we have presented the data either
in a binary model (LIT/HIT) or a 5-zone model were zones 1–2
are below the first lactate threshold (LT
1
), zone 3 between LT
1
and LT
2
, and zones 4–5 above LT
2
[3,44]. The intensity
distribution is classified both according to a time in training zone
approach and a frequency based session goal approach (SG).
These methods and the intensity zones cut-offs have been
described in detail recently [43].
Annual periodization phases and peaking model
General training data from the entire year are either presented
as annual training characteristics or divided into different
periodization phases as presented in Table 3. Peaking character-
istics were quantified based on the final 6 weeks of training prior to
the gold medal winning performance, as delineated in Table 3.
Statistical analyses
All data in text, tables or figures are presented as mean 6
standard deviation (SD) and/or range. Statistical comparisons
between different periodization phases are focused on the general
preparation period (GP), specific preparation period (SP) and
competition period (CP) in addition to comparing the actual
peaking phase with pre-peaking phase, GP and SP. Data were not
normally distributed. Therefore each variable from the GP, SP
and CP (overall, pre-peaking and peaking phase) was tested with a
non-parametric Friedman test, followed by a post-hoc test
(Wilcoxon Signed Rank) to locate statistical differences. Male
and female athlete data are merged, as a Mann-Whitney U Test
revealed no significant differences in any relevant variables across
gender (data not shown). All figures and statistical analyses were
performed using Microsoft Excel or SPSS 18.0 (SPSS Inc,
Chicago, IL, USA) and statistical significance was accepted at
the P,.05 level or Bonferroni adjusted alpha level.
Results
Annual training characteristics
Total training volume was 770699 h (622–942) distributed
across 470668 sessions (375–585) throughout the gold medal year.
Endurance training accounted for 9463% of all training time with
the remaining 562% composed of strength training and 161% ski
sprint training. Time in training zone based intensity distribution
showed that 9161% of all endurance training time was executed
as LIT (zone 1–2) and 961% as HIT (zone 3–5).
Monthly training distribution of specific and non-specific
activity forms during each training phase are presented in
Figure 2. Endurance and sprint training was executed with
sport-specific movement patterns (ski or roller ski) for 6463%
(465656 h/min-max: 376–569 h) of total training time, with the
remaining 3663% (265647 h/min-max: 196–337 h), composed
of non-specific activity forms (running, cycling etc.) throughout the
year. The proportion of sport-specific training increased signifi-
cantly from GP (4866%) to SP (8768%) and CP (9264%) (P,
.01).
The distribution across all five intensity zones was: zone 1:
86.063.4%, zone 2: 5.363.0%, zone 3: 3.360.9%, zone 4:
3.361% and zone 5: 2.161.0%. When all endurance sessions
were nominally categorized using the SG approach, the distribu-
tion was 7762% LIT and 2362% HIT (Figure 3, A). Weekly
training patterns during each training phase are presented in
Table 4.
Total annual HIT duration (including competitions) was
63614 h (46–85 h) distributed across 106620 sessions (85–147)
throughout the year. The relative distribution of HIT duration in
intensity zones 3, 4 and 5 was 39610%, 37613% and 24613%
respectively, and 3266%, 38614% and 30613% according to a
SG distribution. Monthly frequency of HIT sessions increased
Table 1. General characteristics of athletes included in the study.
Subject Gender Age Height Weight _
VVO
2max (ml
.
kg
21
.min
21
)_
VVO
2max (l.min
21
)
1 M 28 180 77 92.5 7.13
2 M 26 190 82 81.9 6.73
3 M 29 189 83 84.8 7.07
4 M 28 179 66 81.2 5.25
5 F 23 172 55 72.9 3.90
6 F 23 176 63 73.6 4.64
7 F 29 173 63 76.6 4.81
8 F 20 175 69 70.4 4.83
9 F 28 166 61 69.1 4.24
10 F 22 162 51 76.0 3.93
11 F 30 169 64 71.4 4.60
Mean 6SD, Male 2861 185667768 85.165.2 6.560.9
Mean 6SD, Female 2564 170656166 72.962.8 4.460.4
Values are reported from the analyzed year in the current study.
doi:10.1371/journal.pone.0101796.t001
Training and Peaking of Gold Medallists
PLOS ONE | www.plosone.org 3 July 2014 | Volume 9 | Issue 7 | e101796
from GP to SP (P,.01). In addition, the monthly frequency of
intensity zone 5 sessions increased from GP to SP and then
remained unchanged in the CP (P,.01) (Figure 3, B). Weekly HIT
patterns during each training phase are presented in Table 4.
Peaking characteristics
Total training time (h.wk
21
) decreased by 9614% from the pre-
peaking to peaking phase, but this did not reach statistical
significance. However, the reduction from GP, when training
volume was highest, to the peaking phase, was 32615% (P,.01).
This decrease in total training volume was entirely due to a
reduction in non-sport-specific training. Individual data for each of
the 11 athletes are presented in Figure 4. The decrease in training
volume from GP and pre-peaking phase to the peaking phase was
achieved via a reduction in both endurance and strength training,
while sprint training time remained stable, although there was a
tendency for sprint training time to increase slightly from the pre-
peaking phase to the peaking phase. There were no significant
changes in total session frequency per week between the peaking
phase and any of the other phases (Figure 5 A and Table 4).
There was non-significant decrease of 9615% in LIT
endurance training (h.wk
21
) from the pre-peaking phase to the
peaking phase. However, LIT training volume decreased by
31617% (P,.01) from GP to the peaking phase. In contrast, HIT
time (h.wk
21
) remained stable from both pre-peaking phase to the
Figure 1. Training distribution methods. Total training time was divided into training forms (endurance, sprint and strength). Endurance time
and frequency were further distributed into 5 intensity zones in line with table 2. Zones 1–2 are LIT and zones 3–5 are HIT. Endurance and sprint time
together were divided into activity forms. Ski and roller ski were classified as specific, and running, cycling or other as non-specific activity forms.
doi:10.1371/journal.pone.0101796.g001
Table 2. The 5-zone, 3-zone, and binary intensity scales used in the current study.
Intensity Zone Typical Blood lactate
A
(mmol. L
21
) Typical Heart Rate (% max) Three zone model Binary model
5.5.8 .94 .LT
2
4 3.7–5.7 89–93 HIT
3 2.1–3.6 84–88 LT
1
–LT
2
2 1.3–2.0 74–83 LIT
1,1.2 54–73 ,LT
1
A
Measured with Lactate Pro LT-1710. Reference values presented are derived from the average self-reported zone-cut offs of 29 elite XC-skiers [43], and individual
adjustments are required.
doi:10.1371/journal.pone.0101796.t002
Training and Peaking of Gold Medallists
PLOS ONE | www.plosone.org 4 July 2014 | Volume 9 | Issue 7 | e101796
peaking phase and from GP to the peaking phase (Figure 5 A and
Table 4).
LIT endurance session frequency decreased from GP to the
peaking phase by 21624% (P= .016) but remained stable from
pre-peaking phase to the peaking phase. Weekly HIT session
frequency increased by 40627% (P,.01) from GP to the peaking
phase, but remained stable from pre-peaking phase to the peaking
phase (Figure 5 B). Training volume and frequency distribution
among zones 3, 4 and 5 through the different phases are presented
in Figure 5 B and Table 4.
Discussion
To the authors’ knowledge, this is the first study to connect
accurate annual day-to-day training data to a specific peaking
period in a group of athletes achieving ultimate international
success in an endurance sport. The main findings of the present
study are: 1) The annual training data for these Olympic and
World champion XC skiers and biathletes conforms to previously
reported training patterns amongst elite endurance athletes. 2) In
contrast, peaking characteristics for these gold medalists did not
conform to suggested best practice for tapering strategies in elite
endurance athletes, as derived from partly experimental studies.
Annual training characteristics
Training volume. High training volume has emerged as a
key commonality in successful endurance training [20,1–3,25].
Athletes in the current study trained ,800 h.year
21
across ,500
annual training sessions although there were individual differenc-
es. This finding is in line with previous studies reporting training
volume in elite XC skiers [1,16–17]. Muscular loading differences
and stress associated with different activities probably explain why
there is large variation in reported annual training volume across
sports. For example, top international runners are reported to
train ‘‘only’’ 500–600 h.year
21
[7–8] while a case study of an
international level triathlete reports .1000 h.year
21
[23]. The
current data show a tendency for developments in training
Table 3. Training phases in annual cycle, including peaking phases.
Period in annual training cycle Duration
Preparation period (PP) May-December
Transition period May
General preparation period (GP) June-October
Specific preparation period (SP) November-December
Competition Period (CP) January-March
Pre-peaking phase 6–3 weeks before championship event
Peaking phase Last 14 days before championship event
Regeneration period April
doi:10.1371/journal.pone.0101796.t003
Figure 2. Annual organization of specific and non-specific activity forms. Endurance and sprint training time (h) distributed into specific (ski
and roller ski) and non-specific (running, cycling and other) activity forms during each month and divided in phases. #Difference in specific training
time vs. GP (P,.01).
doi:10.1371/journal.pone.0101796.g002
Training and Peaking of Gold Medallists
PLOS ONE | www.plosone.org 5 July 2014 | Volume 9 | Issue 7 | e101796
patterns during the time period from 1985 to 2011, with a positive
relationship between total training volume and year of champi-
onship title (r = .59, P= .055). Increased training volume appears
to be mainly due to increased frequency of training sessions from
1985 to 2011, while average duration per training session has
remained relatively stable at 1.760.2 h.
During the entire training year, 94% of all training time was
executed as endurance training. However, strength and sprint
training appear to play an important role in the training of XC
skiers [45]. Strength training was carried out as general, specific or
maximal, while sprints included both specific ski sprint-related
exercises and jumps. Interestingly, ,90% of all strength and sprint
training was executed during the preparation period (PP). In
practice, this means two to three strength and sprints sessions.-
week
21
in PP compared to one weekly session during CP, typically
conducted at the end of endurance training sessions. The main
underlying philosophy for these athletes was to build up a
prescribed strength level during PP and then maintain this level
during CP. Unfortunately, systematic strength testing documen-
tation was not available for these athletes. We are therefore not
able to verify whether strength characteristics of these athletes
were stable during CP. However, previous research suggests that
one bout of strength training per week is sufficient to maintain
strength levels over shorter time frames [46].
Figure 3. Annual training characteristics. A: Total training time (h) distributed into endurance training (zones 1–5), strength and sprint (bars, y-
axis), and total training frequency (sessions) (line, z-axis) during each month and divided into phases. B: HIT frequency (sessions) distributed into
zones 3, 4 and 5 (bars, y-axis) during each month and divided in phases. There was a statistically significant difference (P,.05) in total HIT sessions
and zones 3, 4 and 5 respectively across the GP, SP and CP. Pairwise post-hoc tests showed: * Difference in total HIT sessions across phases (P,.01). #
Difference between zone 5 sessions vs. GP (P,.01).
doi:10.1371/journal.pone.0101796.g003
Training and Peaking of Gold Medallists
PLOS ONE | www.plosone.org 6 July 2014 | Volume 9 | Issue 7 | e101796
Table 4. Weekly training patterns during different phases throughout the season.
Competition phase
Weekly training patterns
Transition phase General preparation phase Specific preparation phase Overall Pre-peaking Peaking Regeneration phase
Total training:
Training time (h.wk
21
)13.965.3 17.962.5 16.562.4 11.662.2
#b
13.563.1
#b
12.162.4
#b
6.163.7
Tr. sessions
.
wk
21
8.162.8 10.161.0 10.361.7 8.362.1
#b
8.961.9
#b
9.262.3 4.362.3
Training forms:
Endurance (h
.
wk
21
)12.764.9 16.562.6 15.762.2 11.262.0
#b
12.962.8
b
11.762.3
#b
5.963.6
Strength (h
.
wk
21
)1.161.0 1.260.5 0.760.4 * 0.360.2
#b
0.560.4
#
0.260.2
#b
0.260.3
Sprint (h
.
wk
21
)0.160.1 0.260.2 0.160.1 0.160.1 0.160.1 0.260.2 0.060.0
Intensity distribution:
Zone 1 (h
.
wk
21
)10.864.7 14.362.6 13.862.1 9.462.3
#b
11.262.8
b
9.762.0
#b
4.962.9
Zone 2 (h
.
wk
21
)1.360.9 0.960.6 0.660.5 * 0.460.3
#
0.460.5
#
0.660.8 0.360.5
Zone 3 (h
.
wk
21
)0.360.2 0.660.2 0.560.2 0.460.3 0.460.2
#
0.460.4 0.260.3
Zone 4 (h
.
wk
21
)0.260.2 0.560.2 0.460.3 0.560.3 0.560.4 0.760.4
b
0.360.3
Zone 5 (h
.
wk
21
)0.160.1 0.260.1 0.460.2 * 0.560.3 0.460.3 0.260.3 0.260.2
Activity forms:
Specific (h
.
wk
21
)4.363.3 8.161.7 13.762.3 * 10.361.5
#b
12.262.6
#
10.861.8
#
4.163.2
Non-Specific (h
.
wk
21
)8.564.6 8.661.5 2.161.2 * 1.060.6
#b
0.860.6
#b
1.061.0
#b
1.861.5
Values are mean 6SD and represent training hours per week in different phases. A non-parametric Friedman test indicated that there was a statistically significant difference in all variables across the GP, SP and CP (P,.05 level). A
pairwise post-hoc test (Wilcoxon Signed Rank) was used to determine whether there was a statistically significant difference between the GP, SP or CP, as well as pre-peaking and peaking phases (P,.01 level).
*P,.01, GP vs. SP;
#
P,.01, GP vs. CP, pre-peaking or peaking;
b
P,.01, SP vs. CP, pre-peaking or peaking;
{
P,.01, pre-peaking vs. peaking.
doi:10.1371/journal.pone.0101796.t004
Training and Peaking of Gold Medallists
PLOS ONE | www.plosone.org 7 July 2014 | Volume 9 | Issue 7 | e101796
Activity forms. During the entire training year, 64%
(,500 h) of all training was executed with sport-specific move-
ment patterns (skiing/roller-skiing). However, over the course of
the training year the amount of specific training increased from
,50 to 90%. That is, in line with the early periodization models
[48], when training load was highest in PP only ,50% of all
training was executed as ski or roller-ski. Otherwise, when training
load was lowest in CP, .90% was performed as sport-specific
training.
Sport-specific training is outlined as a key to improving _
VVO
2max
[51–52]. Hence, a high portion of sport-specific training during
the CP for these athletes appears to be essential in order to reach
an international performance level. However, we maintain that a
large volume of non-specific activity forms during PP serve an
important purpose in increasing trainability and improving
general aerobic capacity [53–54].
Intensity distribution. Recently there has been some debate
regarding findings suggesting that HIT induces superior physio-
logical and performance adaptations compared with LIT [47].
The trend among endurance athletes is to adopt a polarized
intensity distribution model integrating both intensity domains
[7,9,16,20,25]. The present data consistently demonstrate that
these 11 gold medalists executed a large proportion of their total
training as LIT throughout the annual cycle. Total LIT time was
progressively increased during PP, in line with some key features
from the early periodization models of Matwejew [48], before
being reduced dramatically during CP. However, it is important to
emphasize that the marked intensity shift to more HIT described
in Matwejew’s models was not observed in this group of elite
athletes.
The current study contributes unique knowledge to our
understanding of the self-selected duration and distribution of
HIT in elite endurance sports. Depending on the quantification
methods used [43], results from several other studies suggest that
an approximate 80/20% LIT/HIT distribution is optimal,
although the percentage of HIT varies from ,10–30% [7–9,14–
16,44,49–50] using a time in training zone method [43]. However,
in the current study only 9% of annual endurance training time, or
,60 h/,100 sessions were reported to be above LT
1
. This is in
contrast to other top Olympic athletes reported to perform a
greater amount of HIT in addition to high total training volume
[23,28–29]. The total volume of HIT training was evenly
distributed throughout the year with an average of 562h or
963 sessions
.
month
21
. Interestingly, it was also found that HIT
training sessions were distributed virtually equally among zones 3,
4 & 5, with average durations of 0.8/0.6/0.5 h in zones 3/4/5
respectively. However, from the PP to CP, both duration and
frequency in zones 3 and 4 were maintained, while the frequency
of zone 5 training sessions increased. That is, as the main
performance peak came closer, LIT time decreased dramatically
while HIT patterns shifted towards a more polarized model,
despite virtually constant HIT training time.
Peaking practice
Training volume and specificity. Optimizing the reduction
in training load during the peaking phase is believed to be a key to
optimal championship performance [6,30]. Training load is
described as a combination of training volume, intensity and
frequency [27]. A meta-analysis conducted by Bosquet et al [5]
concluded that athletes could maximize taper-associated benefits
by reducing training volume by ,50%, without reducing training
frequency or training intensity.
In line with current best practice [4–6], we defined the peaking
phase as the last 14 days prior to the athletes’ most successful
competition (Olympic/World Championship gold medal), and
compared training patterns in this final training phase to the
penultimate phase beginning 4 weeks prior to the peaking phase
(pre-peaking phase). With regards to training volume, we found
Figure 4. Individual peaking characteristics. Individual (lines) and average (dotted bold line) total weekly training volume during GP, and the
last 6 weeks prior to championship title.
doi:10.1371/journal.pone.0101796.g004
Training and Peaking of Gold Medallists
PLOS ONE | www.plosone.org 8 July 2014 | Volume 9 | Issue 7 | e101796
only a 4 and 15% (NS) decrease in training volume during days -
14 to -8 and days -7 to-1 respectively, compared to the pre-peaking
phase. This is substantially less than the current taper recommen-
dations of a ,50% reduction (Figure 6). It is possible to speculate
as to why these champion athletes chose a strategy very different
from experimentally derived optimum. Bosquet et al [5] reported
no effect on performance if the reduction in training volume was
20% or less. However, there was large individual variation in
peaking behavior in the current study, and no clear patterns
emerged. In fact, 4 of the 11 athletes increased their training
volume during the last seven days. However, existing research has
limitations in terms of narrow focus on one single competition
[32]. In contrast, our results demonstrate that competitions are
frequently integrated into the peaking process in elite sports. The
competition schedule, designed by the International Ski Federa-
tion, is crucial in planning a taper and must be integrated into the
peaking regime. The WC season in these sports typically consists
of two competition days per week over up to 14 weeks with a
maximum of two to four competition free weeks. Such a schedule
may interfere with an optimal tapering process. Rather than
incorporating a single tapering phase, such a schedule may rather
require the athlete to perform repeated ‘‘mini-tapers’’ prior to
each competition.
Figure 5. Peaking characteristics. A: Weekly training time (h) distributed into endurance training (zones 1–5), strength and sprints (bars, y-axis),
and total training frequency (sessions) (line, z-axis) during GP, and during the last 6 weeks prior championship title. B: HIT frequency (sessions)
distributed into zones 3, 4 and 5 (bars, y-axis) during GP, and during the last 6 weeks prior to championship title. There was a statistically significant
difference (P,.05) in total HIT sessions and zones 3 and 5 respectively across GP, pre-peaking phase and peaking phase. Pairwise post-hoc tests
showed: * Difference in total HIT sessions across phases (P,.01). There were no statistically significant differences in zones 3, 4 or 5 across phases.
doi:10.1371/journal.pone.0101796.g005
Training and Peaking of Gold Medallists
PLOS ONE | www.plosone.org 9 July 2014 | Volume 9 | Issue 7 | e101796
Since there was minimal decrease (NS) in overall training
volume during the four-week pre-peaking period, we chose to
compare training performed during the peaking phase to GP,
where weekly training volume was highest. Once the athletes
started their WC season, in either XC or biathlon, their total
training volume was consistently lower than that reported during
GP. Relative to GP, training volume was, respectively, 29 and
35% lower during the penultimate and final weeks before each
athlete’s gold medal race. High competition stress load and
frequent travel may dictate the reduced training volume during
this phase, rather than a predetermined periodization model.
These data indicate that peak training volume for these athletes
was markedly dissociated in time from peak performance by up to
4 months, even accepting individual variations. It is unclear
whether high training volume executed during PP 4–9 months
prior still influences physical capacity during the peaking phase,
following an extended period of reduced training volume where
competitions themselves become a key source of HIT.
Several decades ago, Hickson et al [55] reported that trained
athletes retain most of their physiological and endurance
performance adaptations during 15 subsequent weeks of reduced
training. However, for an Olympic athlete, even a small
performance decrement associated with reduced training could
be the difference between a medal and fourth place. Unfortu-
nately, similar to strength performance, we do not have data for
endurance tests throughout the year. Our objective testing data for
these athletes terminates 3–4 months prior to their gold medal
performances. In elite practice, laboratory testing typically ends
when the competitive season begins. However, in a similar group
of athletes with virtually identical training patterns as in the
current study, Losnegaard et al [17] found that aerobic
physiological adaptations were maintained, and performance
and anaerobic adaptations were even enhanced, several months
after peak training volume.
To our knowledge, no data are available providing mechanistic
links that span such an extended time period. It is possible to
speculate that a prolonged period of high training volume during
PP could favorably alter genomic sensitivity to training during the
season through epigenetic mechanisms [51]. Such cellular level
adaptations to high training volume could be a mechanistic bridge
linking PP training characteristics to training effects several
months later, when high training volumes are precluded by the
competition and travel stress load.
During both the pre-peaking phase and the peaking phase,
virtually all (92%) training was conducted as XC skiing. This shift
towards more specific movement patterns when competition
approaches may explain why peak performance is possible even
after several months with reduced training volume [51–52].
Training frequency. The athletes in the current study
trained, on average, 8–10 sessions
.
week
21
, with no significant
differences in training frequency between the peaking phase and
other phases (Table 2). This finding is in line with current taper
recommendations [4–6]. Nor were there any significant differenc-
es in the number of LIT or HIT sessions from the pre-peaking
Figure 6. Taper comparison. Schematic representation of the actual taper observed in current study compared to recommended volume
reduction. Adapted from Mujika & Padilla [4].
doi:10.1371/journal.pone.0101796.g006
Training and Peaking of Gold Medallists
PLOS ONE | www.plosone.org 10 July 2014 | Volume 9 | Issue 7 | e101796
phase to the peaking phase. However, LIT frequency decreased
from GP (8 sessions
.
week
21
) to the peaking phase (6 sessions
.
-
week
21
), indicating that the observed reduction in total LIT time
was a result of both reduced session frequency and session
duration.
Intensity distribution and rest days. Adaptive stimuli
from HIT sessions appear to be a key component in maintaining
and enhancing physiological and performance adaptations q
during a taper period [36–37,39]. McNeely & Sandler [31]
reported that frequent short HIT bouts .90% _
VVO
2max are more
effective than LIT to enhance endurance performance, and that,
during a taper, steady-state workouts should be replaced by HIT
intervals and short sprints in order to improve performance.
Interestingly, we found that HIT duration did not change
(1.3 h
.
week
21
) during any of the phases. However, HIT frequency
increased from 2 sessions
.
week
21
in GP to 3 sessions
.
week
21
in the
peaking phase (P,.01). In addition, there was a tendency towards
increased sprint training duration from the pre-peaking phase to
the peaking phase. Hence, HIT sessions during the peaking phase
were typically executed more frequently but with shorter duration
than during GP, alongside more frequent bouts of sport-specific
‘‘anaerobic sprint training’’. Examining distribution of training
among intensity zones 3, 4 and 5, we observed a tendency toward
a decrease in zones 3 and 4 and an increase in zone 5 in both
duration and frequency from GP to CP. This suggests that total
HIT duration did not change throughout the year, but that the
actual executed intensity shifted towards a more polarized model
as the major competition approached.
To our knowledge, details regarding best practice models of
HIT patterns and recovery strategies during the final days prior to
peak performance are lacking in the literature. However, the
current data show that short bouts of HIT were performed evenly
throughout the final 14 days (,5 sessions in total per athlete)
(Figure 7). Interestingly, 10 out of 11 athletes performed a HIT
session within 48 h of competition. The exact intensity during
these HIT sessions is somewhat inconsistent, but was typically
above LT
2
. Competitions performed during the final days but not
seen as ‘‘primary events’’ were also integrated into the peaking
strategy. Whether these contribute to a beneficial peaking regime,
or interfere with the optimal strategy is not clear. Eight of 11
athletes in the present study competed in at least one champion-
ship final prior to the event in which they won a gold medal. With
regard to recovery strategies, rest days were typically concentrated
in days 12 to 6. Among all 11 athletes, only 3 athletes took a rest
day during the last 5 days, compared with 14 athlete rest days
taken in the middle period of the peaking phase. That is, rest days
were 3 times more likely to be taken during the middle portion of
the peaking phase (days 12–6) compared with the final 5 days.
However, it is not clear whether this organization of HIT and rest
days during the final 14 days was the result of strategic planning to
optimize performance, or merely coincidental. It has been
previously reported that runners taking a rest every third day
during a six day taper performed worse than those athletes who
trained every day [56], and this topic may be a fruitful area for
future research.
Altitude training. Altitude training is incorporated into the
training of most world-class XC skiers, and is a consistent feature
of Norwegian endurance training. For athletes in the current
study, precise records are not available regarding all days spent at
altitude or the specific altitude at which each training session was
performed. For the last 2–3 decades, 4–6 annual training camps of
14–21 d duration living at 1800–2000 m above sea level and
training at 1200 to 2800 m above sea level, have been integrated
throughout the annual cycle. The aim of these altitude training
camps is to stimulate increased hemoglobin mass, and specifically
acclimate to competition venues located above 1400 m. The
athletes in the current study typically spent 60 to 100 days training
at altitude during the season quantified, although this was likely
somewhat lower for those athletes winning gold prior to 1992. In
addition, where championship events were held at moderate
altitude (e.g. in Salt Lake City, 2002) altitude camps were also an
important feature of the final weeks of training. Based on a
previous study of 29 XC skiers training at altitude [43], objective
data suggest that intensity distribution during altitude camps shifts
towards lower intensity. Training at the highest aerobic intensities
during such camps is essentially absent, unless it is performed at
reduced altitudes. The likely impact of this emphasis on altitude
training was to somewhat reduce the amount of HIT performed
during PP.
Winning an international title in endurance sports clearly
requires outstanding physiological capacity and performance level.
Controlled laboratory trials of world-class elite athletes are
challenging, and training literature based on less well-trained
individuals may be misleading when linking findings to elite
athletes. Our current data outlines unique and accurate day-to-day
training data throughout a season that concluded with each athlete
winning an Olympic or World championship title. Experimental
approaches may in many ways be artificial, while descriptive
training studies allow investigation of elite endurance athletes in a
real-life situation. This may therefore provide a fruitful foundation
from which to generate novel experimental research questions.
We did not find evidence of athletes following the current
tapering recommendations regarding training volume reduction.
However, when comparing training patterns during the peaking
phase to training executed during PP several months earlier, we
found a picture more analogous to that derived from experimental
studies, although the magnitude of training time reduction was still
lower. It is possible to speculate as to whether the medal-winning
performances of these athletes was truly representative of their best
possible performance, or if they could have skied even faster had
they followed recommended tapering strategies specifically for that
one event. On the other hand, the more progressive reduction in
training time from GP to CP observed in the current study,
continued to a lesser degree throughout the CP up until the major
competition, may be the ideal strategy in sports where the
competition schedule is organized as it is in XC skiing and
biathlon. A three month competition phase during which athletes
are typically required to compete once or twice every week,
precludes the application of the recommended tapering strategy
Figure 7. Peaking phase. Number of athletes (of 11) who performed
HIT sessions (line) and number of athletes who took a rest day from
training (bars) during the final 14 days prior to peak performance.
doi:10.1371/journal.pone.0101796.g007
Training and Peaking of Gold Medallists
PLOS ONE | www.plosone.org 11 July 2014 | Volume 9 | Issue 7 | e101796
presented in the research literature. Regardless, the performance
of these athletes was sufficient to beat the rest of the field on the
day, and take home the gold medal.
A central concern in a descriptive study such as this, where
training self-report is the key data source, is whether the data are
accurate and valid. We have recently demonstrated that elite
endurance athletes report their training accurately, although we
found some small discrepancies related to intensity distribution
[57]. We believe the current data represent the same validity as
shown in Sylta et al [57], since both athlete groups used similar
monitoring routines, and some of the athletes are, in fact,
represented in both papers. In addition, athletes recorded their
training on a daily basis, which likely reduced reporting error.
Conclusions
These data show that winning an international title in XC skiing
or biathlon requires a training load of ,800 h/500 sessions
.
-
year
21
, of which ,500 h is executed as sport-specific movement
patterns. Endurance training time for these athletes was distrib-
uted as approximately 90% LIT and 10% HIT, equal to a ,80/
20% SG distribution. Training volume was highest during GP and
decreased progressively during SP and CP. Concurrently, the
proportion of sport-specific training increased markedly. Total
amount of HIT remained stable across all phases, although HIT
training patterns tended to become more polarized in CP.
These athletes did not appear to incorporate a taper in the final
weeks leading up to competition, with training volume, frequency
and intensity remaining unchanged from the pre-peaking phase to
the peaking phase. Hence, we did not observe the recommended
,50% training volume reduction that has been proposed as the
optimal tapering strategy based on previous experimental studies.
However, there was a clear reduction in training volume from GP
to the peaking phase. This reduction was almost entirely due to a
reduction in non-sport-specific LIT with virtually all training
during the pre-peaking phase and the peaking phase composed of
ski training. Only three out of 11 athletes incorporated a rest day
in the final five days leading up to the best athletic performance of
their career, A very large training load during the GP appears to
be an important precondition for exceptional athletic performance
several months later, although exactly how training loads in June-
October are mechanistically connected to performance several
months later remains unclear.
Author Contributions
Conceived and designed the experiments: ET SS. Performed the
experiments: EH IS ET ØS. Analyzed the data: ET ØS SS IS TH.
Contributed reagents/materials/analysis tools: ET EH. Contributed to the
writing of the manuscript: ET EH ØS SS IS TH.
References
1. Tønnessen E (2009) Hvorfor ble de beste best? En casestudie av kvinnelige
verdensenere i orientering, langrenn og langdistanseløp. Oslo: Norges
Idrettshøgskole. 299 p.
2. Seiler S, Tønnessen E (2009) Intervals, Thresholds, and Long Slow Distance: the
Role of Intensity and Duration in Endurance Training. Sportscience 13: 1–27.
3. Seiler S (2010) What is Best Practice for Training Intensity and Duration
Distribution in Endurance Athletes? Int J Sports Physiol Perform 5: 276–291.
4. Mujika I, Padilla S (2003) Scientific bases for precompetition tapering strategies.
Med Sci Sports Exerc 35: 1182–7.
5. Bosquet L, Montpetit J, Arvisais D, Mujika I (2007) Effects of tapering on
performance: a meta-analysis. Med Sci Sports Exerc 39: 1358–65.
6. Le Meur Y, Hausswirth C, Mujik a I (2012) Tapering for competition: A review.
Science & Sports 27: 77–87.
7. Billat V, Demarle A, Slawinski J, Paiva M, Koralsztein JP (2001) Physical and
training characteristics of top-class marathon runners. Med Sci Sports Exerc 33:
2089–2097.
8. Billat V, Lepretre PM, Heugas AM, Laurence MH, Salim D, et al. (2003)
Training and bioenergetic characteristics in elite male and female Kenyan
runners. Med Sci Sports Exerc 35: 297–304.
9. Esteve-Lanao J, San Juan AF, Earnest CP, Foster C, Lucia A (2005) How Do
Endurance Runners Actually Train? Relationship with Competition Perfor-
mance. Med Sci Sports Exerc 37: 496–504.
10. Karp J R (2008) Training characteristics of US Olympic Marathon Trials
qualifiers. New Studies in Athletics 23: 31–37.
11. Ingham SA, Fudge BW, Pringle JS (2012) Training Distribution, Physiological
Profile, and Performance for a Male International 1500-m Runner. Int J Sports
Physiol Perform 7: 193–195.
12. Stellingwerff T (2012) Case study: Nutrition and training periodization in three
elite marathon runners. Int J Sport Nutr Exerc Metab 22: 392–400.
13. Schumacher OY, Mueller P (2002) The 4000-m team pursuit cycling world
record: Theoretical and practical aspects. Med Sci Sports Exerc 34: 1029–1036.
14. Zapico AG, Caldero´n FJ, Benito PJ, Gonza´lez CB, Parisi A, et al. (2007)
Evolution of physiological and haematological parameters with training load in
elite male road cyclists: A longitudinal study. J Sports Med Phys Fitness 47: 191–
196.
15. Seiler KS, Kjerland GØ (2006) Quantifying training intensity distribution in
elite endurance athletes: is there evidence for an ‘‘optimal’’ distribution?
Scand J Med Sci Sports 16: 49–56.
16. Sandbakk O, Holmberg HC, Leirdal S, Ettema G (2011) The physiology of
world-class sprint skiers. Scand J Med Sci Sports 21: 9–16.
17. Losnegard T, Myklebust H, Spencer M, Halle´n J (2013) Seasonal variations in
VO2max, O2-cost, O2-deficit and performance in elite cross-country skiers.
J Strength Cond Res 27: 1780–1790.
18. Gao J (2008) A Study on Pre-Game Training Characteristics of Chinese Elite
Swimmers. J Beijing Sport Univ 31: 832–834.
19. Siewierski M (2010) Volume and structure of training loads of top swimmers in
direct starting preparation phase for main competition. Pol J Sport Tourism 17:
227–232.
20. Fiskerstrand A
ˆ, Seiler KS (2004) Training and performance characteristics
among Norwegian International Rowers 1970–2001. Scand J Med Sci Sports
14: 303–310.
21. Guellich A, Seiler S, Emrich E (2009) Training Methods and Intensity
Distribution of Young World-Class Rowers. Int J Sports Physiol Perform 4:
448–460.
22. Neal CM, Hunter AM, Galloway SD (2011) A 6-month analysis of training-
intensity distribution and physiological adaptation in Ironman triath letes.
J Sports Sci 29: 1515–23.
23. Mujika I (2013) Olympic Preparation of a World-Class Female Triathlete. Int J
Sports Physiol Perform: In press.
24. Hongjun Y, Chen X, Zhu W, Cao C (2012) A Quasi-Experimental Study of
Chinese Top-Level Speed Skaters’ Training Load: Threshold Versus Polarized
Model. Int J Sports Physiol Perform 7: 103–112.
25. Orie J, Hofman N, de Koning JJ, Foster C (2014) Thirty-eight years of training
distribution in olympic speed skaters. Int J Sports Physiol Perform 9: 93–9.
26. Garcia-Pallares J, Garcı
´a-Ferna´ndez M, Sa´nchez-Medina L, Izquierdo M (2010)
Performance changes in world-class kayakers following two different training
periodization models. Eur J Appl Physiol 110: 99–107.
27. Wenger HA, Bell GJ (1986) The Interactions of Intensity, Frequency and
Duration of Exercise Training in Altering Cardiorespiratory Fitness. Sports Med
3: 346–356.
28. Tjelta LI (2013) A Longitudinal Case Study of the Training of the 2012
European 1500 m Track Champion. Int J Appl Sports Sci 25: 11–18.
29. Tjelta LI, Tønnessen E, Enoksen E (2014) A case study of the training of nine
times New York Marathon winner Grete Waitz. Int J Sports Sci Coach 9: 139–
159.
30. Mujika I, Chatard JC, Busso T, Geyssant A, Barale F, et al. (1995) Effects of
training on performance in competitive swimming. Can J Appl Physiol 20: 395–
406.
31. Mujika I, Padilla S, Pyne D (2002) Swimming performance changes during the
final 3 weeks of training leading to the Sydney 2000 Olympic Games. Int J Sports
Med 23: 582–7.
32. McNeely E, Sandler D (2007) Tapering for endurance athletes. Strength Cond J
29: 18–24.
33. Pyne D B, Mujika I, Reilly T (2009) Peaking for optimal performance: Research
limitations and future directions. J Sports Sci 27: 195–202.
34. Thomas L, Busso T (2005) A theoretical study of taper characteristics to
optimize performance. Med Sci Sports Exerc 37: 1615–1621.
35. Mujika I (1998) The influence of training characteristics and tapering on the
adaptation in highly trained individuals: a review. Int J Sports Med 19: 439–46.
36. Hickson RC, Foster C, Pollock ML, Galassi TM, Rich S (1985) Reduced
Training Intensities and Loss of Aerobic Power, Endurance, and Cardiac
Growth. J Appl Physiol 58: 492–499.
Training and Peaking of Gold Medallists
PLOS ONE | www.plosone.org 12 July 2014 | Volume 9 | Issue 7 | e101796
37. Shepley B, MacDougall JD, Cipriano N, Sutton JR, Tarnopolsky MA, et al.
(1992) Physiological-Effects of Tapering in Highly Trained Athletes. J Appl
Physiol 72: 706–711.
38. Houmard JA, Scott BK, Justice CL, Chenier TC (1994) The Effects of Taper on
Performance in Distance Runners. Med Sci Sports Exerc 26: 624–631.
39. Mujika I (2010) Intense training: the key to optimal performance before and
during the taper. Scand J Med Sci Sports 20: 24–31.
40. NSF (1992) Treningsdagbok – Langrenn. Oslo: Norges Skiforbund.
41. NSF (1985) Treningsdagbok - Langrenn. Oslo: Norges Skiforbund.
42. NSSF (1991) Treningsdagbok - Skiskyting. Oslo: Norges Skiskytterforbund.
43. Sylta O, Tonnessen E, Seiler S (2014) From heart-rate data to training
quantification: a comparison of 3 methods of training-intensity analysis.
Int J Sports Physiol Perform 9: 100–7.
44. Lucia A, Hoyos J, Carvajal A, Chicharro JL (1999) Heart rate response to
professional road cycling: the Tour de France. Int J Sports Med 20: 167–172.
45. Sandbakk O, Holmberg HC (2014) A reappraisal of success factors for olympic
cross-country skiing. Int J Sports Physiol Perform 9: 117–121.
46. Rønnestad B, Hansen E, Raastad T (2010) In-season strength maintenance
training increases well-trained cyclists’ performance. Eur J Appl Physiol 110:
1269–1282.
47. Laursen PB (2010) Training for intense exercise performance: high-intensity or
high-volume training? Scand J Med Sci Sports 20: 1–10.
48. Matwejew L (1975) Periodisierung des sportlichen trainings, 2nd ed. Berlin:
Bartels & Wernitz.
49. Lucia A, Hoyos J, Santalla A, Earnest C, Chicharro JL (2003) Tour de France
versus Vuelta a Espana: which is harder? Med Sci Sports Exerc 35: 872–878.
50. Esteve-Lanao J, Foster C, Seiler S, Luc ia A (2007) Impact of training intensity
distribution on performance in endurance athletes. J Strength Cond Res 21:
943–949.
51. Tanaka H (1994) Effects of cross-training – transfer of training effects on VO
2max
between cycling, running and swimming. Sports Med 18: 330–339.
52. Foster C, Hector LL, Welsh R, Schrager M, Green MA, et al. (1995) Effects of
specific versus cross-training on running performance. Eur J Appl Physiol 70:
367–372.
53. Steinacker JM, Lormes W, Lehmann M, Altenburg D (1998) Training of rowers
before world championships. Med Sci Sports Exerc 30: 1158–1163.
54. Bompa T (1999) Periodization: Theory and methodology of training (4
th
ed.).
Champaign, IL: Human Kinetics.
55. Hickson RC, Kanakis C Jr, Davis JR, Moore AM, Rich S (1982) Reduced
Training Duration Effects on Aerobic Power, Endurance, and Cardiac Growth.
J Appl Physiol 53: 225–229.
56. Mujika I, Goya A, Ruiz E, Grijalba A, Santisteban J, Padilla S (2002)
Physiological and performance responses to a 6-day taper in middle-distance
runners: influence of training frequency. Int J Sports Med 23: 367–373.
57. Sylta O, Tonnessen E, Seiler S (2014) Do elite endurance athletes report their
training accurately? Int J Sports Physiol Perform 9: 85–92.
Training and Peaking of Gold Medallists
PLOS ONE | www.plosone.org 13 July 2014 | Volume 9 | Issue 7 | e101796