International Journal of Sports Physiology and Performance, 2006:1:137-149

© 2006 Human Kinetics, Inc.

The Yo-Yo Test: Reliability and Association

With a 20-m Shuttle Run and VO2max
Andrew Thomas, Brian Dawson, and Carmel Goodman

Purpose: The purpose of the study was to determine the reliability of yo-yo inter-
mittent recovery test (yo-yo) scores and their degree of association with a 20-m
shuttle run (20MSR) and VO2max values. Methods: Subjects were elite (Australian
Football League [AFL], n = 23), state-level (hockey, n = 15, and cricket, n = 27), and
recreational team-sport players (n = 33). All performed a 20MSR and the yo-yo at
either level 1 (recreational and state level) or level 2 (AFL). A recreational subgroup
(n = 19) also performed a treadmill VO2max test. Results: Test–retest results found
the yo-yo (levels 1 and 2) to be reliable (ICC = .86 to .95). The 20MSR and yo-yo
level 1 scores correlated (P < .01) in the recreational (r = .81 to .83) and state-level
groups (r = .84 to .86), and 20MSR and yo-yo level 2 scores, in the elite (r = .86)
and recreational groups (r = .55 to .57). The VO2max and yo-yo level 1 scores in
the recreational group correlated (P < .01, r = .87), but no association was found
with yo-yo level 2 (r = .40 to .43, nonsignificant). Conclusions: We conclude that
level 1 (recreational and state level) and level 2 (elite) yo-yo scores were both
strongly associated with 20MSR scores and VO2max (level 1: recreational subjects
only). The yo-yo appears to measure aerobic fitness similarly to the 20MSR but
may also be used as a field test of the ability to repeat high-intensity efforts. Key
Words: aerobic fitness, blood lactate, high-intensity running

The 20-m multistage shuttle run (20MSR) is widely used by team-sport athletes
as a field test for determining aerobic fitness,1 and it has also been demonstrated
to have a strong association with VO2max (r = .92 to .93),2,3 although this relation-
ship may be lower in well-trained athletes.4 However, in team sports, exercise is
intermittent and players must repeatedly perform intense, short sprint exercise.5 For
example, in Australian football (AFL), there are between 150 and 200 high-intensity
(fast running and sprinting) efforts, almost all of 1 to 6 seconds in duration.6
These game demands and required fitness are not directly assessed by the
20MSR. Although aerobic fitness is integral to performance in a team game, the
ability to repeat high-intensity efforts is also very important.5,6 Both Krustrup et al5
and Mohr and Bangsbo7 have demonstrated that soccer play quality is associated
with the amount of high-intensity running performed throughout a game. Hence,
it is important to evaluate team-sport athletes’ fitness for intermittent repeated
high-intensity exercise, as well as their aerobic capacity.

The authors are with the School of Human Movement and Exercise Science, University of Western
Australia, Crawley 6009 Western Australia, Australia

137
138 Thomas, Dawson, and Goodman

The yo-yo intermittent recovery test (yo-yo)8 was specifically designed to

evaluate the ability to repeat high-intensity running efforts over short distances. It
involves repeated 20-m shuttle runs at progressively increasing speeds, with a 10-
second active recovery interspersed between every 20-m (out and back) shuttle. This
differs from the 20MSR, which progressively increases in speed but has no recovery
period. Both tests end when the athlete cannot maintain the required speeds. The
yo-yo has 2 levels, with level 1 designed for low-level athletes and level 2 for well-
trained and elite athletes. Level 2 starts at and maintains higher speeds than level
1. By its nature, the yo-yo appears to have more relevance to team sports than the
20MSR. The demands of games such as Australian football,5 basketball,9 netball,10
rugby,11 and soccer1,7,12 all require multiple high-intensity efforts interspersed with
short recovery periods. Therefore, the yo-yo might be a more specific fitness test
for these sports than the 20MSR.
However, a large degree of similarity exists in the procedures of the yo-yo
and the 20MSR, and it remains to be determined whether there is any significant
association between these test scores. If so, then the tests may not be assessing
different elements of aerobic (and/or anaerobic) fitness and only one or the other
need be performed.
Furthermore, only Krustrup et al5 have reported a coefficient of variation
(4.9%) for test–retest reliability for the yo-yo (level 1), and no data are available
for yo-yo level 2 or the degree of association between levels 1 and 2 of the yo-yo
test. Therefore, the main objectives of this study were to examine the degree of
association between the 20MSR and the yo-yo test scores, the relationship to VO2max,
and to assess the test–retest reliability of the 2 levels of the yo-yo test. Secondary
objectives were to determine the degree of association between levels 1 and 2 of
the yo-yo test scores and their relationship to postexercise blood lactate values. The
results of this study should more clearly define the role of the yo-yo for assessing
the fitness of team-sport athletes.

Methods
Subjects
The sport, stage of season when tested, and physical characteristics of the groups
of subjects used are shown in Table 1. The elite and state-level athletes were tested
after 6 to 8 weeks of preseason training and the recreational subjects 2 to 3 months
into their competitive season. Before commencement, all subjects were informed
of the study requirements, benefits, and risks before giving written, informed
consent. The Research Ethics Committee of the University of Western Australia
granted approval for the study.

Experimental Design
Subjects were recruited from different levels of team-sport competition, namely
elite, state level, and recreational. This design allowed for a determination of any
differences in the degree of association between the 20MSR and the yo-yo among
team-sport players of varying competition and fitness levels.
 Yo-Yo Test and Aerobic Fitness 139

All subjects performed both the 20MSR and the yo-yo (either level 1 or level 2)
in random fashion. Elite and state-level subjects performed each test at least once,
after prior familiarization, involving submaximal efforts to approximately 75% of
maximum. The recreational group performed the 20MSR once, either level of the
yo-yo once, and the other level of the yo-yo twice, thereby performing a test–retest
of level 1 and level 2 of the yo-yo for reliability purposes. Our intention was for the
elite-level players to do a test–retest of level 2, but this could not be arranged because
of their training schedule. The outline of testing is shown in Table 2. In addition,
a subgroup (n = 19) of the recreational subjects also performed a treadmill VO2max
test and had postexercise blood lactates taken after the yo-yo (level 1 and level 2)
to provide further information about the aerobic and anaerobic energy contribution
to the yo-yo. Testing sessions were separated by at least 3 days but no longer than 1
week. For every 3 to 4 subjects, there was 1 official determining successful shuttle
completion and final test score. Subjects wore running shoes for all testing, which

Table 1 Subject Characteristics, Mean ± SD (Range)

When
Subjects n Age, y Height, cm Weight, kg tested

Elite
Australian 23 23.5 ± 3.5 178.5 ± 8.5 87.7 ± 8.8 pre-
football (men) (19.5–30.7) (171.0–204.0) (75.3–106.3) season

State level
U17/19 cricket 27 16.7 ± 1.2 180.0 ± 6.6 74.4 ± 10.9 pre-
(men) (14.5–18.7) (165.0–194.0) (53.5–97.0) season

U/21 hockey 15 19.4 ± 0.7 166.7 ± 8.6 67.7 ± 10.1 pre-

(women) (18.4–21.0) (150.5–179.1) (52.2–95.3) season

Recreational*
men 33 24.4 ± 6.0 182.3 ± 5.6 80.9 ± 9.5 in season
(15.2–41.1) (170.8–192.1) (58.5–104.5)
*Sports played include basketball, cricket, Australian football, hockey, rugby, touch rugby, and soccer.

Table 2 Division of Subjects and Testing Procedures

Elite
Group State-Level Groups Recreational Groups

AFL Cricket Hockey

(n = 23) (n = 27) (n = 15) Group 1 (n = 16) Group 2 (n = 17)

Yo-yo Yo-yo Yo-yo Yo-yo level 1 (× 2) Yo-yo level 2 (× 2)

level 2 level 1 level 1 and level 2 (× 1) and level 1 (× 1)
140 Thomas, Dawson, and Goodman

was performed indoors either in a gym (wooden sprung floor) at the University
of Western Australia (recreational and hockey groups) or on an indoor rubberized
surface at club training facilities (Australian Football and cricket players).

General Procedures
Before and after all familiarization or testing sessions, subjects performed a thor-
ough warm-up and cool-down. In both the 20MSR and the yo-yo, each subject ran
in his or her own lane, to assist with recording of test scores. The correct distances
were clearly marked on the floor by electrical tape for both tests. In addition, mark-
ers were placed along the lines for visual recognition. A maximum of 12 subjects
were tested at any 1 time. In both tests, subjects were instructed to give a maximal
effort by completing as many shuttles as possible and were given strong verbal
encouragement to do so. Subjects were removed from each test when they failed
to reach the 20-m line before the audio pacing signal sounded on 2 consecutive
occasions. Their final score recorded was the last level and shuttle number per-
formed before they were removed from the test (ie, the shuttle on which they were
removed was not counted).
All tests were organized so that subjects were tested at the same time of day to
minimize any effects of diurnal variation. Subjects were also asked to refrain from
vigorous physical activity for 24 hours and from consuming food and beverages
(other than water) for 2 hours before each testing session.

Test Methods
20-m Multistage Shuttle-Run Test. The 20MSR required subjects to run back
and forth between 2 lines set 20 m apart. Subjects started running at an initial speed
of 8.5 km/h, which increased by 0.5 km/h every minute. This pace was established
by audio signal emitted from a 20MSR compact disc.13 The procedures for the
20MSR used here were very similar to those reported by Leger and Lambert.14 The
number of completed levels and shuttles (expressed as a percentage of the level)
was recorded at the end of the test.
The Yo-Yo Intermittent Recovery Test. The yo-yo is similar in format to the
20MSR but has a 2 × 5-m, 10-second active recovery at the end of every 2 × 20-m
shuttle run. Subjects ran the 2 × 20-m shuttle (20 m out and 20 m back) at the speed
designated by the Yo-Yo Compact Disc (Helle Thompson, Copenhagen, Denmark).
At the end of each 2 × 20-m shuttle, subjects walked or jogged (as they wished)
2 × 5 m (out and back) and then waited at the starting line for the next signal to
run again. Yo-yo level 1 starts at 10 km/h and level 2 at 13 km/h, with both levels
progressively increasing in speed throughout the test. Finishing speeds can be over
20 km/h, but this depends on the subject’s fitness and fatigue levels experienced
during the test. Yo-yo level 1 usually takes 6 to 20 minutes, and level 2 takes 2 to
10 minutes. The number of completed levels and shuttles and the total distance
covered were recorded at the end of the test.
Maximal Oxygen Consumption (VO2max). Subjects performed a brief warm-up
on the treadmill (a fast walk at 6 km/h for 2 minutes) and then their own stretching
routine before commencing the test. During the test, subjects ran on a motorized
 Yo-Yo Test and Aerobic Fitness 141

treadmill, commencing at a speed of 10 km/h at a 0% gradient, with a 1-km/h

increase in speed every minute until volitional exhaustion. Strong verbal encourage-
ment was provided to all subjects throughout the test. Expired air was continuously
monitored for the analysis of O2 and CO2 concentrations (Ametek Gas Analyzers,
Applied Electrochemistry, SOV S-3A11 and COV CD-3A, Pittsburgh, Pa). Data
were then averaged over 15-second intervals. Ventilation was also recorded every
15 seconds using a turbine ventilometer (Morgan, 225A, Kent, England). The gas
analyzers were calibrated immediately before and verified after each test using
an alpha-verified β-grade gas mixture (BOC Gases, Chatswood, Australia). The
ventilometer was calibrated preexercise using a 1-L syringe in accordance with the
manufacturer’s instructions. The ventilometer and gas analyzers were connected to
an IBM PC that measured and displayed variables every 15 seconds. The sum of
the 4 highest consecutive 15-second values during the treadmill run was recorded
as the subject’s VO2max.
Blood Lactate. At the conclusion of the yo-yo (level 1 or 2), fingertip capillary
blood samples were taken within 1 minute of finishing from selected recreational
subjects. Krustrup et al5 reported that samples taken 1 to 6 minutes after the yo-
yo were not different, therefore only 1 postexercise sample was taken here. The
first drop was discarded and 35 μL of blood was then collected in a heparinized
glass capillary tube (D957G-70-35, Clinitubes, Radiometer, Copenhagen). Lactate
concentration ([La–]) was determined using a benchtop blood-gas analyzer (ABL
625, Radiometer, Copenhagen), which was regularly calibrated using precision
standards and routinely assessed by external quality controls.

Statistical Analysis
Descriptive statistics (mean ± SD) were calculated on all test results, and then both
interclass (Pearson r) and ICC (1-way random model) were determined on selected
variables. Typical error scores were also calculated using the method of Hopkins.15
In addition, independent t tests (with Bonferroni corrections applied) were used to
examine for significance between the elite, state-level, and recreational group test
scores. Significance for all analyses was set at P < .05.

Results
Test–Retest Reliability
Test–retest reliability of the yo-yo performed in the recreational group for level
1 (n = 16) was ICC = .95 (P < .01), and for level 2 (n = 17) was ICC = .86 (P <
.01). These ICC values were the same for both yo-yo score (level:shuttle %) and
total distance covered. Trials 1 and 2 score and distance completed for yo-yo
level 1 were, respectively (mean ± SD), 15.81 ± 1.49 and 16.35 ± 1.61 and 1030
± 463 m and 1195 ± 513 m. For yo-yo level 2, these scores were 19.21 ± 0.68
and 19.28 ± 0.70 and 325 ± 107 m and 339 ± 113 m. Typical error results for
scores were similar for both level 1 (0.26) and level 2 (0.33), with the results for
distance being 107 m (level 1) and 41 m (level 2), respectively. The CV values
for level 1 and 2 scores were 1.9% and 1.3%, respectively, and for distance
completed, 8.7% and 12.7%.
142 Thomas, Dawson, and Goodman

Performance Scores for Elite,

State-Level, and Recreational Subjects
Table 3 presents the performance results for the recreational, state-level, and elite
athletes for the yo-yo, 20MSR, VO2max, and postexercise blood lactate values. Results
for the whole recreational group (n = 33) and VO2max/blood-lactate subgroup (n =
19) were very similar for yo-yo levels 1 and 2 score and distance. There were no
differences between the recreational group scores (15.89 ± 1.40, n = 33 and 15.77
± 1.33, n = 19) and the state-level cricket group yo-yo level 1 scores (15.90 ±
0.89), and similarly for distance completed (1048 ± 446, n = 33, and 1010 ± 419,
n = 19 vs 1049 ± 285 m (t values all <–0.46). State-level female hockey athletes
achieved yo-yo level 1 scores and distances of 15.23 ± 0.93 and 840 ± 280 m, and
these were similarly not significantly different from the recreational group scores
(t values all <1.58). There were, however, significant differences between the rec-
reational subjects and the AFL players yo-yo level 2 scores (19.16 ± 0.80, n = 33
and 19.18 ± 0.71, n = 19 vs 20.83 ± 0.50, respectively; t = –9.261 and –8.327; P <
.001) and distances completed (322 ± 119, n = 33 and 322 ± 110, n = 19 vs 708 ±
157 m; t = –9.888 and –9.030, P < .001). For 20MSR scores, there was no differ-
ence between the recreational (n = 33 and n = 19) and state-level cricket subjects
(11.29 ± 1.80 and 11.26 ± 1.90 vs 10.99 ± 1.16, respectively) and estimated VO2max
(51.1 ± 6.3 and 51.0 ± 6.3 vs 50.1 ± 4.0 mL · kg–1 · min–1, respectively; t values
all <0.68). Similarly, state-level hockey subjects, who achieved a 20MSR score
of 10.60 ± 1.33 and an estimated VO2max of 48.6 ± 4.4 mL · kg–1 · min–1, were not
significantly different from the recreational groups (t values all <1.47). However,
elite AFL players achieved a significantly higher 20MSR score (14.21 ± 0.85, t =
–7.244 and –6.917, P < .001) and hence a higher estimated VO2max (61.2 ± 3.0 mL
· kg–1 · min–1, t = –7.098 and –6.767, P < .001) than both recreational groups. There
was also a significant difference between elite AFL players and both state-level
groups for the 20MSR (t = –11.235 and –9.270, P < .001). Recreational athletes
(n = 19) achieved a VO2max value of 57.4 ± 7.9 mL · kg–1 · min–1 and a maximum
treadmill speed of 17.58 ± 1.77 km/h. Postexercise blood lactates were similar
and not significantly different for level 1 (10.3 ± 3.5 mmol/L) and level 2 (8.1 ±
1.9 mmol/L; t = 1.67).

Correlational Results for Recreational Subjects

Table 4 presents these results for the recreational athletes for the yo-yo, 20MSR,
VO2max, and postexercise blood lactate data. There was a moderate relationship (r
= .60 to .63, P < .01, and r = .48 to .50, P < .05) between yo-yo level 1 and level
2 scores and distance for both recreational groups (n = 19 and n = 33). A strong
association for the 20MSR and estimated VO2max versus yo-yo level 1 score and
distance (r = .81 to .83, P < .01) was recorded for both groups. However, for level
2, this relationship was only moderate (r = .55 to .57, P < .01) for n = 33 and was
not significant (.42 to .44) in the n = 19 recreational group.
Relationships between VO2max and treadmill maximum speed versus yo-yo level
1 scores were also strong (r = .83 to .87, P < .01) in the recreational subgroup (n =
19), but for level 2 these relationships were weak and insignificant (r = .34 to .43).
The associations between VO2max versus treadmill maximum speed, the 20MSR,
Table 3 Performance Results for the Yo-Yo Intermittent Recovery Test, 20-m Multistage Shuttle Run (20MSR),
Maximal Oxygen Uptake (VO2max), Maximum Treadmill Speed, and Postexercise Blood Lactate Measures

Elite (Australian State (hockey), State (cricket),

Recreational, n = 19 Recreational, n = 33 football), n = 23 n = 15 n = 27

Yo-yo level 1 (level: shuttle %) 15.89 ± 1.40 15.77 ± 1.33 — 15.23 ± 0.93 15.90 ± 0.89
Distance (m) 1048 ± 446 1010 ± 419 — 840 ± 280 1049 ± 285
Yo-yo level 2 (level: shuttle %) 19.16 ± 0.80 19.18 ± 0.71 20.83 ± 0.50‡ — —
Distance (m) 322 ± 119 322 ± 110 708 ± 157‡ — —
20MSR (level: shuttle %) 11.29 ± 1.80 11.26 ± 1.90 14.21 ± 0.85‡ 10.60 ± 1.33 10.99 ± 1.16
–1 –1 51.1 ± 6.3 50.96 ± 6.30 61.2 ± 3.0‡ 48.6 ± 4.4 50.1 ± 4.0
Est. VO2max (mL · kg · min )
VO2max (mL · kg–1 · min–1) 57.40 ±7.90 — — — —
Max treadmill speed (km/h) 17.58 ± 1.77 — — — —
Yo-yo level 1 lactate (mmol/L) 10.3 ± 3.5 — — — —
Yo-yo level 2 lactate (mmol/L) 8.1 ± 1.9 — — — —
‡P < .001; significantly different from the recreational (both n = 19 and n = 33) and state group scores.
Yo-Yo Test and Aerobic Fitness
143
144 Thomas, Dawson, and Goodman

Table 4 Correlational Results for the Recreational Subjects

1 2 3 4 5 6 7 8 9 10

1. Yo-yo level 1 score —

n = 19 .48* .81† .81† .87† .83† –.17
n = 33 .61† .83† .83†
2. Yo-yo level 1
distance
n = 19 — .50* .81† .81† .87† .83† –.19
n - 33 .63† .82† .82†
3. Yo-yo level 2 score
n = 19 — .42 .42 .40 .34 .29
n = 33 .55† .55†
4. Yo-yo level 2
distance
n = 19 — .44 .44 .43 .35 .29
n = 33 .57† .57†
5. 20MSR
n = 19 — .78† .93†
n = 33
6. Estimated VO2max
n = 19 — .79† .93†
n = 33
7. VO2max, n = 19 — .75†
8. Max speed, n = 19 —
9. Yo-yo level 1 —
lactate, n = 10
10. Yo-yo level 2 —
lactate, n = 9
*P < .05.
†P < .01.

and estimated VO2max were also strong (r = .75 to .79, P < .01), but even stronger
between treadmill maximum speed versus the 20MSR and estimated VO2max (r =
.93 and .93, P < .01).
The relationship between yo-yo level 1 score/distance and postexercise blood
lactate was not significant, inverse, and weak (r = –.17 and –.19), and level 2 simi-
larly recorded a weak and nonsignificant relationship (r = .29 and .29).
 Yo-Yo Test and Aerobic Fitness 145

Correlational Results for Elite and State-Level Subjects

Table 5 presents these results for the elite and state-level athletes for both levels of
the yo-yo, the 20MSR, and estimated VO2max. Both the state-level hockey (r = .84
and .84, P < .001) and state-level cricket (r = .86 and .86 P < .001) groups recorded
strong associations between the 20MSR and yo-yo level 1 score and distance.
Similarly, elite AFL subjects recorded a strong relationship (r = .86 and .86, P <
.001) between the 20MSR and yo-yo level 2 score and distance. Similar results were
found for 20MSR estimated VO2max scores and yo-yo level 1 and level 2 scores. The
hockey (r = .83 and .82, P < .001) and cricket (r = .86 and .86, P < .001) subjects
recorded strong correlations with yo-yo level 1, whereas the AFL subjects also
recorded a strong association (r = .84 and .85, P < .001) for yo-yo level 2.

Discussion
The major findings in the present study were that both levels 1 and 2 of the yo-yo
were reliable (test–retest), there was only a moderate association between level
1 and 2 scores, and level 1 (recreational and state level athletes) and level 2 (elite
athletes) scores were strongly associated with 20MSR scores, the estimated VO2max,
and also the directly measured VO2max values (recreational subjects only). There-
fore, it appears that the 20MSR and level 1 or 2 of the yo-yo test measure similar
aerobic fitness abilities, but for level 2, this appears to only be the case with highly
trained athletes.

Test–Retest Reliability
The ICC recorded for level 1 (.95) was greater than that for level 2 (.86), but the
CV values for both shuttle and distance scores for level 1 (1.9% and 8.7%) and
level 2 (1.3% and 12.7%) were quite similar. The faster running speeds required

Table 5 Correlational Results for the Elite and State-Level Subjects

Variable Pearson r

20MSR vs yo-yo 1 score .84‡ (hockey), .86‡ (cricket)

vs yo-yo 2 score .86‡ (AFL*)
20MSR vs yo-yo 1 distance .84‡ (hockey), .86‡ (cricket)
vs yo-yo 2 distance .86‡ (AFL)
20MSR estimated VO2max vs yo-yo 1 score .83‡ (hockey), .86‡ (cricket)
vs yo-yo 2 score .84‡ (AFL)

20MSR estimated VO2max vs yo-yo 1 distance .82‡ (hockey), .86‡ (cricket)

vs yo-yo 2 distance .85‡ (AFL)
*AFL indicates Australian Football League.
‡P < .001.
146 Thomas, Dawson, and Goodman

in level 2 might have been a factor in the slightly lower ICC value because a sub-
ject that “misses” the starting audio beep and is therefore slow to start may find it
difficult to catch up and successfully finish the shuttle. There are no other reports
for test–retest reliability on yo-yo level 2 in the literature. As we were unable to
get elite level athletes to perform a test–retest on level 2, and as only recreational
athletes were used for this purpose here, test–retest reliability of yo-yo level 2 using
elite team-sport players should be further investigated.
The CV value for the level 1 shuttle score here is similar to that reported by
Krustrup et al,5 which was 4.9%. They also used a Pearson r (interclass) correla-
tion to check the test–retest reliability of level 1, reporting a value of .98. The
test–retest reliability scores presented here are considered more appropriate, as the
method used in this study (ICC) is a preferred statistical procedure for measuring
test–retest reliability.17

Yo-Yo Association with 20MSR Scores and VO2max

Strong, significant correlations (r > .80) were observed for all recreational and
state-level athletes between yo-yo level 1 shuttle and distance scores versus
20MSR and 20MSR estimated VO2max values. These results suggest that the aerobic
energy system is important in the performance of level 1 of the yo-yo, as recently
reported.5,12
In contrast, weak and nonsignificant relationships were recorded for the
recreational subgroup (n = 19) between yo-yo level 2 shuttle and distance scores
versus 20MSR and 20MSR estimated VO2max values (r < .50). Although the whole
recreational group (n = 33) recorded a moderate relationship for these variables
(r = .55-0.57), only approximately 31% of the shared variance was accounted for,
perhaps reflecting the lesser impact of aerobic fitness on these test scores for these
subjects. These results may indicate that level 2 requires a different energy-system
contribution to level 1, most likely more anaerobic energy, as the starting speeds
are faster.
In contrast to the recreational athletes’ results, strong, significant correlations
(r > .80) were found between yo-yo level 2 shuttle and distance scores and 20MSR
and 20MSR estimated VO2max values for the elite AFL group. Pearson r values cal-
culated here were very similar to those achieved by the recreational and state-level
athletes for yo-yo level 1 scores. The AFL players had significantly higher estimated
VO2max values (based on 20MSR scores) than did all other groups of subjects (61
vs approximately 50 mL · kg–1 · min–1); therefore, yo-yo level 2 may also require
a significant aerobic energy contribution in athletes with higher aerobic fitness
capacities. Krustrup et al5 only reported data on yo-yo level 1, so the aerobic energy
contribution required by elite athletes to perform level 2 awaits further research.
It may be that in less-trained subjects, with lower aerobic fitness levels, level 2
requires greater contributions from the anaerobic energy systems, although here the
blood lactates recorded after level 2 (8.1 ± 1.9 mmol/L) in the recreational subjects
were not greater than after level 1 (10.3 ± 3.5 mmol/L). However, the faster speeds
required in level 2 most likely account for these results, as some subjects could not
physically keep pace with the speeds set down for this level and were therefore
ejected from the test. There are no other reports in the literature on postexercise
blood lactates for yo-yo level 2 with which to compare these results.
 Yo-Yo Test and Aerobic Fitness 147

As professional athletes, the AFL players used here were better trained for
high-intensity performance than were the recreational and state-level athletes. Their
greater aerobic and anaerobic capacities are likely to have prolonged their time
to fatigue, therefore enabling them to perform level 2 longer than the recreational
subjects could, and perhaps with more aerobic energy contribution.
The recreational group correlations for the 20MSR and 20MSR estimated
VO2max versus directly measured VO2max (r = .78 and .79) were slightly lower than
the correlations recorded for yo-yo level 1 shuttle and distance versus VO2max (r =
.87 and .87). The level 1 results here are in broad agreement with those of Krustrup
et al,5 who found level 1 distance to also correlate strongly with VO2max (r = .71).
The slightly stronger correlations recorded here account for an extra 13% of the
shared variance and suggest that yo-yo level 1 might be an acceptable field test
for VO2max estimation. Further research with a greater number of participants of
various fitness levels is required to test this notion further.
The strong correlations found here between the 20MSR and yo-yo level 1
scores suggest that both tests are largely measuring aerobic fitness, with the rela-
tionship found between yo-yo level 1 and VO2max further supporting the notion
that aerobic capacity is important in performing the test. Our results also suggest
that for elite-level team-sport athletes, yo-yo level 2 also largely measures aerobic
fitness, as the correlation with 20MSR scores (r = .86) and degree of shared vari-
ance (73%) were slightly higher than that obtained for Level 1. Therefore, as has
been recently suggested,1 yo-yo level 1 and/or 2 may be a more relevant test of
team-sport-specific aerobic fitness than the 20MSR or other field tests designed
to measure this ability.
The remaining (approximately) 30% of variance unaccounted for by the
20MSR and yo-yo test scores may relate to the ability to repeat high-intensity (but
not maximal) short-duration intermittent efforts, as is characteristic of many team
sports.6,7,11,12 This ability would appear to be measured to a greater degree in the
intermittent yo-yo test than the continuous 20MSR.
To further examine the relationship between the yo-yo test and aerobic fitness,
research that correlates the yo-yo score and distance completed with VO2max across
a range of athletic abilities is required. If conducted on a sufficiently large scale,
the prediction of VO2max from yo-yo scores, as is available with the 20MSR,14,18
might be possible. At present though, if an estimate of VO2max is required, then
the 20MSR should be used, but it should be recognized that differences between
direct and 20MSR estimated VO2max values are common,4,16 as was found here. The
recreational subgroup had an estimated VO2max of 50 mL · kg–1 · min–1, compared to
a directly measured value of 57 mL · kg–1 · min–1. Because of the error surrounding
such predictions, 20MSR results should be expressed as score (levels:shuttles) or
distance completed, rather than estimated VO2max values.12

Yo-Yo Association With Blood Lactates

Krustrup et al5 found postexercise blood lactates for yo-yo level 1 similar to those
found here. They reported a value of 10.1 ± 0.6 mmol/L in 17 male recreational
subjects who had a mean VO2max of 50.5 mL · kg–1 · min–1 and covered approxi-
mately 1800 m. The relationship found here between yo-yo score/distance and
postexercise blood lactate was weak for both level 1 (–.170 and –.193) and level 2
148 Thomas, Dawson, and Goodman

(.293 and .293). Krustrup et al5 reported significant, inverse correlations of –.41 to
–.81 between yo-yo level 1 distance covered and blood lactate measured during the
run, but they did not report on postexercise values and performance, thus making
direct comparisons with our data impossible.
Our results suggest that blood lactate accumulation may not be limiting in the
performance of yo-yo level 1 in less-trained athletes, although further research is
needed to confirm this notion. Krustrup et al5 found that muscle lactate and pH
were the same 1.5 minutes before the end of yo-yo level 1 as they were at exhaus-
tion, suggesting that inability to repeat high-intensity efforts was not caused by an
accumulation of hydrogen ions. Whether this also applies to yo-yo level 2 remains
to be determined.

Practical Applications
As an alternative to the 20MSR, yo-yo level 1 could be used in low-level team-sport
athletes to assess aerobic fitness and the ability to repeatedly perform high-intensity
running, whereas level 2 could be used to test well-trained and elite team-sport
athletes for the same purpose. The intermittent nature and faster running speeds of
the yo-yo test have more relevance to team sports than does the continuous 20MSR,
but if an estimated VO2max is needed, then the 20MSR would still have to be used.
It should be recognized, however, that VO2max estimates from 20MSR scores may
be 10% to 15% in error from true values.
Further research is also necessary to determine the test–retest reliability and
aerobic/anaerobic energy-system contributions in yo-yo level 2, using elite-level
athletes, who should also be tested on level 1 to better indicate the role of each level
in measuring team-sport fitness. In addition, the relationship between the amount of
high-intensity running performed in soccer matches or other team sports and yo-yo
level 2 scores should be investigated (as it has been for level 1). This is necessary
to further establish the validity of the yo-yo for use with team-sport athletes.

Conclusions
In conclusion, the results found here show that yo-yo level 1 and 2 tests are reliable
(test–retest) and that level 1 (in recreational and state-level athletes) and level 2
(elite athletes) yo-yo scores are strongly associated with 20MSR scores and VO2max
(recreational athletes). Therefore, the yo-yo appears to measure aerobic fitness in
a manner similar to that of the 20MSR but might also be used as a more specific
field test of the ability to repeat high-intensity efforts in team-sport athletes.

References
1. Svensson M, Drust B. Testing soccer players. J Sports Sci. 2005;23(6):601-618.
2. Ramsbottom RI, Brewer J, Williams C. A progressive shuttle run test to estimate
maximal oxygen uptake. Br J Sport Med. 1988;22(1):141-144.
3. Paliczka VJ, Nichols AK, Boreham CA. A multi-stage shuttle run as a predictor
of running performance and maximal oxygen uptake in adults. Br J Sport Med.
1987;21(4):163-165.
 Yo-Yo Test and Aerobic Fitness 149

4. St Clair Gibson A, Broomhead S, Lambert MI, Hawley JA. Prediction of maximal

oxygen uptake from a 20-m shuttle run as measured directly in runners and squash
players. J Sport Sci. 1998;16:331-335.
5. Krustrup P, Mohr M, Amstrup T, et al. The Yo-Yo intermittent recovery test: physiologi-
cal response, reliability, and validity. Med Sci Sports Exerc. 2003;35(4):697-705.
6. Dawson B, Hopkinson R, Appleby B, Stewart G, Roberts C. Player movement patterns
and games activities in the Australian Football League. J Sci Med Sport. 2004;7(3):278-
291.
7. Mohr M, Bangsbo J. Development of fatigue towards the end of a high level soccer
match [abstract]. Med Sci Sports Exerc. 2001;33(5 suppl 1):S160.
8. Bangsbo J. Fitness Training in Football—A Scientific Approach. Bagsvaerd, Denmark:
HO+Storm; 1994.
9. McInnes SE, Carlson JS, Jones CJ, McKenna MJ. The physiological load imposed on
basketball players during competition. J Sports Sci. 1995;13:387-397.
10. Steele J, Chad K. An analysis of the movement patterns of netball players during match
play. Sports Coach. 1992;15(1):21-28.
11. Docherty D, Wenger HA, Neary P. Time motion analysis related to the physiological
demands of rugby. J Hum Mov Stud. 1988;14:269-277.
12. Stolen T, Chamari K, Castagna D, Wisloff U. Physiology of soccer. Sports Med.
2005;35(6):501-536.
13. Twenty Metre Multistage Shuttle Run [CD-ROM]. Canberra, Australia: Australian
Sports Commission; 1999.
14. Leger L, Lambert J. A maximal multistage 20-m shuttle run test to predict VO2max. Eur
J Appl Physiol. 1982;49(1):1-12.
15. Hopkins W. A new view of statistics. Available at: http://www.sportsci.org/resource/
stats/index.html. Accessed August 20, 2004.
16. Sproule J, Kunalan C, McNeill M, Wright H. Validity of 20MSR for Predicting VO2max
of adult Singaporean athletes. Br J Sport Med. 1993;27(3):202-204.
17. Coakes SJ, Steed LG. SPSS: Analysis Without Anguish: Version 11.0 for Windows 2003.
Milton, Qld: John Wiley & Sons Australia Ltd; 2001.
18. Leger L, Mercier D, Gadoury C, Lambert L. The multistage 20 metre shuttle run test
for aerobic fitness. J Sports Sci. 1988;6:93-101.