Climatic Influence on Home Advantage in Gulf Region Football

Statistical Analysis using International Match Outcomes

Franck Brocherie

, Olivier Girard

, Abdulaziz Farooq

and Grégoire P. Millet

Institute of Sports Sciences, Department of Physiology, Faculty of Biology and Medicine,

University of Lausanne, Lausanne, Switzerland

ASPETAR, Qatar Orthopaedic and Sports Medicine Hospital, Research and Education Centre, Doha, Qatar

1 OBJECTIVES

While the relevance of time-motion analysis to

determine fatigue occurrence in match-play has

recently been questioned (Carling, 2013), reductions

in match-running performance during football

competition in hot vs. cool conditions highlights the

importance of heat acclimatisation. This is of utmost

consideration for teams playing away matches at

higher temperature and/or humidity (Grantham et

al., 2010), as it can adversely affect players’

thermoregulatory control (Cheuvront et Haymes,

2001; Vihma, 2010). However, no study has yet

focused on the influence of heat stress on football

home advantage at an international level.

Therefore we investigated the impact of climatic

variables on international football results and scores

in the specific context of the World Cup 2022

perspective by applying statistical analyses aiming

to control for factors including the home advantage

and the difference in FIFA ranking between national

squads.

2 METHODS

2.1 Data Collection and Analysis

Relevant information on football match outcomes

and environmental conditions data were extracted

from two websites: (i) the official internet website of

FIFA in order to collect FIFA-recognised Olympic

and A level football results, scores and ranking for

six national teams representing Gulf Cooperation

Council (GCC), and (ii) “Weather Underground”

website centralizing climatic data from weather

stations owned by government agencies referenced

by the World Meteorological Organization (WMO).

It was used to collect both home countries’ and

opponents’ average dry bulb temperature (°C) and

relative humidity (%) of the month preceding the

matches. Temperature and humidity of the day for

all matches for the three different venues (home,

played in GCC; away, played in the opponent’

country; or neutral, played neither at home nor

away) were also granted.

Six variables were defined: (i) the probability of

a favourable outcome (i.e. win or draw vs. loss) (ii)

the difference between the number of goals scored

and the number of goals conceded (ΔGoals), (iii) the

difference in team FIFA-ranking (ΔRank), (iv) the

home advantage along with (v) the temperature (ΔT)

and (vi) the humidity (ΔH) differences between the

home venue of a specific team and that of the match

venue. Two additional variables were used to

determine the best environmental predictor for the

probability of favourable outcome and ΔGoals: (vii)

the heat index (ΔHI) and (viii) the wet bulb globe

temperature (WBGT) (ΔWBGT) difference between

the home venue of a specific team and that of the

match venue.

2.2 Statistics

Generalised linear mixed models with a logit link

function for a binomial residual distribution and a

random intercept for country were developed. The

parameter estimates were reported as odds ratios

(OR) for the favourable outcome or beta coefficient

(β) for the ΔGoals with 95% of confidence interval

(95% CI). For all procedures, a P-value <0.05 was

considered as cut-off for significance.

3 RESULTS

A total of 2008 games over 55 years between 1957

and 2012 were used.

In GCC region, home teams have greater

probability of a favourable outcome (P<0.001) and

higher ΔGoals (P<0.001) than their opponents. After

adjustment for ΔRank, home advantage and ΔH, ΔT

Brocherie F., Girard O., Farooq A. and P. Millet G..

Climatic Inﬂuence on Home Advantage in Gulf Region Football - Statistical Analysis using International Match Outcomes.

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

was clearly identified as a significant explanatory

variable. With every 1°C increase in temperature,

both the probability of favourable outcome

[OR=1.02 (95% CI 1.01; 1.02, P<0.001] and ΔGoals

[β=0.02 (0.01; 0.04), P<0.001] increased. ΔH

appeared to be of lower importance than ΔT in

affecting the favourable outcome OR=1.00 (1.00;

1.01); P>0.05] and ΔGoals [β=0.01 (0.01; 0.01);

P<0.001]. Meanwhile, the probability of favourable

outcome and ΔGoals decreased OR=0.99 (0.98;

0.99) and β=-0.02 (-0.02; -0.02), respectively; both

P<0.001 when playing against a stronger opponent.

To determine the best climatic predictor on the

likelihood of a probability of favourable outcome or

on ΔGoals, the different environmental variables

collected or calculated were compared after

adjustment for ΔRank and home advantage. While

ΔT was significant (P<0.001), adding ΔH was found

to be of lower importance in affecting the match

outcomes (see above). Replacing these two variables

by ΔHI decreased the model fit and appeared non-

significant. Finally, when using ΔWBGT, both the

probability of favourable outcome OR=1.02 (95%

CI 1.01; 1.03); P<0.001 and ΔGoals β=0.03 (95%

CI 0.02; 0.05); P<0.001 increased with a similar

reading than as for ΔT. These relationships were

highlighted by plotting the different climatic

variables against ΔGoals (figure 1).

4 DISCUSSION AND

CONCLUSIONS

Our results showed that the differences in heat stress

conditions between home and away teams

significantly affect the outcome of international

football matches in the GCC region and therefore

represent an integral component of the home

advantage in these hot countries. Hot weather teams

(i.e. GCC), presumably better heat-acclimatised,

tend to have greater probability of favourable

outcome and higher ΔGoals at home with an

increase in ΔT. However, ΔRank plays an important

role in our dataset which tends to disguise the impact

of the environmental conditions.

Finally, our results suggest that temperature and

WBGT approximation, but not humidity or heat

index, are more likely to reflect the impact of

environmental conditions on match performance.

Figure 1: Relationship between temperature (ΔT),

humidity (ΔH), heat index (ΔHI), wet bulb globe

temperature (ΔWBGT) differences and the adjusted

ΔGoals.

REFERENCES

Carling C., 2013. Interpreting Physical Performance in

Professional Soccer Match-Play: Should We be More

Pragmatic in Our Approach? Sports Med [Epub ahead

of print].

Cheuvront S. N., Haymes E. M., 2001. Thermoregulation

and marathon running: biological and environmental

influences. Sports Med 31(10): 743-62.

Grantham J., Cheung S. S., Connes P., Febbraio M. A.,

Gaoua N., Gonzalez-Alonso J. et al., 2010. Current

knowledge on playing football in hot environments.

Scand J Med Sci Sports 20, Suppl 3: 161-7.

Vihma T., 2010. Effects of weather on the performance of

marathon runners. Int J Biometeorol 54(3): 297-306.