TABLE 2.

Relationships between selected variables and match outcomes among elite German soccer players.

Parameters	Close games (≤1)
	F	WM	CM	FB	CD
Total distance covered (increase by 0.1 km)		55.94	+*
Distance covered below 11 km/h (increase by 0.1 km)			+*	+*
Distance covered below 14.4 km/h (increase by 0.1 km)	+*
Distance covered at 14–17 km/h (increase by 1 %)			8.81
Distance covered at speeds >24 km/h (increase by 0.1 km)		60.71
Distance covered at speeds >24 km/h (increase by 1 %)	7.99
Distance covered at speeds >24 km/h - seasons (increase by 1%)	4.08
Number of sprints (unit increase)	1.41
Mean running speed (increase by 0.1 km/h)			–*
Total number of shots (unit increase)	11.44
Total number of shots - seasons (unit increase)	2.72
Number of passes (unit increase)		0.41	–0.41
Number of crosses (unit increase)		–0.54	–9.16	–0.28
Number of one-on-one plays (unit increase)	0.88
Playing at home	51.75	41.99	91.34	49.34	59.93

Results marked with an asterisk (“^∗”) were identified by visual inspection as unreliable in terms of values for the coefficient. These results have a very large confidence interval. This means that even though the model selected the variable as a contributor to the match outcome, more-precise estimate cannot be attained, but it can merely point out the direction of the variable’s effect as helpful (positive) or harmful (negative) for the match outcome.