This study examined distances covered at low (1–2ms −2 ), moderate (2–3ms −2 ) and high (>3ms −2 ) acceleration (L ACC , M ACC and H ACC respectively) and deceleration (L DEC , M DEC , and H DEC respectively) during competitive football games. Temporal and transient patterns of acceleration and deceleration were also examined.Observational, repeated measures.Thirty-six professional male professional footballers were monitored using a 10Hz non-differential global positioning system (NdGPS). Match data was organised into six 15min periods (P1: 1–15min, P2: 16–30min, P3: 31–45min, P4: 46–60min, P5: 61–75min, and P6: 76–90min) for analysis of temporal patterns, and into eighteen 5min periods for analysis of transient patterns. ANOVA with Bonferroni post hoc tests were used to identify significant (p<0.05) differences between periods.Distance covered at L ACC , M ACC , H ACC , L DEC , M DEC , and H DEC was 424±75m, 242±25m, 178±38m, 365±54m, 210±23m and 162±29m respectively. Between period decrements ranged from 8.0% to 13.2% from P1 to P3, 9.2% to 16.3% from P4 to P6, and from 14.9% to 21.0% from P1 to P6. Following PEAK H ACC (148% of mean 5min H ACC ), H ACC at 5 POST was 10.4% lower than mean (p<0.01).Time-dependent reductions in distances covered suggest that acceleration and deceleration capability are acutely compromised during match play. Further, the occurrence of transient fatigue may be supported by the findings that H ACC and H DEC performance following PEAK was approximately 10% lower than mean values.