||AIM: To describe characteristics of Thoroughbred training stables in Matamata and in all other locations in New Zealand combined, over two 19-month time periods in 1996-1997 and 1998-1999, representing equal length periods immediately prior to and after the construction of a new training surface at the Matamata Racing Club. METHODS: Retrospective records covering all horses training and racing in New Zealand during two 19-month time periods (1996-1997 and 1998-1999), covering 161 locations, were obtained from New Zealand Thoroughbred Racing (NZTR). Outcome variables included whether a horse was raced again in the 6 months following any start in the first 13 months of either time period, number of race starts for every horse, and finishing position. Summary measures with confidence intervals (CI) and unadjusted odds ratios (OR), measuring strength of associations for various factors, were computed. RESULTS: The datasets contained information on 45,446 horses, 11,336 races, 5,110 trials and a total of 110,643 race starts. Horses trained at Matamata represented 8% (3,715) of the total horse datasets, and accounted for 11,977 race starts (10.8%). They were more likely to start in a race or trial in either time period and were 1.4 and 1.3 times as likely to finish first, second or third compared with horses trained at other locations in 1996-1997 and 1998-1999, respectively. A 6-month no-race period occurred for 9,306/12,584 (74%) horses that started at least once in the first 13 months of either time period. Horses trained at Matamata were less likely to have a 6-month no-race period than horses trained at other locations in both time periods. There was no effect of time period within each location on the probability of either a horse having a 6-month no-race period or of a race start being followed by a 6-month no-race period, but there was an overall effect of time and more 6-month no-race periods were observed in 1998-1999 relative to 1996-1997. CONCLUSION: Summary statistics are presented for Thoroughbred racing in New Zealand over two 19-month time periods. Differences between the populations of horses trained in Matamata compared with those trained at other locations were attributed, in part, to the fact that many of the more successful racehorse trainers in the country have stables at Matamata. As a result, the population of horses in Matamata may not be representative of the racehorse population in New Zealand. Although more likely to win or place in both time periods, the magnitude of the advantage to horses in Matamata was reduced in 1998-1999 relative to 1996-1997, and this could be due, in part, to effects of the new track surface at Matamata. There was no evidence of a rise in risk of a 6-month no-race period following any race start in those horses trained in Matamata in 1998-1999 relative to either horses trained at other locations or to horses trained in Matamata during the earlier time period.