We designed a laboratory test with variable fixed intensities to simulate cross-country mountain biking and compared this to more commonly used laboratory tests and mountain bike performance. Eight competitive male mountain bikers participated in a cross-country race and subsequently did six performance tests: an individual outdoor time trial on the same course as the race and five laboratory tests. The laboratory tests were as follows: an incremental cycle test to fatigue to determine peak power output; a 26-min variable fixed-intensity protocol using an electronically braked ergometer followed immediately by a 1-km time trial using the cyclist's own bike on an electronically braked roller ergometer; two 52-min variable fixed-intensity protocols each followed by a 1-km time trial; and a 1-km time trial done on its own. Outdoor competition time and outdoor time trial time correlated significantly (r = 0.79, P < 0.05). Both outdoor tests correlated better with peak power output relative to body mass (both r = -0.83, P < 0.05) than absolute peak power output (outdoor competition: r = -0.65; outdoor time trial: r = -0.66; non-significant). Outdoor performance times did not correlate with the laboratory tests. We conclude that cross-country mountain biking is similar to uphill or hilly road cycling. Further research is required to design sport-specific tests to determine the remaining unexplained variance in performance.