Why it exists
A Sharpe ratio of 1 measured over three months means much less than the same value over ten years, and fat-tailed or negatively skewed returns make it even less reliable. The PSR, proposed by Bailey and López de Prado, turns the observed Sharpe into a probability that takes sample length, skewness and kurtosis into account.
How Quant Arena uses them
Above 95% is strong evidence of a positive edge in that window; around 50% means the result is indistinguishable from luck. PSR is the significance component of the Arena Score.
PSR alone ignores that we test many strategies at once: with enough tries, one of them will look good by chance. The Deflated Sharpe Ratio (DSR) raises the bar to the best Sharpe you would expect from luck among all strategies on the same market and window, then asks the same question. Strategy pages show both. In our own data no strategy clears 95% on the DSR over five years, which is a useful reminder of how much of any leaderboard is noise.