The Regime Clock — How Long Has This Regime Lasted?

Historical Averages

The mean is significantly higher than the median in both cases because a small number of very long episodes — extended bull or bear markets — drag the average up. Nearly half of all episodes end before day 20. This is why the dashboard shows your position within the historical distribution rather than just comparing to the mean.

Percentile Position

The clock shows what percentage of historical episodes have already ended by this point. If the current risk-on episode is 25 days old, and 54% of historical risk-on episodes ended before day 25, the clock reads "54% of episodes ended."

This is a more honest measure than comparing to the mean. A regime at the 75th percentile has outlasted three quarters of all historical episodes — that is genuinely unusual, regardless of what the average says.

Conditional Flip Probability

The flip percentage shown on the clock answers: given the regime has survived to this point, what is the probability it ends in the next 10 trading days? This is a conditional probability — it accounts for the fact that episodes which have already lasted longer are, in some sense, self-selecting for durability.

From 113 completed episodes per regime type, 1996–2026.

The Progress Bar

The bar fills to show the percentile position — how far through the historical distribution the current episode sits. Markers at the 50th and 75th percentile give reference points. The colour shifts from purple through yellow to orange and red as the episode becomes increasingly unusual by historical standards.

Sticky Regimes

Episodes that survive well past the 90th percentile have proven structural resilience — they have outlasted 9 out of 10 historical episodes. These are typically extended bull markets or prolonged risk-off periods during multi-year deleveraging cycles. The conditional flip probability moderates for these episodes because regimes that have survived this long have already weathered multiple stress events.