The same signal does not mean the same thing in every regime. How Apex defines market regimes, what changed in 2026, and why correlation is the regime's fingerprint.
A regime is a period in which the relationships between assets are stable — until they are not. In one regime Bitcoin trades with technology stocks; in another with gold; in a third with nothing. A signal calibrated in one regime and used in another is not a signal, it is a memory. Apex treats the regime as the first thing to establish and the last thing to trust.
Apex tracks nine instruments on four axes — against equities, against gold, against long bonds, against the real rate — on 60-day and 180-day windows, and expresses each correlation as a percentile of its own history. The pattern of those percentiles is the regime.
Bitcoin's 60-day correlation with gold reached 0.65 — above 99.8 % of readings since 2021 — while the Nasdaq's correlation with gold fell from 0.55 to 0.22. Bitcoin, not the market, had moved. In calm conditions it was being priced as a monetary asset. But the correlation rises monotonically with the depth of an equity drawdown (+0.02, +0.12, +0.14, +0.24 across the bands) and the gold link does not. Across 158 stress days Bitcoin fell 1.94 % on average against gold's 0.12 %. Same asset, two regimes, and the one that matters for a portfolio is the stress one. Calm regime · stress regime.
If regimes are stable, does one market lead another inside them? Apex measured cross-correlations at every lag for 28 instrument pairs on daily data and 132 on minute data. 28 of 28 daily pairs peak at lag zero; so do 92 of 132 minute pairs, and the 40 that do not have correlations under 0.15. Moves are simultaneous. The follower has 0.008–0.045 % left to travel by the time the leader's move is observable — less than the cost of the trade. Research page.
Regimes also govern how many independent bets a portfolio really holds. On calm days the crypto book's instruments correlate at around 0.24; on violent days at 0.89. Eight positions become one position exactly when it hurts. Apex publishes the cluster count on its portfolio section for that reason.
As context and invalidation, not as a trigger. A read of "pressured, not confirmed risk-off" comes with the regime that makes it so — crypto cycle neutral, pressure elevated, correlations high — and the conditions under which the regime, and therefore the read, would change. See macro regimes for the macro layer and scenario analysis for what Apex does with the combination.
All numbers on this page were published in Apex’s own record at the time; they describe the past under stated conditions and promise nothing about the future. Research · Learn · The Apex day