Liminal / Relative Value
Dislocation Scanner
Thesis
KRE–XLF divergence at a 6-week extreme while HY spreads sit near tights
The scanner ranks cross-asset pairs by how far their current relationship sits from its trailing norm, and reports the ranking itself as the finding. This week's board, in order: regionals against broad financials, running roughly twice the stretch of the gold/real-yield pair and three times the BTC–Nasdaq pair, and the only one where the confirming market is actively pricing the other direction. Dislocations either mean-revert or regime-change; the board's job is to say which side each pair is on, not to re-argue the desks' own cases.
Tape
Evidence
Ranked by z-score against its trailing relationship, KRE–XLF sits furthest from norm of the three tracked pairs: about twice the gold/real-yield stretch, three times the BTC–NDX gap. It is also the only pair where the disagreeing market — liquid credit, with HYG at the year's tights — would have to reprice for the gap to close upward instead of down.
The gold/real-yield break carries regime-change characteristics, because sustained central-bank flow can hold it there structurally. The KRE/XLF spread looks mean-reverting on the history, where funding-cost repricings have resolved within a quarter. Sorting pairs into those two buckets is the entire edge of the scanner.
Market Read
Liquid credit isn't quiet by accident: a market that believed the regional-funding story would show it in spread levels, and it hasn't. That non-confirmation is a standing judgment already priced into credit, which means the edge here sits specifically in equities, not in the credit complex the scanner also watches.
The Scanner's Selection Bias
Extreme-spread scanning has a survivorship problem. Pairs that re-converge quietly never make the board, so the scanner over-samples relationships that were breaking for a durable reason, and the mean-reversion base rate ends up inflated by the dislocations nobody ever saw.