[TRACK RECORD]

Signal performance, measured

Same Profit Factor math as the screener. 3,236 symbols, 1,038,304 scored events since 1997. Pick a signal below (White Arrow BUY is the default). Judge Profit Factor first; win rate alone misleads on trend rules.

Universe. Liquid US common stocks. excludes ETF/crypto/index/commodity/preferred/CEF/BDC/REIT; price≥$5; dollar ADV≥$1,000,000.

Backtest: today's rules on today's universe (survivorship bias). Price-only and no dividends. Best Combo is net of 20 bps round-trip costs; other signal records do not model costs. Updated 2026-08-29.

Select a signal

BEST COMBO · Best Combo V5 stable-evidence state machine. Confirmed close-final decisions fill at the next session open; holding duration is determined only by evidence, with an immediate CRISIS risk override and no legacy fallback.

Showing 6,290 scored Best Combo events · 1997-2026 (69 still open, excluded)

PROFIT FACTOR
2.76
gross wins / gross losses
AVG RETURN
+5.99%
winsorized to +/- (95/500)%
MEDIAN RETURN
-0.6%
typical trade (often negative on trend rules)
WIN RATE
47.5%
not the score when PF > 1
CLOSED TRADES
6.3k
69 open (excluded)
AVG HOLD
50 bars
trading days

How to read. Trend rules often win about 35-45% of trades. Use profit factor and average return. A negative median with PF > 1 means a few large winners carry the edge.

'97'99'01'0361%'0515632%'0724257%'0940137%'1129860%'1324831%'1527259%'1724450%'1937144%'2130053%'2334450%'2532537%'26132100%50%0%
Grey bars = trades · amber line = win rate. Win rate is descriptive; profit factor is the edge. Dim 50% line is a coin-flip reference, not a target.
EDGE BY MARKET REGIME · color = profit factor (green ≥1)
REGIMEBEST COMBO
UNKNOWN-
BULLPF 2.5349%1.4k
RECOVERYPF 2.8848%4.7k
BEARPF 1.7037%105
CRISISPF 0.2025%83
How the edge holds across market regimes. Color = profit factor.
BY ENTRY YEAR · PF · win rate · median · trades
YEARBEST COMBO
20261.5637%-3.1%132
20253.4750%+0.1%325
20242.3250%-0.0%293
20233.2353%+1.1%344
20221.3940%-2.4%395
20211.6844%-1.4%300
20205.2656%+2.5%401
20192.7950%+0.1%371
20181.1431%-3.5%248
20175.5959%+3.5%244
20166.2860%+3.7%332
20150.7631%-3.1%272
20141.9039%-2.0%293
20135.4460%+3.6%248
20123.4252%+0.7%279
20111.3837%-2.7%298
20103.8352%+1.2%307
20094.0757%+3.2%401
20080.7830%-4.2%218
20070.9832%-2.9%242
20063.3150%+0.1%191
20056.0861%+3.3%156
2004-
2003-
2002-
2001-
2000-
1999-
1998-
1997-
Color = profit factor (green ≥1). Win rate and median for context.
RETURN DISTRIBUTION · % per closed trade · Best Combo
<=-503-50..-3021-30..-20631-20..-101.2k-10..-5896-5..-2508-2..03610..24132..55675..1067810..2038020..3033130..5020050..10057>100median -0.6%
Left of 0 = losers, right = winners. Left skew with PF > 1 is normal for trend rules (few large winners). Tails clipped at -95% / +500%.

[METHODOLOGY]

  • White Arrow BUY. Long round trips on White Arrow direction. Next-open entry and exit. 250-session warmup.
  • White Arrow SELL. Exit-timing score (move avoided until re-entry). Not a short. Often weak alone.
  • Exhaustion. Fixed 21-session horizon after a release.
  • Best Combo V5. One stable-evidence state-machine replay. Qualified point-in-time out-of-fold transitions, net of 20 bps round-trip costs, become published calls; otherwise the model abstains.
  • Regime. Hold BULL/RECOVERY; exit on BEAR/CRISIS (same as screener Regime Profit Factor). Next-open entry and exit, 250-session warmup.
  • Caps. Returns clipped to -95% / +500% before aggregates. Open trades excluded from closed stats.

Full model detail: Methodology · Docs · Signal performance

Research only. Not investment advice. Past backtests do not guarantee future results. Terms.