Trade Signals · Backtests

The numbers behind the signals.

Every trading day, Reprisa scans a large-cap universe for two rule-based setups. This page publishes the backtest results behind them — rendered directly from the same constants the live alerts quote, with the sample sizes, exit context, and caveats attached. Nothing re-typed, nothing cherry-picked, and one exit family deliberately withheld where it lost money.

Window2016-2026Universelarge-capSetupstwoSourceproduct constants

[01]Methodology — read this first

What these numbers are — and what they are not.

The window

Both setups were validated over the 2016-2026 window on a large-cap US universe — a single decade published whole, spanning bull runs, crashes, and everything between. No favorable sub-window is quoted anywhere on this page.

Base rates are frequencies

A base rate is how often a defined setup resolved a given way over that window — a historical frequency averaged across varied market regimes. It is not a probability for the next trade, and it is not a prediction.

Win rate means nothing without the exit

The same entry quotes very different win rates under different exits — you can raise a win rate and lose money doing it. Every figure below carries its exit context, and the one exit family that backtested to a high win rate with a negative return is suppressed exactly where it failed.

One source of truth

Every statistic on this page is rendered directly from the product’s hard-coded signal constants — the same object the live alerts, briefs, and emails quote. Nothing is re-typed for marketing, so this page cannot drift from what the product says.

These are rule-based simulations over history, not live trading results, and past performance is no guarantee of future results. Historical parallels are context, not predictions. Not financial advice.

[02]Setup · long

Pullback to rising 20-day MA.

A stock in an established uptrend — rising 20-day and rising 200-day moving averages, holding above its 20-day line for most of the past month — dips just far enough for the day’s low to tag the rising 20-day average while the close holds near it. The setup buys that routine dip inside an intact trend. It only fires when the 200-day average is rising: the backtested edge exists in that regime and no other, so the filter is part of the rule, not an option.

DirectionlongPeriod2016-2026Universelarge-capRegime filter200-day MA risingDefault hold10 trading days

Standard tier

Today is the touch day: the low tags the rising 20-day average and the close holds near it, inside the rising-200-day regime.

Win rate · 10-day hold
55.1%
Avg return / trade
+0.59%
Sample size
4,105 trades

Exit for these figures: the 10-day time exit. The tight-take-profit exit styles shown for the high tier are not quoted here by design — see the note under the high tier.

High tier SPY regime + reclaim + quiet volume

Yesterday tagged the 20-day line, today closed back above it (the reclaim), the touch came on quieter-than-average volume, and the broad market (SPY) is in a healthy regime.

Win rate · 10-day hold
56.5%
Avg return / trade
+0.87%
Sample size
866 trades

Exit styles — quoted on the high tier only

TP 1×ATR / SL 2×ATR
71.7% win · +0.34%/trade
TP 1.5×ATR / SL 2×ATR
63.3% win · +0.47%/trade

Why those exit styles stop at the high tier: the identical tight-take-profit exits applied to standard-tier signals backtested to a high win rate with a negative average return — a win rate that loses money. So they are quoted only where they held up, never on standard-tier alerts, and this page follows the same rule the product enforces.

R:R ladder — stop: 1×ATR below entry

2:1 target
37.3% win · +0.12R expectancy
3:1 target
29.7% win · +0.19R expectancy

The honest pattern to notice: as the target widens, the win rate falls — and the expectancy improves anyway. Expectancy is in R, risk units, where one R is the distance from entry to stop. This is why a win rate quoted without its exit is close to meaningless.

Suggested structure — both tiers

Entry
signal-day close
Stop
1×ATR below entry
Target
R:R ladder (see rrLadder) or the 10-day time exit

The structure every alert ships with — levels are computed per signal from the stock’s own volatility (ATR), so the plan scales with the name instead of using fixed percentages.

[03]Setup · short

Extension fade at declining 200MA (short).

A stock that has spent most of the past three months below a declining 200-day moving average rallies hard — stretched far above its 20-day average — straight into the 200-day zone overhead. The setup shorts that extension, targeting a retrace back toward the 20-day line. It is a short against a bear-market rally, not a bottom call.

DirectionshortPeriod2016-2026Universelarge-cap

All signals

Every qualifying extension into the declining 200-day zone over the window.

Win rate
33.9%
Avg return / trade · short side
+0.93%
Profit factor
1.23
Sample size
118 trades

High conviction smaller sample

The same setup with a stricter filter: the close stretched at least 20% above the 20-day average.

Win rate
37.3%
Avg return / trade · short side
+2.44%
Profit factor
1.53
Sample size
59 trades

Flagged “smaller sample” in the constants and in every alert — a tighter filter on fewer trades, so read the improvement with the sample size in view.

Read that win rate for what it is: most of these trades lose. The setup still carried positive expectancy over the window because the average winner outweighed the average loser — that is what a profit factor above one encodes. Per-trade averages are quoted on the short side, so a positive figure is a profitable short.

Suggested structure

Entry
signal-day close, short
Stop
signal-day high + 0.5×ATR
Target
retrace to the 20-day MA

The stop is what validated this setup: in the parameter sweep, the wider stop above the signal-day high beat tight volatility stops in every pairing tested. Same lesson as the ladder above — the exit is the strategy.

[04]Omissions — on purpose

What you won’t find on this page.

  • 01Drawdown or equity-curve figures. The published constants don’t contain them, and this page renders nothing the constants don’t contain.
  • 02A separate average-win / average-loss split. Returns are stored as per-trade averages, so per-trade averages are what you see.
  • 03Tight-take-profit exits on the standard pullback tier. They backtested to a high win rate that loses money there, so the product never quotes them outside the high tier — and neither do we.
  • 04Live trading results. Everything here is a rule-based simulation over history; past performance is no guarantee of future results.

Trade signals are informational, not financial advice. Base rates are historical frequencies across the 2016-2026 window — context, not a prediction. Historical parallels are context, not predictions. Not financial advice.

See them live

The same numbers, attached to a live signal.

On the Pro plan, every alert arrives with its entry, stop, and target levels and the exact base rates above — quoted from the same constants, with the same caveats.