Evaluating double-bottom breakouts as a testable system

From a two-trough sketch to one sequence

A double-top-bottom is a two-trough or two-peak price structure that becomes a long or short hypothesis only after a defined breakout through the intervening extreme. The archive studied the long case as a breakout-system: a complete entry, exit, and abstention procedure that turns the chart condition into one testable sequence rather than a discretionary sketch.

Editorial reading: lock the pattern rules, vary one exit or structure choice at a time, and judge the setup by whether the same procedure still earns its keep after those changes.

How the double bottom was coded

A coded double-bottom required the structure to last no more than 90 trading days, placed the intervening high at least five days before the breakout, and limited the vertical gap between troughs to 10% of the pattern height.

The tested long signal was a daily close above the neckline, with entry on the next open and a stop a few cents under the lower trough. The baseline take-profit was the measure-rule: a take-profit placed one pattern height above the breakout price, used as the baseline reward target against which other exit multiples are compared.

The procedure traded only when the S&P 500 was treated as bullish. It skipped names under $5 or over $500 and names whose breakout-day volume was below 100,000. Commissions were ignored.

The same rules on two universes

On the January 2000 through mid-2017 window, the same S&P 500 rules produced a higher profit factor, win rate, average trade, win/loss ratio, and Kelly value, plus a smaller maximum drawdown, than the Russell 3000 run. The broader universe still posted a larger total dollar profit from more trades.

Profit factor here means gross profit divided by gross loss. It is used as a reliability score for the whole procedure, not as a forecast of future return. The Kelly criterion is a sizing filter that combines the share of winning trades with the average win-to-loss ratio so a procedure can be ranked before capital is committed.

Holding time after entry was nearly the same in both universes, about 32 to 33 days, so those universe-level differences were not explained by a different typical holding period.

What changed when one rule moved

Patterns with the left trough above the right trough outperformed the opposite height order. Patterns whose right trough carried more volume than the left trough also scored better on the same reliability measures.

Raising the take-profit to two pattern heights while cutting losses just under the neckline left a low Kelly value, so that exit pair was not presented as a recommended combination.

Splitting patterns into small, medium, and large height bands relative to price produced no monotonic size effect. Most detections were small because the lookback was 90 days, and the largest band had too few trades for a firm ranking.

Editorial reading: universe choice, trough order, and the exit pair all moved the reliability scores, while holding time did not. The measure-rule remains the baseline against which those other exits were compared, not a claim about what a future double bottom will pay.