Profitable Stock Trading Strategies: What Actually Survives a Backtest

Profitable stock trading strategies are not the ones with the best story — they are the ones that still make money after commission, slippage and several hundred trades. That gap, between a good-looking chart and a durable edge, is the subject of this page.

Everything below comes from backtests Northtape has published with a certificate attached. You can open any of them at the verification index, or see the full leaderboard on the public track record. None of it is a projection.

What actually makes profitable stock trading strategies work

A strategy is profitable when its edge — the repeatable statistical tilt it exploits — is larger than the cost of harvesting it. Three things decide that, and none of them is the indicator you choose.

1. Edge has to clear costs, not just exist

Every Northtape backtest runs on the same conservative engine: $0.62 per contract on futures, 0.10% per equity trade, one tick of slippage, pyramiding set to zero, calculation on bar close only, and fixed position size. Those settings are deliberately unflattering.

This is where most published "profitable trading strategies" quietly fail. A scalping system with a 60% win rate and an edge of two ticks becomes a losing system the moment one tick of slippage is applied. A strategy that only works at zero cost is not a strategy; it is a chart pattern.

2. Sample size separates an edge from an anecdote

Compare two real Northtape results. Golden Hour returned a 6.71 profit factor on equities daily bars — over 16 trades. Open Strike returned a 2.02 profit factor on ES/NQ 5-minute bars — over 403 trades. The lower number is the more trustworthy one. Sixteen trades cannot tell skill apart from a favourable year.

Treat trade count as a confidence dial, not a footnote. Under roughly 30 trades you have a hypothesis. Over a few hundred, you have something worth risking capital on.

3. Out-of-sample testing, or you are just describing the past

Fitting rules to historical data is trivial. The test that matters is whether the same rules hold on data they were never tuned against. Frontier was built and tested on equities daily bars from 2000 to 2026, then independently re-verified in Python on a separate 10-symbol universe covering 2023 to 2026 — where it produced a 1.29 Sharpe against 1.38 for SPY buy-and-hold on the same window, at well under half the drawdown (−14.7% versus −19.0%).

The profitable stock trading strategies that survive testing

Four families of profitable trading strategies come through honest testing again and again. Each has a different personality, and those differences matter more than the labels do.

Mean reversion — buying short-term panic inside a trend

The classic is Larry Connors' RSI(2): go long when the 2-period RSI closes below 10 while price sits above its 200-day moving average. Backtested across 10 liquid names over roughly 3.6 years of daily bars it produced a 73.4% win rate, a 2.45 profit factor and +1.42% per trade over 406 trades. The full breakdown is in our RSI-2 article.

Northtape's own mean-reversion engine, Reflex, runs the same family across equities daily and NQ 60-minute bars, reaching up to an 80.0% win rate and up to a 6.82 profit factor on its best configuration — though that configuration covers only 10 trades, which is exactly the sample-size trap described above.

Trend and stage analysis — owning the middle of the move

Trend strategies win less often but hold longer. Ascent applies weekly multi-timeframe stage analysis across a 50-name universe, tiering names by strength; at the last run, 28 of 50 names were in live Stage 2. Trend Lock, an intraday trend system on NQ 60-minute bars across 2025–2026, tested at a 72% win rate and a 2.53 profit factor. If the framework is new to you, start with Stage Analysis explained.

Breakout — the workhorse of profitable stock trading strategies

Breakout systems are the least comfortable to trade and often the most durable. Open Strike, Northtape's flagship, traded ES/NQ 5-minute bars from April 2025 to March 2026 for a 52.9% win rate, a 2.02 profit factor and $137,607 net across 403 trades. Frontier, the equities-daily breakout, won only 48.6% of 218 trades and still returned a 2.61 profit factor and +85.6% total return. Inside Job, a GC 5-minute breakout tested from September 2024 to March 2026, produced a 74.5% win rate and a 3.00 profit factor over 47 trades.

Momentum and swing trading strategies — fewer trades, fatter tails

Momentum and swing trading strategies hold for days to weeks and depend on a handful of large winners. Overdrive tested at an 84.6% win rate and a 6.47 profit factor on NQ 15-minute and GC 5-minute data — across 13 trades. The Turn, run on NVDA daily bars from January 2016 to March 2026, produced a 79.3% win rate, a 5.69 profit factor, a 1.32 Sharpe and +1,813% total return over 29 trades, with a −37.5% maximum drawdown. That drawdown is the price of admission, and it is not optional.

Every row above was profitable in backtesting over the period shown, with modelled costs. None of them is a forecast. The full strategy index lists all 15.

Why win rate alone is a misleading metric

Win rate is the number beginners quote and professionals ignore. It tells you how often a strategy is right, not how much it makes when it is.

Frontier won under half its trades — 48.6% — and still returned a 2.61 profit factor, because the winners were far larger than the losers. Meanwhile a system winning 90% of the time while losing ten times its average win is a slow-motion account failure.

Profit factor is gross profit divided by gross loss. Above 1.0 the strategy made money; above roughly 1.5 it made money with room to absorb execution error. Expectancy is the average result per trade — RSI-2's +1.42% across 406 trades is an expectancy statement, and expectancy is the number that scales with size.

How Northtape verifies profitable stock trading strategies

Every strategy on the desk goes through double-confirmation verification before it earns a name.

• Pass 1 — by hand. The strategy is built as Pine Script and run on TradingView's engine with the exact entry, exit and risk settings, checked bar by bar. If it does not hold up visually, it never reaches pass 2.
• Pass 2 — by machine. The same rules are re-implemented independently in Python and re-run over the raw historical data with realistic costs. Two engines, one edge.

If the two passes disagree, the strategy goes back to the bench. Passing strategies get a certificate carrying a SHA-256 hash of the results record, so if a published number ever changes, the hash changes with it. The certificates live at loomiai.io/verify/, and membership terms are on the pricing page.

Honesty is part of the standard. Golden Hour's certificate carries a note stating it did not beat SPY buy-and-hold on risk-adjusted return (0.72 Sharpe versus 1.38); its only advantage was a shallower drawdown, −14.2% against −19.0%. Night Shift is published as "in test — double-verify pending" rather than quietly shelved.

Position sizing: where profitable strategies become profitable accounts

Sizing is arithmetic, not opinion, and it decides more outcomes than entry logic does. All published Northtape tests use fixed size with pyramiding set to zero — never adding to a position — precisely so the reported edge belongs to the strategy and not to leverage.

Two consequences follow. First, a strategy's maximum drawdown is your sizing constraint: The Turn's −37.5% drawdown means a fully allocated trader had to sit through losing more than a third of the account to collect the +1,813%. Second, risking a fixed fraction of equity per trade rather than a fixed dollar amount is what keeps a losing streak survivable. Positive expectancy only compounds if you are still trading when it arrives.

Why a profitable backtest can still lose money

This is the section most sites skip. Three failure modes turn a profitable backtest into a losing account.

Overfitting

Add enough filters and any dataset can be made to look excellent. The tell is a spectacular result on a thin sample — a 6.71 profit factor over 16 trades, or a 6.82 profit factor over 10. Those numbers are real, and they are also the most likely to be curve-fitted. The defence is out-of-sample confirmation plus a trade count that supports the claim.

Regime change

An edge is a description of how a market behaved, and markets change how they behave. Trend Lock and Coil were tested on NQ across 2025–2026 — one specific volatility regime, in which Coil returned a 69% win rate and a 1.82 profit factor. If trend structure compresses, trend strategies bleed while mean reversion improves. Nothing in a backtest tells you which regime tomorrow belongs to.

Survivorship bias

The Turn's headline +1,813% comes from NVDA — a name that happened to be one of the decade's biggest winners, over a window identified by exact trade-set match. Applied to a randomly chosen ticker in 2016, the same rules would not have produced that curve. Whenever a result is tied to a single symbol or a hand-picked period, discount it accordingly.

FAQ