Win Rate Is Not Enough

Ask anyone about their trading framework and they will lead with win rate.

Seventy percent. Eighty percent. Ninety-two percent on Tuesdays during a full moon.

Win rate is the metric the trading internet has chosen as its universal scoreboard. It is simple, easy to compare, and satisfying to post. There is only one problem with it.

Win rate, by itself, tells you almost nothing about whether a framework actually makes money.

The problem with a single number

Here is a framework with an 80% win rate. It wins eight out of every ten trades. That sounds exceptional.

Now here are the details. It risks 10 points to make 1 point. Every winner earns 1 point. Every loser costs 10. Eight winners produce 8 points. Two losers cost 20. Net result: negative 12 points per ten trades.

An 80% win rate that loses money on every cycle.

Here is another framework. It wins 40% of its trades. That sounds terrible. But it risks 2 points to make 8. Four winners produce 32 points. Six losers cost 12. Net result: positive 20 points per ten trades.

A 40% win rate that generates consistent profit.

Win rate without context is not a metric. It is a marketing number.

I needed something better than that. I needed the full battery.

What the full battery looks like

When I built the research infrastructure behind the Algorithmic Suite, I tested our framework across 18 years of ES futures data — over 89,000 qualifying trade interactions evaluated under 45 different target and stop combinations. The research was not designed to produce a win rate. It was designed to answer a harder question: does this framework have a durable, quantifiable edge after every cost is accounted for?

Answering that question requires more than one number. It requires a battery of metrics, each one testing the edge from a different angle. If any single metric fails, the framework fails.

Here is the battery.

1. Win rate — the starting point, not the finish line

The framework produces a win rate consistently above the breakeven threshold across all tested configurations during Regular Trading Hours.

That is a strong result. But I already showed you how a strong win rate can lose money. So this is where most people stop, and where the real work begins.

Win rate tells you how often you are right. It does not tell you how much you make when you are right, how much you lose when you are wrong, or whether the gap between those two numbers is wide enough to survive real-world costs.

For that, you need the next metric.

2. Profit factor — the ratio that cannot lie

Profit factor is the total gross profit divided by the total gross loss. It is the single most honest number in any trading evaluation.

A profit factor of 1.0 means you are breaking even. Below 1.0, you are losing money regardless of your win rate. Above 1.0, you are making money. The higher the number, the wider the margin.

The Algorithmic framework produces a profit factor well above 1.0 across every tested configuration, with the highest configurations exceeding 3.0.

That means the framework generates multiple dollars in gross profit for every $1.00 in gross loss — across every target/stop combination tested.

This is the metric that most indicator vendors cannot provide. Not because they choose not to. Because they have not built the research infrastructure to compute it. They have win rate screenshots from favorable sessions. They do not have profit factor across 18 years of continuous data.

I wrote about the scale of that research infrastructure in The Research Behind the Algorithmic Suite. The profit factor is one of its outputs.

3. Expected value — what a trade is worth before you take it

Expected value is the most important number in the battery. It tells you the average outcome of a trade before the trade is placed.

The formula is straightforward:

EV = (Win Rate x Average Win) - (Loss Rate x Average Loss)

The raw expected value of the framework is positive across all tested configurations.

That means every time the framework identifies a qualifying setup, the average outcome — across thousands of occurrences over 18 years — is a net gain in points.

But raw EV is not the number that matters. The number that matters is EV after friction.

4. Friction — the tax that never stops

Every real trade carries costs that do not appear in a backtest unless you put them there. Most backtests do not.

I put them there.

For MES (Micro E-mini S&P 500), total friction per round-trip trade is 0.848 points ($4.24). That breaks down into two components:

Commission: $1.74 per round trip (Tradovate MES rates).

Slippage: 1 tick per side, both entry and exit. That is the industry standard assumption for ES/MES — the market moves against you by one tick on each fill. Two ticks total, or 0.50 points.

Combined: 0.348 points (commission) + 0.500 points (slippage) = 0.848 points of friction on every single trade.

After friction, the expected value remains positive across all tested configurations.

That is the real number. Not the theoretical edge. The edge that survives the market's toll booth.

I will write a dedicated post on friction modeling — why most backtests ignore it, how slippage assumptions change everything, and what happens to frameworks that look profitable until you add real costs. (Coming soon: The Friction Tax: Why Most Backtests Lie About Profitability)

5. Breakeven win rate — the line between edge and illusion

This is the metric that puts win rate in context.

Breakeven win rate is the minimum win rate required to produce zero profit — the exact point where the framework stops making money and starts losing it. It is determined by the ratio of risk to reward and the cost of friction.

The breakeven win rate is calculated from the target/stop ratio. After friction, it rises to the breakeven threshold after friction — a number meaningfully below the observed win rate.

That leaves a meaningful cushion above breakeven. Real daylight between the framework's actual performance and the line where the edge disappears.

Is that a wide cushion? It depends on your standard. But it is a measurable, quantified cushion derived from 18 years of data. That is more than most frameworks can show — because most frameworks have never calculated their breakeven win rate in the first place.

I will dedicate a full post to breakeven analysis and what that cushion means under stress. (Coming soon: Breakeven Win Rate: The Metric That Separates Real Edges from Curve Fits)

6. Consistency — profitable days, profitable years

An edge that only works in certain market conditions is not an edge. It is a coincidence that has not yet expired.

The framework was profitable on nearly four out of every five trading days — the vast majority of sessions produced a net positive result.

It was profitable in 18 of 19 years tested (2008 through early 2026).

That span includes the 2008 financial crisis, the 2020 COVID crash, the 2022 rate hike bear market, and every regime in between. The full scope of what those 18 years contain is documented in The Index Futures Research Behind the Algorithmic Suite. The framework was not tuned on favorable periods. It was run across all of them.

7. Survivability — maximum consecutive losses

Win rate tells you how often you win. It does not tell you how bad the losing streaks get.

A framework with a strong win rate will still produce runs of consecutive losses. The question is how long those runs last and whether they are survivable.

The maximum consecutive losing streak across the full dataset was 12 trades.

Twelve losses in a row, out of tens of thousands of trades over 18 years. That is the worst it got.

Whether twelve consecutive losses is tolerable depends on position sizing, which is its own discipline. But the number exists, it is quantified, and it is available — which is more than can be said for most frameworks that advertise their win rate without ever mentioning their worst drawdown.

I will cover maximum consecutive losses, drawdown analysis, and what Monte Carlo simulation reveals about tail risk in a future post.

8. Configuration robustness — how many combinations survive?

A framework that only works with one specific target and one specific stop is fragile. Change either parameter by a point and the edge might vanish. That is a sign of overfitting — the research found a number, not a pattern.

The Algorithmic framework was tested across 45 target/stop combinations (targets from 3 to 15 points, stops from 2 to 6 points).

After applying full friction — commission and slippage on every trade — 33 of those 45 combinations remained profitable.

That is 73% of all tested configurations surviving the full cost burden. The edge is not brittle. It does not depend on one precise setting. It persists across a wide parameter space.

This is one of the strongest signals that the framework captures a real market behavior rather than a statistical artifact. I will write about parameter robustness and walk-forward validation in a dedicated post.

What most vendors cannot tell you

Here is the uncomfortable truth about most indicator products in the futures space.

Ask the vendor for their profit factor. Ask for the expected value per trade after slippage and commission. Ask for the breakeven win rate. Ask how many consecutive losses their framework produces at its worst.

Most cannot answer. Not because they are hiding the numbers. Because the numbers do not exist. The research was never done. The infrastructure to compute these metrics was never built.

What exists is a collection of favorable chart screenshots, a win rate calculated from a handful of sessions, and a price tag.

That is not research. That is not a framework. That is a claim without evidence.

I built the Algorithmic Suite differently because I trade it myself. I needed to know — not believe, not hope, not feel confident — whether the edge was real. The only way to know is to measure it from every angle, across every market condition, with every cost included.

The battery of metrics in this post is the result of that requirement.

The series ahead

This post is the overview. Each metric deserves its own examination, and each will get one.

The series will include:

• The Research Behind the Algorithmic Suite — The data foundation, verification methodology, and permutation scale

• Transaction Costs in Futures Trading — Commission, slippage, and why theoretical edges collapse under real costs

• Drawdown, Consecutive Losses, and Monte Carlo Simulation — Tail risk analysis and survivability testing

• Walk-Forward Validation: How to Test a Framework Without Fooling Yourself — Out-of-sample testing and regime robustness

• Sharpe Ratio and Risk-Adjusted Returns in Futures Research — Why raw returns are not enough either

• MAE and Trade Efficiency: How Much Heat Does Each Trade Take? — Maximum adverse excursion and what it reveals about entry quality

Each post will use the same data, the same methodology, and the same standard: numbers derived from 18 years of continuous futures data, independently verified, with all costs included.

Start with the evidence

The Algorithmic Suite is built on the framework described in this series. Every metric, every verification layer, every cost assumption is available to subscribers through the platform.

If you want decision support built on quantified, verified research rather than curated screenshots and unsubstantiated win rates, start your 7-day free trial.

Disclaimer: Past performance does not guarantee future results. All metrics presented are derived from historical backtesting across ES futures data from January 2008 through March 2026. Trading futures involves substantial risk of loss and is not suitable for all investors. The Algorithmic Suite provides decision support tools for research purposes. It does not provide financial advice, trade recommendations, or signals. Users are solely responsible for their own trading decisions.

Internal Cross-Links

Link Text | Target URL | Status

The Research Behind the Algorithmic Suite | /blog/the-research-behind-the-algorithmic-suite | Published

Transaction Costs | /blog/transaction-costs-futures-trading-analysis | Published

Monte Carlo Simulation | /blog/monte-carlo-simulation-futures-trading | Published

Walk-Forward Testing | /blog/walk-forward-testing-index-futures | Published

Sharpe Ratio | /blog/sharpe-sortino-ratio-futures-trading | Published

MAE and Trade Efficiency | /blog/maximum-adverse-excursion-futures-trading | Published