Historical base rates can be incredibly useful. They can also give you a false sense of confidence if you use them carelessly.
Suppose the SPX has closed green on 62% of Mondays following a red Friday over the past five years. That matters. It gives you a historical starting point and may influence how you approach Sunday night or Monday morning.
But it does not mean the market has a 62% chance of closing green this Monday.
That distinction is where a lot of traders get into trouble.
The regime matters
A base rate calculated from 2019 through 2024 blends together several very different markets: a steady bull run, the COVID crash, the recovery, high-inflation volatility, and long stretches of choppy price action.
The resulting percentage tells you what happened across all of those environments combined. The problem is that the current market may look nothing like the average environment in that sample.
If the market is currently in a low-volatility uptrend, data from other low-volatility uptrends is probably more relevant than data from panic-driven selloffs.
The answer is not necessarily to shorten the time window. A shorter window usually means fewer observations, which makes the result less reliable. A better approach is to filter the data according to the current regime.
That might include:
- The current VIX range
- The previous week's direction
- Market breadth
- Recent realized volatility
- Whether the broader trend is bullish, bearish, or sideways
TensaFi's Probability Hub makes these filters available directly, so you can examine the conditional base rate without calculating everything manually.
A base rate is not a prediction
A historical hit rate describes a sample. It does not predict the next outcome with certainty.
The market does not know that a setup worked 62% of the time in the past, and it has no obligation to repeat that behavior.
What the number does give you is context. If you already plan to take a directional position on Sunday night, a 62% historical green rate suggests that the long side has been favored in similar situations.
That is useful information. It is not a guarantee.
Think of a base rate as a prior assumption that should be updated with current information, not as a standalone trading signal.
Sample size matters more than the headline percentage
A 71% hit rate sounds impressive until you realize it came from five wins in seven observations.
That is not a dependable edge. It is a tiny sample that could easily be the result of chance.
As a general rule, samples below 30 observations should be treated cautiously. The result may suggest a direction, but the exact percentage is not especially meaningful.
Once the sample exceeds 100 observations, the reported percentage becomes more useful. Even then, it should still be evaluated alongside the market regime, the size of the average move, and the distribution of outcomes.
A high win rate with tiny average gains and occasional large losses may be less attractive than a lower win rate with better overall expectancy.
Always check what is behind the percentage.
Cherry-picking is the biggest trap
The easiest way to misuse a probability tool is to keep adding filters until the result looks impressive.
With enough combinations, you can almost always find a setup showing an 80%, 90%, or even 100% historical hit rate. Usually, the sample has been reduced to a handful of observations.
At that point, you have not discovered a reliable edge. You have found a pattern that happens to fit past data.
A better process is to choose the filters before viewing the result.
Start with a real hypothesis:
Monday gaps following red Fridays may show directional continuation.
Then test that specific idea.
Do not test dozens of unrelated combinations and keep only the one with the highest percentage. That is data mining, and it creates strategies that often look excellent in hindsight and fail immediately in live trading.
How to use base rates effectively
Base rates are most useful for adjusting conviction and position size, not making the entire decision for you.
A 60% historical base rate may justify slightly more size on a trade you already like. It should not convince you to enter a weak setup that you would otherwise avoid.
A few practical rules help:
- Always read the sample size alongside the percentage.
- Filter for the current market regime before relying on the result.
- Decide what you are testing before running the query.
- Review the average return, not just the win rate.
- Use the result as supporting evidence rather than a complete strategy.
Base rates are a lens, not a crystal ball.
Used carefully, they can add meaningful context to a trading decision. Used carelessly, they simply make a weak idea feel more scientific.