A hit rate on its own tells you nothing

“I’m right about 60% of the time” is the most common self-assessment in forecasting, and on its own it is close to meaningless.

Not because people are lying about it. Because the number has no denominator that matters. Sixty percent accuracy is a strong record at some entry prices and a steady decline at others, and nothing in the figure itself tells you which one you have.

The threshold is the price

An event contract resolves at one dollar or at zero. Buy a share at 30 cents and you are risking 30 to make 70. Buy at 70 and you are risking 70 to make 30.

So the accuracy you need to break even is not some fixed bar like 50%. It is the price itself. At 30 cents you need to be right more than 30% of the time. At 70 cents you need to be right more than 70%.

Put those together and the self-assessment collapses. A 60% record built entirely on contracts bought around 40 cents is a genuine, compounding edge. The same 60% record built on contracts bought around 75 cents is a losing strategy that feels good, because two thirds of the outcomes go your way and the account still shrinks. Commission shifts the threshold slightly further against you in both cases.

This is why an overall hit rate is the wrong statistic to keep. The number worth tracking is your accuracy within each price band, measured against the band’s own break-even.

The per-event cap does not change the threshold

It is reasonable to ask whether Your Call’s structure changes the arithmetic, since maximum realisable profit on any single event is capped at 0.5% of account size while losses are not capped at all.

It does not change the break-even point. If you size each position so that a win lands exactly on the ceiling, then the capital you commit at price p is ceiling × p ÷ (1 − p), and that is also what you lose if the contract goes against you. Work the expectation through and the price cancels out: the break-even accuracy is still the entry price. A cap on the upside and a proportional cap on what you commit move together.

What the cap does change is how much capital each unit of progress costs you, and that is a separate problem.

Where the cap actually bites

Progress toward the 10% Equity Growth Target is measured in events, because no single event can contribute more than half a percent. The ceiling is the same whether you paid 25 cents or 75.

The capital required to reach it is not. At 25 cents you commit roughly a third of the ceiling amount to earn the full ceiling. At 75 cents you commit three times it. That is a nine-fold difference in capital tied up — and in capital exposed — for an identical maximum contribution.

That matters because the drawdown limits are absolute numbers, not proportional ones. Daily drawdown is 3% of account size; maximum drawdown is 6%, trailing a high-water mark that never locks. Your budget for being wrong is fixed. Expensive entries spend it far faster per unit of progress, even when they are correct more often.

The conclusion is not “only trade cheap contracts”. Low prices demand a lower hit rate for a reason — those outcomes usually do not happen. The conclusion is that entry price is a real dimension of the decision, with real consequences for how many errors you can survive, and treating one accuracy figure as your scorecard hides it entirely.

What a usable record looks like

The published requirement gives the scale of the problem. Against a 0.5% cap and a 10% target, twenty events resolving perfectly is the theoretical floor. With entries around the middle of the permitted range, roughly 50 events at a 70% hit rate, roughly 100 at 60%, roughly 200 at 55%. At 50% it is not reachable at all, because a coin flip has no edge to compound.

You cannot tell which row you are on from memory. Memory keeps the forecasts that resolved well and quietly discards the near-misses, and it does not record what you paid. So the record has to be written before the outcome, and it needs five fields:

• The question, in the resolution language rather than your paraphrase
• Your probability, written down before you look at the price
• The price you actually paid
• The outcome
• One line on why, written at entry, not reconstructed afterwards

The third field is what makes the whole thing work. Without it you can compute accuracy but not whether accuracy was enough.

Calibration, briefly

Once a few dozen entries exist, the question stops being “how often was I right” and becomes “when I said 70%, did it happen about 70% of the time”. That is calibration, and it is the property that actually transfers. A forecaster who says 70% and is right 70% of the time can size positions rationally. A forecaster who says 90% and is right 65% of the time cannot, however impressive the raw accuracy sounds.

The standard measure is the Brier score — the average squared distance between your stated probability and the outcome scored as one or zero. Lower is better, and it punishes confident errors much harder than hedged ones, which is the correct incentive. But you do not need the formula to get most of the benefit. Bucketing your forecasts by stated confidence and checking whether each bucket came in near its label will show you the problem.

Writing your probability before you see the price is the part most people skip, and it is the part that makes the record honest. Once you have seen the market, your number is no longer independent of it, and you have lost the only comparison that tells you whether you know something the market does not.

Evaluations are conducted on simulated accounts. Not financial advice.