Can a well-calibrated prediction strategy still lose money?

Yes. A trader whose forecasts always equal the market price can be perfectly calibrated, if the market is, and still lose exactly the fees and spreads on every trade. Accuracy that matches the market earns nothing; only accuracy beyond the price, net of costs, pays.

How large an edge is needed?

Larger than it looks. If a trader correctly estimates 70% and buys at a 68-cent ask, the expected profit on 100 contracts is $70.00 - $68.00 - $1.53 Kalshi fee = $0.47, under half a cent per contract. A one-cent move in the fill or a slightly wider spread removes it.

How should a strategy be evaluated?

On both measures. Score the forecasts with a Brier or skill score against the market price, and separately compute profit at the actual entry prices after all costs. AltStreet's Kalshi study warns that edges measured against stale or placeholder quotes are 'an artifact of the sampling schedule,' not tradable profits.

Can being right 94% of the time lose money?

Yes. A trader who buys favorites at 95 cents and wins 94% of the time collects $94 per 100 contracts for $95 plus a $0.34 Kalshi fee: a loss of $1.34. A high win rate says nothing about profit until it is compared with the price paid.

Why do fees matter more on cheap contracts?

Because the fee is a larger share of the price. AltStreet's Kalshi study (July 2026) calculated the standard fee at 6.7% of a five-cent contract versus 0.4% of a ninety-five-cent one, and noted that rounding up imposes a one-cent-per-order floor that weighs on small low-price orders.

Did AltStreet find a profitable edge on Kalshi?

No. AltStreet's Kalshi review (August 2026) reports that the study of 15,685 markets found no systematically exploitable edge after fees: four apparent edges all proved to be sampling artifacts, three of them traced to stale or placeholder quotes and one that faded as the sample grew.

Accuracy vs Profitability

Prediction Markets & Event Contracts

Definition

Accuracy versus profitability is the gap between forecasting well and trading profitably. Accuracy, measured by calibration and Brier scores, compares a forecaster's probabilities with outcomes. Profitability compares those probabilities with the prices actually paid, after fees, spreads, slippage and capital lock-up. A trader profits only when their probability beats the executable price by more than the costs.

Why it matters

Good forecasting is necessary but not sufficient. Profit is the gap between your probability and the price you actually pay, minus fees, spreads, slippage and the cost of tied-up capital, and that gap is usually small. Evaluate your own or someone else's results on profit at real entry prices after costs, not on win rate or accuracy alone. Treat backtests against midpoints, last trades or old quotes with suspicion, since AltStreet's own study saw apparent edges vanish under better data. And remember that on a zero-sum venue, a strategy's profit has to come from other traders.

Common misconceptions

•Calibrated forecasts are not profitable by default, because a forecaster who matches a calibrated market price pays fees and spreads with no offsetting edge.
•Being right more often than not can still lose money at the prices paid, because a 94% win rate on contracts bought at 95 cents loses before and after fees.
•Backtests against stale prices overstate profits, because placeholder quotes far from the event were never available to trade at size.
•A small edge is not a small profit at scale, because fees, a one-cent worse fill or a wider spread can remove an edge of half a cent per contract.
•Accuracy and profit are measured against different things, because accuracy compares forecasts with outcomes while profit compares them with executable prices after costs.

Technical details

Worked comparison (illustrative)

Trader A forecasts exactly the market price on every trade: same Brier score as the market; profit equals minus fees and spreads.

Trader B is 2 points more accurate than the market at 68 cents: about +$0.47 per 100 contracts after Kalshi fees, before spread and slippage.

Trader C is 2 points less accurate: about -$3.53 per 100 contracts after fees.

Worked example: high accuracy, negative profit (illustrative)

Buy 100 favorites at $0.95; Kalshi standard fee $0.34.

Win rate 94%: expected payout $94.00.

Expected result: $94.00 - $95.00 - $0.34 = -$1.34 per 100 contracts.

Buy 100 longshots at $0.05 that win 5% of the time: expected payout $5.00, fee $0.34, result -$0.34, a fee of about 7% of the stake.

What AltStreet's reviews and filings show

Kalshi study (July 23, 2026): four apparent edges, including a 33-point gap at 46.4 cents implied and a series at z = +3.90 across 110 contracts, all failed validation; the second's quotes were captured 4,199 minutes before settlement.

Kalshi study: median quoted spreads across 1,873 liquid contracts were 1.0 cent for deep longshots, 3.0 cents at 10 to 34 cents and 1.0 cent for heavy favorites; with spreads that tight, fees are the main cost.

Kalshi study: standard fee 6.7% of a five-cent contract versus 0.4% of a ninety-five-cent one.

Kalshi review (August 2026): the realistic expectation for a directional trader without a durable edge is a negative return after costs.

How investors apply the term

Investors usually apply this term by tying the definition back to the controlling documents and then asking which cash flows, rights, restrictions, or valuation assumptions it changes. A useful review looks at the term alongside fees, leverage, liquidity, reporting frequency, and manager discretion rather than treating it as a standalone concept.

The practical diligence test is simple: identify who benefits if the assumption is wrong, who has control when conditions change, and what evidence the investor receives after closing. Terms that cannot be reconciled to documents, cash movements, and monitoring data should be treated as marketing shorthand until proven otherwise.

Frequently asked questions

Can a well-calibrated prediction strategy still lose money?: Yes. A trader whose forecasts always equal the market price can be perfectly calibrated, if the market is, and still lose exactly the fees and spreads on every trade. Accuracy that matches the market earns nothing; only accuracy beyond the price, net of costs, pays.
How large an edge is needed?: Larger than it looks. If a trader correctly estimates 70% and buys at a 68-cent ask, the expected profit on 100 contracts is $70.00 - $68.00 - $1.53 Kalshi fee = $0.47, under half a cent per contract. A one-cent move in the fill or a slightly wider spread removes it.
How should a strategy be evaluated?: On both measures. Score the forecasts with a Brier or skill score against the market price, and separately compute profit at the actual entry prices after all costs. AltStreet's Kalshi study warns that edges measured against stale or placeholder quotes are 'an artifact of the sampling schedule,' not tradable profits.
Can being right 94% of the time lose money?: Yes. A trader who buys favorites at 95 cents and wins 94% of the time collects $94 per 100 contracts for $95 plus a $0.34 Kalshi fee: a loss of $1.34. A high win rate says nothing about profit until it is compared with the price paid.
Why do fees matter more on cheap contracts?: Because the fee is a larger share of the price. AltStreet's Kalshi study (July 2026) calculated the standard fee at 6.7% of a five-cent contract versus 0.4% of a ninety-five-cent one, and noted that rounding up imposes a one-cent-per-order floor that weighs on small low-price orders.
Did AltStreet find a profitable edge on Kalshi?: No. AltStreet's Kalshi review (August 2026) reports that the study of 15,685 markets found no systematically exploitable edge after fees: four apparent edges all proved to be sampling artifacts, three of them traced to stale or placeholder quotes and one that faded as the sample grew.