Risk Management and Trading Psychology

Trading Expectancy

Q: What counts as good expectancy?

In R multiples, the benchmarks are clear. Below zero, the strategy loses and ruin is a matter of time. Between 0.2R and 0.5R is marginal profitability requiring great discipline not to turn negative. Between 0.5R and 1R counts as sustainable professional level. Between 1R and 2R is excellent. And above 2R is exceptional and very rare.Bear in mind that after commissions and slippage, retail strategies typically start with a drag of 0.2R to 0.5R that must be overcome just to break even.Realistic objectives: in the first six months, simply clearing zero is enough; in the first year, aim for 0.3R; from the second or third year, 0.5R; and from there on, continuous refinement.

Q: Can a strategy profitable in simulation have negative expectancy live?

Yes, and it happens very frequently. There are seven common causes. Survivorship bias, because simulations ignore the companies that disappeared. Look-ahead bias, using information that was not available at the time. Underestimating slippage, since real execution is always somewhat worse. Commissions modelled poorly. Emotional drift, when the trader modifies the strategy under pressure. Regime change, if the strategy was optimised for a market that no longer exists. And psychological factors, such as holding losers past the stop or closing winners early.The practical recommendation is to paper trade any new strategy for three to six months before committing real capital, then compare the real results with the theoretical ones.

Q: How do I calculate expectancy in options?

The same way as in shares, but measured in premiums. For each trade, the risk is the premium paid on a long position, or the maximum loss on a credit spread. The result is the premium received less the premium paid, including any adjustments.From there you calculate hit rate, miss rate, average gain and average loss, and apply the same formula.In multi-leg structures — iron condors, calendars — record each complete position as a single trade, from opening to close or expiration. Keeping the books leg by leg completely distorts the calculation.And it is worth segmenting by strategy type, because mixing long options with premium selling in one calculation produces an average that describes neither well.

Q: How many trades do I need for a reliable calculation?

The statistical minimum is 30 trades, though at that sample size the 95% confidence interval around expectancy is very wide, on the order of ±40% of the estimate. At 100 trades the interval narrows appreciably, and at 300 or more the calculation becomes robust.One important nuance: expectancy is not stationary. Market regime changes alter it, so a figure calculated across the whole history may not describe the current situation.The professional approach is to calculate a rolling expectancy over the last 50 or 100 trades and watch its trend: a sustained decline signals that the strategy is degrading. Segmenting it by market conditions also helps identify which environments it actually works in.

Q: Can I combine expectancy with the Kelly criterion?

Yes: Kelly is built on it precisely. The formula can be written as f* = (hit rate × R − miss rate) / R, or equivalently as expectancy divided by average gain.The result is the theoretically optimal growth rate, but accompanied by drawdowns that can exceed 50%. Half Kelly reduces both growth and drawdown, and quarter Kelly is more conservative still.A practical warning: if the calculation returns a very large position size, it almost always indicates the estimated edge is unrealistically high and the numbers deserve review.For the retail trader, the sensible approach is a fixed fraction of 1% to 2% regardless of what the formula suggests. Kelly is useful as a theoretical reference, not a prescriptive rule; only professionals with heavily validated strategies work at half Kelly.

ES: Esperanza Matemática (Expectancy) PT: Expectativa Matemática

The definitive metric in professional trading: the average expected gain per trade, combining hit rate, average gain and average loss. Positive expectancy means a profitable strategy over the long run; negative expectancy means mathematically guaranteed ruin.

What Expectancy Is

A strategy’s expectancy is the average gain you can expect per trade, calculated by combining the probability of success with the average size of gains and losses. Its formula is E = (Hit rate × Average gain) − (Miss rate × Average loss).

Positive expectancy means the strategy is profitable over the long run, because every trade has a positive expected profit. Negative expectancy mathematically guarantees eventual ruin, however much skill, patience or capital you have. And zero expectancy amounts to breaking even before costs, which net of commissions and slippage is a loss.

This concept is the compass of professional trading because it settles the false debate between the "high hit rate" and "high reward-to-risk" philosophies: both can produce positive expectancy if their parameters are properly balanced.

Two examples. Strategy A hits 70% with an average gain of $100 and an average loss of $80: E = 0.70 × 100 − 0.30 × 80 = +$46 per trade. Strategy B hits only 30% but gains $400 on average and loses $80: E = 0.30 × 400 − 0.70 × 80 = +$64 per trade. Strategy B is superior despite being right far less often.

Van Tharp’s R-multiple system expresses gains and losses as multiples of the initial risk: if you risk $100, a $200 gain is 2R and a $50 loss is −0.5R. On that scale, the usual reference is that E ≥ 0.5R counts as profitable professional level, E ≥ 1R as excellent and E ≥ 2R as exceptional.

Hit Rate and Reward-to-Risk

The trade-off between these two variables defines each strategy’s character.

Those with a high hit rate and a low ratio — scalping, mean reversion, credit spreads — hit 70% to 80% with ratios of 1 to 0.5 or 1 to 1.2. An iron condor hitting 75% with a 1 to 0.5 ratio — gaining 0.5R and losing 1R — has an expectancy of 0.75 × 0.5 − 0.25 × 1 = 0.125R: positive but modest, and highly sensitive to any deterioration in the hit rate.

Those with a low hit rate and a high ratio — breakouts, trend following, long options — hit 30% to 40% with ratios of 1 to 3 or 1 to 10. Breakouts hitting 35% at a 1 to 4 ratio give 0.35 × 4 − 0.65 = 0.75R, a far more solid expectancy.

Neither is intrinsically better: it depends on the trader’s psychology, the strategy’s characteristics and the market regime. High-hit-rate strategies are psychologically easier and have lower variance, but the losses hurt more when they arrive and demand discipline to cut. High-ratio strategies need fewer trades and build wealth on the big winners, but they force you to endure 60% to 70% losing trades.

The threshold table is useful for calibration. At a 50% hit rate you need a ratio above 1 to 1: at 1 to 1.5 expectancy is +0.25R and at 1 to 2, +0.5R. At 40%, you need better than 1.5 to 1: at 1 to 2 you get +0.2R, at 1 to 3 it rises to +0.6R and at 1 to 5 it reaches +1.4R. At 30%, the threshold is 2.33 to 1: at 1 to 3 you get +0.2R and at 1 to 5, +0.8R. And at a 70% hit rate it works even at ratios of 0.43 to 1: at 1 to 1 it already gives +0.4R.

Practical Application and Tracking

Calculating real expectancy requires rigorous record keeping. For each trade you need to note seven things: entry price and date, stop and target, exit price and date, result in R multiples, amount risked, strategy name, and the reasons for entry documented before opening.

The minimum sample size is 30 trades for the calculation to mean anything, and more than 100 for reasonable confidence. Below that, the result is statistical noise.

Every 30 or 50 trades it is worth recalculating hit rate, average gain, average loss and expectancy, and tracking how they evolve: expectancy trending down signals that the strategy is deteriorating.

Segmented analysis adds a great deal: breaking expectancy down by market regime, by instrument type, by time of day or by day of the week reveals where the strategy works best and worst, and lets you concentrate on the most favourable conditions.

One factor is often forgotten: opportunity cost. Frequency-adjusted expectancy is obtained by multiplying expectancy per trade by the number of trades a year. A +0.5R strategy with only 10 trades a year produces 5R annually; another at +0.2R with 100 trades produces 20R. A lower expectancy per trade can be superior in annual terms, provided the frequency does not blow up commissions or wear you down.

And a warning about bad runs: after a drawdown, the right move is to resume trading normally, not aggressively to win it back. The arithmetic of expectancy is merciless in both directions, and forcing size only accelerates the deterioration.

The Relationship With the Kelly Criterion

The Kelly criterion uses expectancy to calculate optimal position size. Its formula is f* = W − (1 − W) / R, where W is the hit rate and R the ratio between average gain and average loss. At a 60% hit rate and a 1 to 2 ratio: f* = 0.60 − 0.40 / 2 = 0.40, that is, 40% of capital.

That result is mathematically optimal for maximising long-run growth, and simultaneously completely inapplicable in practice: it produces drawdowns above 25% routinely. Which is why the norm is to use half or a quarter of that figure.

Edward Thorp demonstrated it empirically, first counting cards and later at his Princeton-Newport fund, and his real experience confirms that full Kelly generates swings most people cannot stomach psychologically.

The criterion carries three lessons. That you need positive expectancy, because a negative f* simply means do not trade. That expectancy alone is not enough: you also need the hit rate and the ratio. And that size scales with the edge, but variance scales with it too.

One distinction is frequently confused: Kelly’s f* is the fraction wagered, not the "risk per trade" of 1-2% from classical risk management. They only coincide when the maximum loss equals 100% of what is committed, as in a long option.

The professional approach is to calculate full Kelly as a theoretical reference and then apply a fraction of it, integrating it with risk of ruin: a size that produces an acceptable probability of ruin — below 1% — without giving up the effect of compounding.

Expectancy Benchmarks by Level

Expressed in R multiples, which allows comparison across strategies of different risk.

Expectancy (R)	Annual return	Profile
< 0	Negative	Most people before training
0 to 0.2R	0-5%	Experienced retail trader
0.5-1R	20-50%	Established trader
1-2R	50-150%	Elite fund manager
> 2R	>150%	Isolated and very rare cases

Frequently Asked Questions

What counts as good expectancy?

In R multiples, the benchmarks are clear. Below zero, the strategy loses and ruin is a matter of time. Between 0.2R and 0.5R is marginal profitability requiring great discipline not to turn negative. Between 0.5R and 1R counts as sustainable professional level. Between 1R and 2R is excellent. And above 2R is exceptional and very rare.

Bear in mind that after commissions and slippage, retail strategies typically start with a drag of 0.2R to 0.5R that must be overcome just to break even.

Realistic objectives: in the first six months, simply clearing zero is enough; in the first year, aim for 0.3R; from the second or third year, 0.5R; and from there on, continuous refinement.

Can a strategy profitable in simulation have negative expectancy live?

Yes, and it happens very frequently. There are seven common causes. Survivorship bias, because simulations ignore the companies that disappeared. Look-ahead bias, using information that was not available at the time. Underestimating slippage, since real execution is always somewhat worse. Commissions modelled poorly. Emotional drift, when the trader modifies the strategy under pressure. Regime change, if the strategy was optimised for a market that no longer exists. And psychological factors, such as holding losers past the stop or closing winners early.

The practical recommendation is to paper trade any new strategy for three to six months before committing real capital, then compare the real results with the theoretical ones.

How do I calculate expectancy in options?

The same way as in shares, but measured in premiums. For each trade, the risk is the premium paid on a long position, or the maximum loss on a credit spread. The result is the premium received less the premium paid, including any adjustments.

From there you calculate hit rate, miss rate, average gain and average loss, and apply the same formula.

In multi-leg structures — iron condors, calendars — record each complete position as a single trade, from opening to close or expiration. Keeping the books leg by leg completely distorts the calculation.

And it is worth segmenting by strategy type, because mixing long options with premium selling in one calculation produces an average that describes neither well.

How many trades do I need for a reliable calculation?

The statistical minimum is 30 trades, though at that sample size the 95% confidence interval around expectancy is very wide, on the order of ±40% of the estimate. At 100 trades the interval narrows appreciably, and at 300 or more the calculation becomes robust.

One important nuance: expectancy is not stationary. Market regime changes alter it, so a figure calculated across the whole history may not describe the current situation.

The professional approach is to calculate a rolling expectancy over the last 50 or 100 trades and watch its trend: a sustained decline signals that the strategy is degrading. Segmenting it by market conditions also helps identify which environments it actually works in.

Can I combine expectancy with the Kelly criterion?

Yes: Kelly is built on it precisely. The formula can be written as f* = (hit rate × R − miss rate) / R, or equivalently as expectancy divided by average gain.

The result is the theoretically optimal growth rate, but accompanied by drawdowns that can exceed 50%. Half Kelly reduces both growth and drawdown, and quarter Kelly is more conservative still.

A practical warning: if the calculation returns a very large position size, it almost always indicates the estimated edge is unrealistically high and the numbers deserve review.

For the retail trader, the sensible approach is a fixed fraction of 1% to 2% regardless of what the formula suggests. Kelly is useful as a theoretical reference, not a prescriptive rule; only professionals with heavily validated strategies work at half Kelly.