Options trading journal: win rate vs expectancy
Why a high win rate says little about an option seller's results, what to record for each trade, and how to read the journal once it has enough rows.
Win rate is the wrong headline
Option sellers win most of the time by construction. A 25-delta short put expires worthless most of the time whether or not it is a good trade. What decides the result is expectancy: the average win times the win rate, minus the average loss times the loss rate.
In our paper book of cash-secured puts the win rate was 85.6%, the average win about +1.4% of collateral and the average loss about -2.8%, a ratio of 2 to 1. That nets to about +0.8% a trade. Change the loss ratio to 7 to 1, the shape of a credit spread at its worst, and the same win rate is not enough.
Early exits raise the win rate and can lower the result
Taking profits early converts some trades that would have lost later into wins. In a replay of put credit spread management rules, closing at 25% of the credit lifted the win rate to about 97% while lowering the dollars earned in every cell compared with holding to expiry. A journal that tracks only wins and losses would call that an improvement.
Track return per trade on collateral or on the dollars at risk, and compare rules on that, never on win rate alone.
What to record
For each trade: the date, the underlying and the factor it belongs to, the structure, expiry, strikes, the fill price against the bid and ask at entry, the collateral or maximum loss, the reason for the trade, the exit and why, and the result as a percent of collateral. Note whether the trade followed your written rule or was a one-off.
That last field matters. In our own records, the largest losses came from trades placed outside the tested rules and at sizes the research never covered. A journal that separates rule trades from discretionary ones shows where the money actually goes.
Reading it
Group trades that share a market draw. Ten puts that expired the same week are close to one observation. Report the average with an interval clustered by week or month, and wait for enough independent weeks before deciding anything. One bad week moved every strategy we were tracking from a pass to a hold, which says the sample was thin, not that something broke.
Also compare against something simple: the same collateral in Treasury bills, or the plain unfiltered version of your trade. The guide on backtesting options honestly covers why the lower bound, not the mean, is the number to act on.
Historical replays are not a forecast. Options can lose more than the premium collected. This is education, not a recommendation to trade anything.