What to Actually Measure in Your Trading Behavior (Beyond P&L)

Two positions are open. One is green, one is red. The cursor moves toward the sell button on the green one first — it feels good to close, to lock something in, to have a small win to show for the session. The red position stays open. The thesis, whatever it was, gets a little quieter in the back of the mind, replaced by a simpler idea: it’ll come back.

Nothing about this moment shows up on a P&L statement. The account balance at the end of the day looks the same whether the trader closed the winner in four minutes or four hours, whether the loser sat open for a day or a month. P&L records the outcome. It says nothing about the sequence of decisions that produced it. If you want to see the behavior itself — the actual mechanics of how a trader interacts with open risk — you have to look somewhere else.

Here’s a short tour of what that “somewhere else” actually looks like in trade data.

Holding time

The simplest behavioral metric is also the most revealing: how long does a position stay open, broken out by whether it’s currently above or below entry?

In 1985, Hersh Shefrin and Meir Statman published a paper naming and formalizing something traders had long suspected about themselves: a tendency to sell positions that have appreciated too early and hold positions that have declined for too long. They called it the disposition effect, and the theoretical case for it drew on prospect theory’s asymmetric treatment of losses and appreciation.

Terrance Odean tested this directly in 1998, working through discount-brokerage account records. The finding was concrete: investors realized appreciated positions at roughly 1.5 times the rate at which they realized declining ones. Not a vague inclination — a measurable ratio, sitting right there in the account history.

Holding time, split by position status at exit, is how that ratio becomes visible in an individual’s own data. It doesn’t require a survey or a confession. It’s just two numbers sitting side by side: median time-to-close for winners, median time-to-close for losers. The gap between them is the footprint.

Re-entry window

The second metric is less about a single position and more about the space between two of them: how much time passes between closing a trade and opening the next one, especially right after a loss.

Coval and Shumway looked at this from a different angle in 2005, studying proprietary traders at the Chicago Board of Trade. Traders who were down money in the morning session were about 16% more likely to take above-average risk in the afternoon than traders who were up money in the morning. The pattern wasn’t about a specific trade setup — it was about what happened to risk-taking after an emotionally charged session, independent of the market conditions those traders were actually facing.

A re-entry window metric captures the same shape at a smaller scale: the interval between a losing exit and the next entry, compared against the trader’s own baseline interval. A short window doesn’t announce itself the way a losing streak does. It’s quiet. It just looks like normal trading, faster. The only way to notice the compression is to measure the gap itself, trade after trade, and compare it to what that same trader’s gap usually looks like on an ordinary day.

Frequency

Zoom out further and you get frequency — not the timing of any single re-entry, but the total volume of activity over a period, and how it moves.

Barber and Odean’s 2000 study, “Trading Is Hazardous to Your Wealth,” is the reference point here. Looking at 66,465 households over 1991–1996, they found the average household turned over about 75% of its portfolio annually — and the households that traded the most had an average annual performance of 11.4%, compared to 17.9% for the broad market benchmark over the same period. The paper’s contribution wasn’t a judgment about any individual trade. It was a structural observation: activity itself, measured simply as turnover, tracked with a gap between individual results and the benchmark.

Frequency as a personal metric works the same way — total trades per week, turnover per month, charted over time rather than judged in isolation. A rising frequency trendline doesn’t say anything about whether any particular trade was well-reasoned. It says something about tempo. And tempo, per Barber and Odean’s data, is a variable worth watching in its own right, separate from the reasoning behind any single decision.

Sizing variance

The last metric is the one that gets talked about least, maybe because it’s less dramatic than a fast re-entry after a loss: how much position size varies, trade to trade, relative to a trader’s own typical size.

There isn’t a named study in the canon measuring this the way there is for the disposition effect or turnover, so this one stays qualitative. But the shape of it is intuitive to anyone who’s watched their own order tickets over a few months: size tends to compress during a cautious stretch and expand during a confident one, sometimes tracking account balance, sometimes tracking something less measurable — a recent string of favorable exits, a feeling of being “on.” A sizing-variance chart doesn’t say why the expansion happened. It just shows that it did, and when, relative to the trader’s own historical range.

None of these four metrics — holding time, re-entry window, frequency, sizing variance — tell a trader whether a decision was sound. That’s not what they’re built to do, and it’s not what the research behind them claims either. Shefrin and Statman theorized a pattern. Odean measured a ratio. Coval and Shumway measured a shift in risk appetite after a losing session. Barber and Odean measured turnover against a benchmark. In every case, the finding was descriptive: this is what the data looked like, across a population, over a defined period.

What a trader does with their own version of these numbers — a personal holding-time gap, a personal re-entry window, a personal turnover trend, a personal sizing curve — is a separate question, and not one the data itself answers. The metrics just make the pattern visible. Whether it’s worth acting on, and how, is the part that stays entirely with the person looking at the mirror.