How the record is calculated

When a creator opens a call, the platform records the live delayed quote at that moment. When it settles — either closed by hand or auto-settled at expiry — the platform records the quote again. A creator can choose what to call and when, but never what price it opened or closed at.

(exit − entry) ÷ entry for a long, and (entry − exit) ÷ entry for a short.
⚠ A short has no floor: if the price more than doubles against it, the loss is worse than −100%. Those are real results and they are shown as they happened.

Option returns are the change in a modelled Black-Scholes premium, not an executed fill. Anywhere an option record appears it carries a “modeled” tag, and option records are kept in a separate tab from shares rather than blended with them.

The record page covers calls that settled in the last 90 days. A creator’s own page shows a shorter 30-day view.

A plain average treats a call from ten weeks ago as today’s news. Each closed call is instead weighted 0.5 ^ (age in days ÷ 21), so a call that settled today counts fully, one from 21 days ago counts half, and one from 42 days ago counts a quarter. The figure shown is the weighted mean.

The sample standard deviation of that creator’s individual call returns — how far their results spread out around their own average. Two creators can average the same return while one earned it in small repeated steps and the other by alternating large gains and large losses. This is the number that separates them.

The worst peak-to-trough fall on the compounded run of closed calls.
⚠ It is measured across the sequence of settled calls, not across calendar time — the curve only moves when something settles. A −100% drawdown means a single call was large enough to wipe out the simulated starting bankroll; the curve floors at zero rather than going negative.

Gross winnings divided by gross losses. Above 1 means the wins outweigh the losses in total size. It is shown as “—” when a creator has no losses at all, because the ratio is undefined rather than infinitely good.

An earlier version opened on a score of Crumpert’s own construction — roughly (weighted return ÷ volatility) × sample discount — which ranked creators by how consistently they had earned their returns.
It was removed. Ranking people you can pay to follow, using a risk-adjusted figure the platform invented, is Crumpert making a claim about them that this data cannot substantiate. The measurements it was built from are all still on the page; they are just reported rather than combined into a verdict.

Most recent (the default) — whose call settled most recently. Return — highest weighted return first. Most calls — largest number of settled calls first.
None of these is a recommendation, and none accounts for risk. A creator at the top of the return column may be there because they took positions you would not want to hold.

A record with fewer than 5 closed calls is marked provisional and sorts below every established record, whichever column you choose. Its win rate is withheld — “100% from one call” is a claim, not a record. This is a sample-size caveat, not a risk rating.

Nothing on the record page rates, grades, or verifies the risk of anyone’s approach. Before following or subscribing to a creator, do your own research and consider whether what they do is something your own risk tolerance can absorb.

An earlier version of this page sorted creators into “Steady”, “Balanced”, “Aggressive” and “High risk” bands by volatility. Those labels were removed.
The thresholds had to be widened twice before the records spread across more than one band — which means they were fitted to the data that happened to exist, not derived from anything. Volatility is also not the same thing as risk. A creator who loses 2% on every single call has almost no volatility and would have been labelled “Steady” while reliably losing money.
The measurement is a fact. The category was a judgement this data cannot support, so you get the number and draw your own conclusion.

Every return here is a percentage. Nothing records how much of anyone’s capital was actually behind a call. A creator risking a tenth of their book and one risking all of it produce identical numbers on this page.

Ten open calls on ten different semiconductor names is close to one large bet, not ten independent ones. Nothing here detects that.

A creator chooses which calls to open on the platform. Ideas they had and never opened here are not part of the record, and cannot be.

Everything on that page describes calls that have already settled. None of it forecasts the next one. Past performance is not a reliable indicator of future results.

Crumpert is a publishing platform. Posts and trade records are creators’ own opinions — not investment advice, not a personal recommendation, and not financial promotions approved under FSMA s.21. The measurements and the default ordering on the record page are computed by Crumpert from recorded prices; they describe past results and are not an assessment of any creator, nor a suggestion to subscribe to one.