An equity curve is a drawing of decisions about what to count. The smooth ones are usually the ones counting the least, and the shape tells you which decision was made.
What is an equity curve actually plotting?
Whatever the person drawing it decided to count. That sounds like a quibble until you see the two curves you can draw from the same account, one rising steadily and one falling, with no dishonesty in either.
The difference is when a trade enters the picture. A curve on closed trades moves only when a position is shut, so an open loser sits outside the line for as long as it stays open. A curve on account equity moves every minute, because it includes what the open positions are currently worth.
Both are legitimate. Only one of them can be gamed by not closing things, which is why the shape of a curve tells you which one you are looking at more reliably than the label does.
Shape one: what does a smooth rising line hide?
The most attractive shape, and the one that most often means the losses have not been counted yet.
Real trading is lumpy. Losers arrive in runs, and at the loss rate in our as-published replay, 48.1% of the 1,514 outcomes carrying a profit-and-loss figure, as of 7 August 2026, a run of five consecutive losses is more likely than not inside a hundred calls. A curve with no visible steps down has either avoided that arithmetic or is not drawing it.
Three mechanisms produce the look without any editing.
- Closed-trade curve plus open losers. Winners get closed and booked, losers stay open and unbooked. The curve climbs while the account does not.
- No stop published, so nothing ever resolves as a loss. 1,647 of the 3,227 signals we have parsed carry no stop at all, which is 51%. A call with no invalidation level can be held indefinitely, and only the author decides when it ended.
- A wide stop nothing touches. The position survives, the curve stays smooth, and the risk sits in the gap between the entry and a stop that would hurt.
The question that separates an excellent trader from this shape is one line: does the curve include open positions at their current value?
Shape two: what is a staircase telling you?
Flat shelves and clean vertical steps up. It is what a closed-trade curve looks like when winners are realised promptly and losers are not.
The shelf is the tell. During a flat stretch nothing was closed, which means either nothing was open, or something was open and going the wrong way. The curve cannot distinguish those two states, and a follower reading it has no way to ask.
Our own outcome categories show the pattern in aggregate. In the as-published replay, 167 trades took the first target and then stopped out, and 362 stopped out without touching any target. Against that, 293 walked a full ladder. A curve drawn only at target hits would show those 293 and part of the 167 as steps up, and defer the rest.
Shape three: why is a sawtooth the honest shape?
Up, down, up, down, with visible drawdowns. The least impressive shape and usually the most honest one.
A sawtooth means losses are being taken and recorded when they happen, which is what a stop does. The depth of the teeth tells you the risk taken per trade, and the recovery after each one tells you whether the arithmetic works. Both are things you can measure; a smooth line offers neither.
What to look at is the depth of the worst tooth against the account. A drawdown of 20% needs a 25% gain to recover, 50% needs 100%, and 80% needs 400%. The recovery asymmetry is why the largest tooth matters more than the average one, and it is worked through in A losing streak is normal: leverage decides what it costs.
Shape four: what makes a straight line end in a cliff?
A long steady rise ending in a single vertical drop. This is the profile of a strategy that produces small regular gains while carrying an unbounded tail.
The mechanism is always the same regardless of the label attached to it: something is being sold or held that pays a little most of the time and a lot occasionally, and the occasional payment goes the other way. Adding to a losing position produces it. Holding through an adverse move without a stop produces it. High leverage on a mean-reversion idea produces it.
The reason it fools people is that the curve looks better and better right up to the end, and the longer the straight section runs, the more convincing it is. In our index the highest stated leverage is 200x, where the whole margin behind a position is a move of 0.5%, and what happens to an account at the cliff is set out in What account blow-ups really mean.
Shape five: what if one trade carries the whole curve?
A long stretch of nothing followed by a single leap that carries the whole record. The curve is real and the average it implies is not.
One trade producing most of a year's result means the sample is one, and everything before and after is noise around it. The percentage advertised on top of such a curve is arithmetic on a distribution with one observation that matters.
This shape is also the easiest to manufacture without lying: publish many calls, let most go nowhere, and let the one that ran carry the marketing. Whether a record has enough scored outcomes to support any percentage at all is the subject of How many trades a hit rate needs before it means anything.
What does a curve from a signal feed look like?
Full of holes, if it is drawn honestly, because most calls never resolve into anything a curve can plot.
Of 3,167 replayed outcomes in our corpus as of 7 August 2026, 1,583 carry a scored verdict. The rest are entries the price never reached and calls overtaken before the entry was available. In the as-published replay, 1,653 signals cannot be resolved because the channel published no stop.
So a curve built from a channel's feed has long flat sections that are not quiet markets. They are calls in an unknown state, and whoever draws the curve decides what to do with them. Dropping them makes the curve smoother and the record shorter, which is the same trade-off a hand-built track record faces, described in How to check a signal channel's track record in 10 minutes.
Which questions does a curve not answer?
Four, and all four have to come from somewhere else.
| Question | Why the curve is silent |
|---|---|
| How large was the position each time? | A percentage curve normalises size away, so one reckless trade and one careful one look the same |
| Was leverage constant? | A curve steepening because leverage rose looks identical to one steepening because the trading improved |
| Were the trades takeable? | Entries that were never available produce no curve movement and no record of the miss |
| Is this one account or a selection? | A curve shows the account it was drawn from, not the ones that were not shown |
The last row is the one that does the most damage. A trader with several accounts has several curves, and the one that goes on the sales page is not chosen at random.
How do you read a curve in two minutes?
Four checks, in this order. None of them requires the underlying trades.
- Find the axis definition. Closed trades or account equity, including open positions or not. If it is not stated, assume closed trades and treat the curve as a claim about bookkeeping.
- Look for the deepest trough, not the trend. Measure it as a percentage of the peak before it. That number is the risk the strategy actually ran.
- Count the flat stretches. Ask what was open during each one, and whether anything about the curve would look different if a position had been sitting underwater throughout.
- Check what one trade contributes. If removing the single largest move flattens the curve, the record is one observation with decoration.
A curve that survives all four is worth taking seriously, which is a lower bar than it sounds and a higher one than most published curves clear.
What this does not prove
A smooth curve is not evidence of dishonesty. Some strategies genuinely produce them for long stretches, and some traders are simply good. The claim here is narrower: smoothness is compatible with several mechanisms that have nothing to do with skill, and the curve alone cannot tell you which one you are looking at.
Nor is a sawtooth evidence of quality. It shows that losses are being recorded, which is a property of the accounting rather than of the results.
Our figures describe our index as of 7 August 2026: 3,227 parsed signals, 3,167 outcomes, 1,583 scored, 1,653 unresolvable as published, and category counts from the as-published replay. Those come from one replay run under one rule set and are never mixed with our one-take ranking metric. The index is 50 channels, not a random sample of the market, and every figure moves as the replay catches up.
Nothing here is financial advice, and nothing here alleges misconduct by any channel or trader.
The practical read
Three questions turn a picture into evidence, and the first one decides the other two.
What is on the vertical axis, and does it include open positions? How deep is the worst trough as a share of the peak before it? And what happens to the shape if the single best trade is removed?
Current ratings for the channels we can score are at Signal Providers, and how our own scores are built is set out in How to read a Life Score.