You took the first target and still lost money

Banking the first target feels like a win, and in most of the signals we parsed the geometry says the position is still underwater once the remainder stops out.

Why does a winning first target end in a loss?

Because the distance to the first target is usually shorter than the distance to the stop, so a small slice of a small gain has to cover a large slice of a larger loss. In the signals we could parse, it usually does not.

This is the trade everybody recognises. Price runs, the first target prints, you take something off, you move on with the day feeling fine about it. Later the remainder stops out. The screenshot of the first fill is real, the feeling of having won is real, and the account is smaller than it was that morning.

Nothing here tells you what to trade or how to split a position. This is arithmetic on numbers the channels themselves published, and it exists because the format encourages a specific misreading of what a partial exit is worth.

The two distances a channel publishes

Every signal with a stop contains the whole answer already. It states where you get in, where the first target sits, and where the trade ends if it goes the other way. Those three numbers fix the outcome of the partial exit before any price moves.

From parse_tp1.py run over the channel archives in this repository, 126 posts across 5 channels publish an entry, a stop and an ordered ladder, as of August 2026:

In 97 of the 126 signals the stop sits further from entry than the first target. The trade risks more than its nearest reward, which is a perfectly ordinary way to structure a trade when the far targets are the point. It becomes a problem the moment the first target is treated as the reward.

What the partial exit actually nets

Take the first target on part of the position, let the rest ride, and the rest stops out. As a move in the underlying, with no leverage:

At the two most commonly advertised splits, a quarter or a third off at the first rung, every single one of the 126 signals nets a loss on this path. Not most of them. All of them, because the geometry the channel published makes it arithmetically unavoidable when the stop is further away than the target.

The number that summarises the whole article sits underneath that table. For the position to come out flat, the median signal needs 60.1% of it closed at the first target, and 77% of the signals need more than half the position closed. Whatever fraction you had in mind when you read the ladder, it was probably smaller than that.

Two comparisons for scale. The same trade with no target taken at all, simply stopped out, is a median -3.82%. The same trade with the whole position closed at the first target is +2.61%. The partial exit lands between them and, at the usual fractions, much closer to the first.

If you use leverage, multiply every figure in that table by it, in both directions. A -1.86% move at 10x is -18.6% of the margin committed, from a trade that hit its first target. What leverage does across a run of these is worked through in what leverage does to a losing streak.

How often does this actually happen?

Often enough to be one of the larger buckets in our replay.

Replaying the signals we can resolve, using each channel's own published stop and its own ladder, 167 trades took the first target and then stopped out, as of August 2026. Widening that to trades that reached any target before stopping, the count is 331, against 293 that walked the ladder to the end.

More trades in that replay hit a target and gave it back than finished the ladder they were sold on. That is the shape of the thing: reaching a target is common, keeping it is the hard part, and a record that counts the first as a win describes very few complete trades. How targets are spaced, and why the advertised number belongs to the far end of the ladder, is covered in partial take profit rules.

The two sets of numbers on this page come from different places and are never combined. The geometry comes from raw channel archives; the outcome counts come from our replay over a parsed corpus. They are not the same trades.

Why it feels like a win anyway

Three reasons, and they are worth naming because they are what the format is exploiting.

The win is realised and the loss is not, for a while. The first fill is a completed event with a number attached. The remainder is an open position, and open positions do not feel like losses until they close. By the time the stop fills, the win has been mentally filed.

The channel's own scoring often stops at the first target. A feed that marks a trade as a win when the first rung prints is describing something that did happen. The rest of the position is simply outside what it chose to count. What a track record can and cannot establish is set out in what a backtest can and cannot prove.

Cutting a winner early is the easier decision. Realising a gain settles an open question, and letting a loss run postpones one. Trading education has described that asymmetry for decades, and this run could not open a primary study to cite a figure for it, so treat it here as the shape of the argument and nothing more. The partial-exit format sits on top of that asymmetry, and the geometry above is what it costs when the remainder does not survive.

That last point is why the fix is not "hold longer". Holding longer changes which error you make, and the numbers here do not say which error is cheaper for you.

What this does not prove

The sample is small and concentrated. 126 posts from 5 channels, and one channel supplies 55 of them. These are the signals that published all three numbers, which is already a minority of what these channels post, and they are not a sample of the market.

The magnitude moves with the parse, and the direction does not. Taking the entry at the near edge of the published zone instead of the midpoint moves the median ratio from 0.66 to 0.54 and the share of losing half-and-half splits from 77% to 98%. Taking the far edge moves them to 0.58 and 61%. Dropping the largest contributing channel leaves 71 posts with a median ratio of 0.96 and 59% losing at a half-and-half split. Every variant keeps the sign and moves the size, so the sign is published as a measurement and the sizes as the ranges they are. The 77% figure in particular falls towards a bare majority once the biggest channel is removed, and it should be read as "a majority, in this sample" instead of a constant.

This is geometry, not outcomes. These 126 numbers describe what the trade would net if the first target were reached and the remainder then stopped. They say nothing about how often that path occurs for any given channel. The replay counts in the previous section address frequency, over a different set of trades.

Costs are excluded. Fees and slippage apply on every leg, and a partial exit has more legs than a single close. Against targets this size that matters, as what fees and slippage do to a 2% target works out in detail. Including them would move every figure in the table down.

We are not recommending a split. There is a genuine trade-off here with a real cost on both sides, and where you sit on it depends on things this article knows nothing about.

The practical read

Before taking a signal, do the one calculation the post has already given you the inputs for: the distance to the first target divided by the distance to the stop. If that ratio is below one, no partial exit at the first target can save the trade unless most of the position goes with it. In our sample that ratio has a median of 0.66 and sits below one in about three quarters of the signals.

That single number tells you what the ladder is really offering. A ratio well below one means the format is asking you to be right repeatedly to stay level, and the first rung is a consolation rather than a result. It also tells you what a channel's win rate is worth, since a feed scoring the first target as a win can be right most of the time while its followers are flat. Why the percentage on its own establishes very little is covered in why win rate alone means nothing.

Write the number down when you take the trade, along with what you actually closed and where. A month of that is the only record that answers whether a channel pays you, and it takes about fifteen seconds per trade.

What the channels we can score have historically published is at Signal Providers. Nothing here recommends any channel, any asset or any way of splitting a position.

Sources

• Entry, stop and ladder geometry, the net-outcome table, the breakeven split and every sensitivity run: work/tp1-then-stopped-out/parse_tp1.py, run against data/channel_dumps/ (39 channels, 23,707 messages, 16,694 after exact deduplication, January 2021 to August 2026). 126 posts from 5 channels publish all three numbers. As of August 2026.
• Outcome counts from the as-published replay, quoted as line items: work/_snapshot-2026-08-07-batch2.md. TP1_THEN_SL 167; all THEN_SL buckets 331; full-ladder completions 293 (68+58+53+50+37+27). As of August 2026.
• Twelve parsed signals plus four cross-channel checks read against their raw posts before publication; longs and shorts both verified. Notes in work/_batch8-shared-research.md.