Where to Put Your Stop Loss in Crypto

10 November 20267 min readBy Alex Marzell, founder of CoinGPT

Put your stop loss at the exact price where your trade idea is proven wrong, then a little past it. Not at a fixed percent, and not at whatever loss you can afford. Get that one price right and your position size and target follow from it. Here's how to find it on a real Bitcoin chart, the mistake that stops you out just before the move, and how the Trade Agent plans it in seconds.

Watch: Where to Put Your Stop Loss in Crypto (Let AI Plan the Trade) by Alex Marzell, founder of CoinGPT. Chapters are on YouTube.

The mistake: stops based on you, not the chart

Most traders choose their stop from the wrong place. They use a fixed number, like "always 1%", or they ask how much they can afford to lose and put it there. The market doesn't know your rule or your account. Normal price movement goes straight through a stop that's based on you rather than the chart.

An example from the Bitcoin 1H chart on 5 November 2026: price had just dropped from above 80,000 to about 75,500. Buy at 75,890 with a stop $140 below and, on the next candle, price dips through it. You're out. Then Bitcoin rallies more than $800 without you.

How to place a stop loss in five steps

  1. Ask what has to happen for the trade to be wrong. If you're buying a bounce, the idea is that the low holds, so the low is your line. If you're shorting, it's the last high. On this chart the low was 75,568.
  2. Go just past the level, not on it. Price often wicks just past an obvious low, so give it room. Here that means 75,520.
  3. Check it against normal movement. An hourly candle was moving about $700 on average. This stop is only $370 away, so it's tight and normal noise can reach it. Know that before you go in.
  4. Let the stop set your size. Risk $10 with a 0.5% stop and the position is about $2,050. Use a 1% stop and it's about $1,000. The loss at the stop is still $10. (Work it out in the position size calculator.)
  5. Check the target. Many traders want a target at least twice as far away as the stop. Here that's $740 up, around 76,630. If the reward isn't there, skip the trade.

On a short, it flips: the stop sits above price, just past the last high, and the idea is wrong if price climbs back above it.

How the Trade Agent plans the stop for you

That's five checks on every trade while the market is moving. In CoinGPT the Trade Agent writes the first draft. Pick a coin and a timeframe and ask for a setup. It reads the chart, finds the key levels and drafts the entry, stop and target, drawn on the chart with the reward to risk.

  • Wrong if. The line to read first. It's the same test as step 1: where the idea stops making sense. That's your stop.
  • Watch out. What could go wrong, such as a bigger structure still making lower highs.
  • Heads up. Shown at the top when the trade goes against the bigger trend.
  • Full analysis. The key levels, trend, momentum and what would cancel the setup.

You can question it too. Ask for other stop levels and it will tell you, for example, that a tighter stop has no clean swing low behind it and would risk being stopped out by normal noise. When you're happy, Review & Execute sends the stop and target with the order, sized from your risk (see how much to risk per trade).

Case study, 5 November 2026: what happened to this setup

It lost. Bitcoin rallied to about 76,560, just $70 short of the target. Around five hours after the setup, it fell through the stop. At 1% risk on a $1,000 account, that's a $10 loss: exactly the number agreed before going in. Price bounced that night, which hurts, but the next day it made a new low at 74,888.

A stop won't always feel right in the moment. It's there so that one wrong idea costs you 1%, not your account.

One mistake to avoid: never squeeze your stop just to get a bigger position. It looks better on paper, but now your stop sits inside normal noise and is much easier to hit.
The whole rule
  1. Find the price where the idea is wrong.
  2. Go just past it.
  3. Check it against normal movement.
  4. Size from it.
  5. If the target isn't worth it, skip the trade.

Next in the series: how to find these setups in the first place, with the Scanner.

Questions

At the price where your trade idea is proven wrong, then a little past it to allow for wicks. For a long bought on a bounce, that's just below the low; for a short, just above the last high.
Not on its own. A fixed percentage ignores the chart, so normal movement can hit it. Find the level first, then use your risk amount to set the position size.
Many traders want the target at least twice as far away as the stop (2:1). If the chart doesn't offer that, skipping the trade is a valid choice.

This guide is for education only and is not financial advice. AI analysis can be wrong, and in a fast market price can move past a stop. Trading involves substantial risk of loss. Review every setup before you trade.

Let AI draft the plan

Entry, stop, target and where the idea is wrong, on any coin, in seconds.

Open the App