A stop loss is an order placed in advance that closes your trade automatically once the price moves against you by a defined amount. It is not a prediction — it is a ceiling. Once set, it means no single trade can take more than a fixed slice of your capital, no matter how wrong the call turns out to be.
It is the simplest risk tool in trading. It is also, in practice, the most ignored.
When you open a trade, you set a stop-loss level before entering. If BTC drops to that level and you're long, the exchange closes the position. Full stop. You don't have to watch it, decide, or argue with yourself about whether to "give it more room." The decision was made before emotion entered the picture.
On an exchange like Binance, a stop-loss order sits in the exchange's system and executes automatically — it doesn't depend on an app being open or a bot being online. That's why exchange-native stops are the only kind worth relying on for overnight or multi-hour trades.
Most traders spend enormous energy on picking the right moment to enter a trade. Very few spend equivalent time on where the trade is wrong. Yet the stop loss is the only variable that determines your actual downside.
Consider two traders who both enter BTC at $85,000. Trader A has no stop and "will decide when it feels right." Trader B has a stop at $83,500 — 1.76% away. When BTC drops to $78,000, Trader A watches a 8.2% loss grow while rationalising. Trader B closed at −1.76% three days earlier and moved on to the next trade.
The stop loss doesn't make Trader B right. It makes their wrong trades survivable.
Percentage-based: close the trade if price moves X% against you (e.g. −2%). Simple, consistent. Doesn't account for the asset's actual volatility.
Structure-based: place the stop just below a support level or above a resistance level that the trade's thesis depends on. If price closes below that level, the original thesis is invalidated — so exit. This is how systematic strategies typically do it, because the stop becomes part of the signal logic, not an afterthought.
ATR-based: use the asset's Average True Range (a volatility measure) to set a stop that breathes with the market. In a calm BTC session, a stop at −0.5% might make sense; in a volatile session, the same setup might warrant −2%. ATR scales the stop to current conditions.
The most cited failure case is "stop-hunting" — the idea that large players deliberately move price to your stop level to take you out, then reverse. This happens, but it is far less common than people assume, and the solution (wider stops) has its own cost in risk taken.
The more common failure is the gap. On weekends, thin liquidity hours, or during sudden news events, price can move from $85,000 to $80,000 in seconds, skipping over your stop entirely. The exchange fills you at the next available price — which might be $80,100. That is a "slippage" event, and no stop prevents it. Position sizing (keeping each trade small) is the real defence against gaps.
The third failure is human: moving the stop further away once price approaches it. "Just give it a little more room." This is almost always a mistake. The stop was set on the original thesis; if you're moving it because of hope rather than new information, you are erasing the entire protection the stop was designed to provide.
Across the 272 closed trades published on /lab, 100% of losing trades exited via a pre-set stop loss or trailing stop — none were manually held past the original exit point. A recent three-week window included 14 losses. Every one was capped. A cluster of simultaneous stops in one session (three SHORTs stopped within hours of each other) looks painful in a table. In the account, each loss was a planned, finite amount.
Stop discipline doesn't prevent losing windows. It prevents losing windows from becoming losing accounts.
Stops cost you trades. There will be sessions where your stop fires and price reverses immediately — you got "stopped out" at the worst possible moment. This is real and it happens. Systematic backtests show that even very tight stops sacrifice some trades that would have recovered.
The question is: which cost is more survivable? The occasional stop-out that would have turned back into a winner, or the rare unprotected loss that goes much further than you planned? Every trader who has held through a −30%, −50%, or −90% drawdown while telling themselves "it'll come back" has answered this question wrong, in real money.
Every trade at Glasshouse runs with a pre-committed stop. The full live and paper book — stops, exits, wins and losses — is published openly at /lab.
📘 Related: what is a trailing stop — the exit that locks in gains as momentum extends, and position sizing — why how much you bet matters more than which coin.
Join the Telegram → · see the live research · or talk to us.