The most trusted rule in retail trading
Ask anyone how to manage risk in crypto and the first answer is a trailing stop. Set it 10–15% below the high, let it ride, sell when it breaks. My strategy has had them since day one: 10% on Bitcoin and Ethereum, 15% on everything else.
After six months of live trading down 11%, I was sure they were the problem. I was about half right, and the wrong half is the interesting part.
Half right, still losing
The system grades every trade a week after it happens: did the price move the way the rule expected? Here's the scorecard for the trailing stops:
Trailing stops (14 fires)
7 right (price fell another ~16% after selling) · 6 wrong (price rose ~3% after) · 1 pending
On paper that's a coin flip with a good payoff: the right calls dodged big drops and the wrong calls missed small gains. If the story stopped at the sell, the stops would look great.
It doesn't stop at the sell. Every stop-out has to be followed by a re-entry.
Every round trip lost money
Here's every round trip this year:
Bitcoin
Sold $73,323 → bought back ~$79,000 → stopped out again $77,378
Ethereum
Sold $1,874 → bought back $2,509 → stopped out again $2,421
Solana
Sold $81.79 → bought back $106.24 → stopped out again $104.77
Chainlink
Sold $9.00 → bought back $11.70 → stopped out again $11.15
The pattern is the same every time: sell, watch it recover past the sale price, buy back higher, get stopped out again below the buy-back.
My conclusion was obvious: a 10% stop on Bitcoin is about one bad week of normal volatility, so I was selling noise. Get rid of the stops on the majors, let the 200-day trend gate handle real exits, and widen the stops on the small coins based on how volatile each one is.
I had the post half-written. Then I tested it.
What the test said
I replayed the live window from late June to now, starting from the real account's state, and switched on each fix one at a time. The replay of my current rules came out at -3.1%. Here's what each idea did on its own:
Drop the stops on BTC and ETH
-5.7%. Worse. Without a stop, the majors rode the October slide all the way down, because the gate never closed.
Volatility-scaled stops on the alts
-4.7%. Also worse. Wider stops just meant bigger losses on the same whipsaws.
Anchor re-entry to the low, not the sale
Slightly worse on its own (-3.8%). Combined with the timer fix from the companion post, it's worth about 4 points.
My two favorite ideas, the ones I was ready to publish, both lost money. The stops weren't selling noise. In a market that ground down 10% in October, they did their job.
The real leak: the buy-back
My re-entry rule bought a coin back once it was up 12% from where I last sold it, with positive momentum. That sounds prudent: wait for proof the bounce is real.
But it means every re-entry is at least 12% above the exit, by construction. However much a coin fell after I sold, the rule waited until it was 12% above my sale price. A coin that bottomed 25% below my exit still got bought back above it.
The fix is one word: measure the bounce from the low since I sold, not from the sale price. Same 12% threshold, same momentum check. It just starts counting from the bottom.
Stop-losses decide when you leave. The re-entry rule decides what leaving costs you. I spent months tuning the first and never questioned the second.
The bigger point
Stop-losses get all the attention because they're the scary moment, the sale you watch happen. The buy-back gets almost none, because by then the price is going up and everyone is relieved to be back in.
But the round trip is one trade, and you have to judge both legs. My stops were roughly fine. My buy-back rule had a 12% loss built into it.
The re-anchored rule is going into the live agent along with two other fixes; the check-in post has the full results. If you run trailing stops on anything volatile, line up each sale against its buy-back. That's where I found the leak, and it wasn't where I was sure it would be.
The live numbers are on the dashboard.
Comments
Leave a comment
The Gate Opened. My Strategy Still Lost 11%.
In June I promised to report back when Bitcoin reclaimed its 200-day line, good or bad. It did on August 20. Here is the honest ledger, and why the rule I trusted most wasn't the problem.
My AI Agent Followed Every Rule Perfectly. That Was the Problem.
516 runs. 188 trades, each with a written justification, and not one rule broken. And a strategy that kept fighting itself. What six months of an autonomous trading agent taught me about rules versus intent.
Opus 4.8 Stress-Tested My Crypto Strategy. Then It Made Me Admit I Was Wrong.
Two months ago I published a backtest claiming my crypto strategy returned +35%. This week I sat down with Opus 4.8 to pressure-test that claim — and the honest answer was more humbling, more robust, and a better story than the one I'd told.