Grid trading, properly: what the mechanics actually are

Most people here treat grid bots as a crypto toy. They have been run on stocks and funds for far longer than that, and the mechanics are worth stating properly rather than dismissing, so here is the honest version. A grid divides a price range into levels. You buy one unit at each level going down, sell one unit at each level going up. That's it. You are not predicting direction — you are being paid for providing patience inside a range. Three parameters decide everything: the upper and lower boundary, the spacing between levels, and the size per level. Getting the boundary wrong is fatal, and I'll post about that separately. What it is good at: harvesting oscillation in a market that goes nowhere. What it is terrible at: a market that leaves the range and doesn't come back. Those are the same market on different days, which is the whole problem.
CryptoKarl
30 posts · 0 followers
+ Follow

Related reading

5 replies

N
NightOwl_Yuki· Jul 2026 ago
This is the framing I wish I'd had. "Paid for providing patience inside a range" explains both why it works and why it stops.
I
IndicatorSkeptic· Jul 2026 ago
It's also a short volatility position wearing a friendly interface, which people should know before they run one.
CryptoKarl· Jul 2026 ago
That's the most accurate one-line description of it I've read, and yes, that's exactly what it is.
F
FiveMinFiona· Jul 2026 ago· edited Aug 2026 ago
The part I'd add for anyone setting one up: spacing should come from how far the thing actually travels in a session, not from a round number you like. I check the average 5m range over a couple of weeks first, and if my grid spacing is smaller than that, the levels just get chopped through without ever really being 'patient'. Also worth deciding in advance what you do when price exits the range. Not deciding is itself a decision.
Q
QuietVol· Aug 2026 ago
Agreed on the short vol framing. The way I hold it: you're collecting a small, fairly reliable payoff many times and paying it back in one large draw when the range breaks, so the mean can look fine while the tail does all the work. Which makes the honest question not "does it profit" but "how many range-bound periods do I need to fund one breakout?" Karl, when you size per level, are you sizing off the grid range or off what a full exit would cost you?
Sign in to reply →
← All automated trading