Pionex Guide

Why Grid Bots Bleed in a Crash — and the Settings That Decide How Bad It Gets

A market crash doesn't break most trading bots. It breaks the assumptions you built them on. Grid bots in particular stay active during a sharp drop, buying as price falls. That behavior is either the whole point or a slow-motion disaster depending on how you set them up.

Why grid bots behave differently in a crash

A grid bot doesn't predict direction. It places buy and sell orders at fixed intervals around a price range, then collects the spread as price moves back and forth. In calm markets, that's a quiet income stream. In a crash, the same mechanism fires buy after buy as price punches through each lower level.

The bot isn't malfunctioning. It's doing exactly what you told it to. The problem is that a grid assumes mean reversion — that price will bounce back through the range. When it doesn't, you're left holding a growing position at prices that keep looking worse.

The difference between a dip and a crash

A 3% dip and a 30% crash look identical to the grid logic in the first few minutes. Both trigger buys. The distinction only matters hours later, when you find out whether price recovered or kept sliding. Crash performance is really a question about your range and your capital, not your bot's code.

What actually degrades during a crash

Three things go wrong at once, and they compound.

Liquidity thins out. Order books get shallow fast. Your fills may happen at worse prices than the grid line suggests, especially on smaller pairs. Slippage during a violent move isn't a rounding error.

Funding and fees keep accruing. If you're running any leveraged or futures-based strategy, funding rates can spike during volatility. On Pionex, grid bots can run on spot or futures, and the futures versions carry funding costs that don't care whether you're winning.

Your capital runs out before the range does. This is the big one. A grid with a $1,000 budget split across 50 levels has roughly $20 per level. A crash that blows through 20 levels deploys $400 in minutes, and you still have 30 levels of falling knife ahead of you.

A concrete scenario

Say you set a grid on a token trading at $2.00, range $1.50 to $2.50, 40 levels, $2,000 total. Each level is about $50. Price drops to $1.60 in two hours. You've now bought roughly 16 levels — around $800 — and your average entry sits near $1.80. If price recovers to $2.00, great, you profit. If it keeps falling to $1.20, you're holding a position 33% underwater with 8 more levels still armed and only $400 left. That's the trap.

Settings that decide whether you survive

You can't control the market. You can control these.

  • Range width. A wider range means fewer levels triggered per dollar of drop, but it also means smaller profits per cycle. Narrow ranges look great in backtests and get shredded in crashes.
  • Number of grids. More grids means more frequent small trades. It also means your capital per level shrinks, so you hit the bottom of your range faster.
  • Arithmetic vs geometric spacing. Geometric grids place more levels near the top, which can reduce buying into a deep drop. Arithmetic grids distribute evenly. Neither is universally better.
  • Stop-loss or no stop-loss. Some platforms let you set an automatic stop. Others, including many spot grid setups, don't. Know which one you're running before the crash, not during it.

Spot vs futures grids in a downturn

FeatureSpot gridFutures grid
Liquidation riskNoneYes, if leverage is used
Funding costsNoYes, can spike in volatility
Capital neededFull position costMargin only
Crash behaviorHolds falling bagsCan get liquidated

That last row matters more than anything else in this table. A spot grid can't liquidate you. It can only leave you holding something you don't want. A leveraged futures grid can take your entire margin.

What to check before the next drop

Don't wait for red candles to think about this. Do it now, while things are calm.

  1. Calculate your worst-case fill: how much capital gets deployed if price hits the bottom of your range?
  2. Confirm that amount is money you can genuinely leave parked for weeks.
  3. Check whether your bot has a stop mechanism, and if so, at what level.
  4. Look at your pair's historical drawdowns. Has it ever fallen 40% in a day?

If the answer to number four is yes, size accordingly. Platforms like Pionex make it easy to launch a grid in a few clicks, which is exactly why so many people launch one without running these numbers first.

Frequently asked questions

Do grid bots lose money during a crash?

They don't lose money in the sense of a bad trade — they accumulate a position that may be worth less than what you paid. Realized losses only happen if you sell at a loss. Unrealized losses can be significant, especially if price exits the bottom of your range and the bot stops trading.

Should I turn my bot off during a crash?

Turning it off stops new buys but doesn't undo the ones already filled. Whether to stop depends on whether you still believe in the asset's recovery. If you've lost conviction, stopping prevents further capital deployment. If you haven't, stopping just locks in a worse average entry.

Can a bot outperform a human during a crash?

Sometimes, in a narrow sense. A bot executes its plan without fear or hesitation, which means it doesn't freeze or panic-sell at the bottom. That discipline is real. But discipline applied to a bad plan just loses money more efficiently.

What's the biggest mistake people make with grid bots?

Setting the range too narrow and the budget too small. It feels safe because the numbers look modest. In practice it means the bot exhausts its capital early in any real move and sits idle while price keeps falling.

Before your next crash, write down the exact dollar amount your bot would deploy if price hit the bottom of your range. If that number makes you flinch, your grid is too big.