As rarely as its strategy allows. Every trade costs a fee, and fees are the one certainty in trading — profits are not. A bot that trades constantly has to out-earn a constant fee bill before it makes anything at all. Frequency is a cost, not an achievement, so the burden of proof sits on every extra trade.
Why does every trade cost money, even when it wins?
Because exchanges charge a fee on both sides of every trade — once when you buy, once when you sell. A typical crypto exchange charges around 0.1% of the trade size per side, so one complete "round trip" (buy, then sell) costs about 0.2%. That fee is collected whether the trade wins, loses, or goes nowhere. There's also a quieter cost called the spread: the small gap between the price you can buy at and the price you can sell at in any given moment, which you cross every single time you trade. Fees and spread are the only guaranteed numbers in trading. Profit is a maybe; the cost of participating is a certainty, and it scales directly with how often your bot pulls the trigger.
What does the fee math look like at different frequencies?
Multiply the round-trip cost by the number of trades and the picture turns stark fast. Assume a $1,000 account, a bot that uses the full balance on each trade, and a 0.2% round-trip fee:
| Frequency | Round trips per month | Monthly fee bill | Strategy must earn this just to break even |
|---|---|---|---|
| 2 trades a month | 2 | $4 (0.4%) | 0.4% per month |
| 1 trade a day | 30 | $60 (6%) | 6% per month |
| 10 trades a day | 300 | $600 (60%) | 60% per month |
Read that last row again. A bot trading ten times a day must out-earn a fee bill equal to 60% of its account every single month before it produces its first dollar of profit. Very few strategies anywhere clear that bar, and none clear it by accident.
Doesn't trading more often mean more chances to profit?
More trades mean more chances — but chances at what? Over minutes, most price movement is noise: random wobble with no direction behind it. A fast-trading bot is mostly reacting to static while paying a real fee for each reaction. An edge, if a strategy genuinely has one, is a small average advantage per trade. Trading more often doesn't grow that advantage; it just runs the same small edge — or non-edge — more times while the fee meter runs continuously. If the edge per trade is smaller than the cost per trade, higher frequency simply means losing money faster and with more precision.
