Risk management: the only edge you fully control

You cannot control whether the market goes up or down. You cannot control whether your strategy's edge is real or how long it lasts. The one variable entirely in your hands is how much you lose when you are wrong — and in a market where you will be wrong constantly, that variable dominates long-run outcomes.

Position sizing: the 1–2% rule

The standard professional convention: risk a fixed small fraction of your account — commonly 1–2% — on any single trade. Note that this is the loss if your stop is hit, not the position size. The formula:

Position size = (account × risk fraction) ÷ stop distance

Example: a 5,000 USDT account risking 1% (50 USDT), entering a coin with a stop-loss 4% below entry → position of 50 ÷ 0.04 = 1,250 USDT. A wider stop means a smaller position, automatically. Volatile coins get smaller positions than quiet ones for the same account risk — this is what ATR-based sizing formalizes.

Why so small? Losing-streak arithmetic

Suppose your strategy wins 45% of the time — a realistic figure for a decent trend system. Over 100 trades, the probability of hitting a streak of six or more consecutive losses is high enough that you should simply plan on it happening. At 1% risk per trade, that streak costs ~6% — annoying. At 10% per trade, it costs ~47% — and the account then needs to nearly double to recover. Sizing is not about the average trade; it is about surviving the inevitable bad run with your capital and judgment intact.

Drawdown math is asymmetric

Losses compound against you. Every percentage point of drawdown you prevent is worth more than a percentage point of profit you chase — which is the entire mathematical case for stop-losses and caps.

Stop-losses that actually protect

• Decide the invalidation before entering. A stop is the price at which your trade idea is objectively wrong — set it when you are calm, not during the move.
• Put it on the exchange, not in your head. Mental stops fail in the two situations that matter: when you are asleep, and when you are rationalizing. An exchange-side reduce-only stop order keeps working even if your connection, your software, or your discipline fails. (This is why the CROT bot places its hard stop on the exchange at entry.)
• Expect slippage. In a crash, a stop at −4% may fill at −5%. Budget for worse-than- stop fills in your sizing; in futures, liquidation is just the involuntary, maximally slipped version.

Leverage without self-deception

Leverage and risk are linked through position size, and controlling size — not the leverage dial — is what matters. A 2x-leveraged position sized at 3% of your account risks less than an unleveraged position sized at 50%. The practical rules: derive size from the 1–2% rule above, and keep your liquidation price far beyond your stop-loss, so that the exchange's liquidation engine never acts before your own exit does.

The diversification trap

Ten positions in ten altcoins feel diversified. In a crypto crash they behave like one big position, because correlations converge when everything sells off together. Position-level limits are not enough; you also need portfolio-level caps — a ceiling on total margin deployed and on aggregate exposure — sized for the day the whole market moves against you at once.

Expectancy: the only score that counts

Expectancy = (win rate × average win) − (loss rate × average loss) − costs. A strategy is profitable only if that number is positive after fees, slippage and funding — and you can only estimate it from a record of real trades. Variance dominates small samples: twenty trades tell you almost nothing; a few hundred begin to. Keep a log of every trade, review it, and size down whenever you are uncertain — uncertainty is the normal state.