Crypto doesn't sleep. Neither does the volatility.
A Bitcoin position held overnight in traditional markets would be a non-event. In crypto, "overnight" can mean a 12% drawdown before your first coffee. That asymmetry changes everything — especially how you size positions.
This post explains exactly how we handle position sizing for crypto signals at Trade Forge. No vague talk about "risk management principles." Real methodology. The same logic that runs inside our live system right now.
---
Why Crypto Volatility Is a Different Problem
Most position sizing frameworks were built for equities or forex. Both have relatively bounded intraday ranges. Crypto doesn't.
Annualised volatility for BTC regularly hits 60–80%. Equities average closer to 15–20%. That's not a small gap — it's a different category of instrument entirely.
The standard approach — risk a fixed percentage of account per trade — breaks down fast when volatility is this unstable. A rule that works on a 1.2% average daily range falls apart when the daily range suddenly expands to 6.4%. You're now risking three times what you planned.
So we don't use fixed percentage risk alone. We use volatility-adjusted sizing.
---
Volatility-Adjusted Position Sizing: The Core Logic
The idea is simple. Your position size should shrink when volatility expands, and grow (within limits) when volatility compresses.
We calculate position size using ATR — Average True Range, which measures how much an asset actually moves over a set period, accounting for gaps. It's the most practical volatility proxy for this purpose.
The formula we use:
Position Size = (Account Risk in $) ÷ (ATR × Multiplier)
If your account is $10,000 and you're risking 1% per trade ($100), and BTC's 14-period ATR is $2,800 with a multiplier of 1.5:
$100 ÷ ($2,800 × 1.5) = 0.0238 BTC
That multiplier (the 1.5 above) is calibrated per asset class and signal style. Scalp signals use tighter multipliers. Swing signals use wider ones. The logic: a swing trade needs room to breathe. A scalp needs precision.
---
The 24/7 Problem Nobody Talks About Enough
Equities have a close. Forex has sessions. Crypto runs continuously — and that continuity hides a trap.
Liquidity isn't uniform across a 24-hour period. At 3am UTC on a Sunday, BTC spreads widen. Slippage increases. The same signal that fills cleanly at 2pm London time might fill 0.4% worse at 3am. Over hundreds of trades, that difference compounds into something meaningful.
In our data, signals fired during low-liquidity windows showed a 9.3 percentage point higher slippage rate compared to peak-session signals. That's not a rounding error. It's the difference between a signal with edge and one without.
Our system flags this. Signals tagged as "liquidity-sensitive" include a session filter. We don't suppress them — we surface the context so you can decide how to act on them.
---
How We Recalibrate Weekly
Our models don't set sizing rules once and walk away.
Every week, our calibration engine retrains on live signal outcomes — currently across 2,400+ signals in the dataset. We use Platt calibration (the technique that turns raw model scores into actual probability estimates) to keep our win rate forecasts accurate rather than optimistic.
Why does that matter for position sizing? Because optimal sizing formulas — including Kelly-based variants — are inputs that depend on your actual win rate and reward-to-risk ratio. If your win rate forecast drifts, your sizing drifts. Getting that input wrong doesn't just hurt individual trades. It degrades the whole system over time.
Most signal providers don't publish their calibration methodology because they don't have one. We do, and we update it weekly.
---
Signal Styles and How Sizing Differs
Not all signals are sized the same way. Here's how our three styles differ:
| Signal Style | Typical Hold Time | ATR Multiplier | Max Risk Per Signal |
|---|---|---|---|
| Scalp | Minutes to 2 hours | 0.8–1.2× | 0.5% of account |
| Day Trade | 2–12 hours | 1.2–2.0× | 1.0% of account |
| Swing | 1–5 days | 2.0–3.5× | 1.5% of account |
These aren't arbitrary. They reflect the statistical distribution of drawdown during each hold period, measured against our live signal history.
R-multiples (how much you made relative to what you risked — the only number that matters for evaluating signal quality) differ meaningfully across styles too. Scalp signals have lower average R but fire more frequently. Swing signals have higher target R but wider variance.
---
Who Doesn't Need This
If you're a long-term crypto holder — buying and holding BTC or ETH on a multi-year thesis — none of this applies to you. Position sizing frameworks for active signals are irrelevant to buy-and-hold strategies. You're solving a different problem.
If you're trading with amounts small enough that a 1% loss is genuinely inconsequential to you, the precision of volatility-adjusted sizing matters less. The methodology still holds, but the urgency is lower.
This approach is designed for active traders who care about R-multiples, drawdown control, and compounding edge across hundreds of signals. If that's not you right now, no problem. Bookmark it for when it is.
---
Hot Take
Most "position sizing guides" online teach you to risk 1–2% per trade and call it risk management. That's not risk management. That's risk ignoring. Fixed percentage sizing applied to volatile assets without volatility adjustment is just uniform recklessness dressed up in a spreadsheet.
---
FAQ
Does Trade Forge provide forex signals too?
Our forex signal infrastructure covers 24 pairs and is built to the same algorithmic standards as our crypto signals. Forex signal generation is currently paused while we finalise live compliance processes. Our crypto, commodities, indices, and ETF signals are live and running now. We'll update the community when forex goes live.
How is this different from what other algorithmic trading signal providers do?
Most algorithmic trading signals providers don't publish their sizing logic or calibration methodology. We do. You can evaluate our approach, not just our results.
Is this relevant if I'm looking for the best forex signals in 2026?
Our forex signal architecture — including this volatility-adjusted sizing framework — will apply when live forex signals launch. The methodology is the same across asset classes. The specific ATR multipliers are calibrated per asset.
Do free tier members get the same signals?
Yes. Free tier members receive the same signals as VIP members via Telegram and Discord, with a short delay. VIP delivers in real time across all asset classes and signal styles.
---
If you want to see how this plays out in practice — real signals, real asset classes, real-time — our free Telegram channel is the easiest place to start. No commitment. Just the signals. Find us at tradingforge.io
---
This article is for educational purposes only. It does not constitute financial advice. Trading involves risk of loss.