The Hidden Truth About Risk Thresholds
Every Automated Trader Must Know.

Here's a sentence that will offend most trading YouTubers: your win rate doesn't matter nearly as much as your risk framework.

You can be right 70% of the time and still blow up an account inside a quarter if your loss band logic is wrong. You can be right 45% of the time and compound beautifully if it's right. This article is about the second scenario — and specifically about how the SilvaBit AI threshold system works, what it does for you, and the three ways we see new users accidentally sabotage it.

What a risk threshold really is.

Strip it to the bone and a risk threshold is a line in the sand. Below that line, the agent trades. Above it, the agent pauses. SilvaBit AI operates two of these lines — one on deposited capital, one on trading profits — and the distinction matters more than most new users realise.

The deposit band.

Your deposit band protects your original capital. Default: 20% of deposited funds drawn down. Reach it and the agent freezes new orders and asks you to review. The band exists because capital preservation is the one game you can't lose — if you preserve capital, you get another month to be smart.

The profit band.

Your profit band protects, specifically, the money the agent has already earned you. Default: 30%. It is deliberately looser than the deposit band — because the role of accrued profit is to give the engine room to breathe through drawdowns without starving the next winning streak.

Why two bands instead of one.

One of the most common pieces of feedback we get from new users is: "why don't you just use one combined loss band?"

The answer is a mental-model answer, not a technical one. A single combined band pushes you toward either over-caution (hitting the band means closing the whole account) or over-greed (treating profit the same as deposit encourages holding losers). Splitting the band keeps the two instincts from fighting.

"Protect what's mine. Risk what's earned." Not a bad rule to run an account on.

What thresholds do NOT do.

This is the part most marketing pages skip. A loss band is not a hard ceiling.

• Execution happens at market prices, not threshold prices.
• A fast gap can breach a band before the agent gets a confirmed close.
• Thresholds pause future trades. They do not retroactively undo past losses.

Said differently: a 20% deposit band does not mean "I will only lose 20% at most." It means "if 20% is drawn, the engine freezes and asks you to look at it." A fast market event can push the realised loss further than the band itself. We state this up front because it would be dishonest not to.

The three mistakes we see every week.

Mistake 1 — Setting bands too tight because "safer is better."

A deposit band at 5% sounds responsible. In practice it causes the agent to freeze inside normal market volatility, missing recovery and locking you out of winning sequences. If 5% draws routinely breach you, your fundamental expectation about automated trading is wrong — not the agent.

Mistake 2 — Treating the profit band as a target.

New users sometimes see "30% profit loss band" and interpret it as "the engine is OK with losing 30%." It is not a target. It's a tolerance before mandatory review. Profits should not be routinely drawn down 30%.

Mistake 3 — Resetting bands manually after every pause.

If every pause ends with you resetting the bands wider so the agent can "just try again", you're not using thresholds — you're removing them. Pauses are supposed to be inconvenient; that's the whole feature.

How to tune your bands sensibly.

If you're brand new to the platform, keep the defaults. 20% on deposit and 30% on profit were set after live testing with hundreds of accounts; they are not arbitrary.

After 90 days, if you have placed at least one withdrawal, revisit them with these questions:

1. Did my bands pause the agent during drawdowns that then recovered? If yes, loosen slightly. If no, keep as-is.
2. Did my bands fail to pause before a loss became painful? If yes, tighten.
3. Am I resetting them often? If yes, I'm not treating them as thresholds — I'm treating them as polite suggestions.

What SilvaBit AI does on your behalf.

Behind the two user-visible bands sits a layered pre-trade check. The engine evaluates projected drawdown risk against both bands before it opens an order, and sizes down (or passes) if the next trade could push either band meaningfully. You never see this unless you dig into the rationale panel — but it's the quiet work that keeps most accounts stable.

The honest conclusion.

Risk thresholds are the single most under-sold, over-performed feature of any automated trading product. They don't promise a win rate. They don't promise zero losses. They promise that the engine has a brake pedal, and that you're the one with your foot on it.

Use them properly and SilvaBit AI runs for years without crisis. Fight them, loosen them, override them — and you'll find yourself the main character of a story we read every week in the support queue.

If you're new, read the platform overview and look at the risk framework section. If you're experienced, revisit your bands this weekend. It is the single most undervalued hour you can spend on the account.