Trade levels
Trade levels decide where stops and targets may sit once a coin qualifies — not whether it qualifies (that’s conditions / score gates), and not how the stop moves after fill (that’s Exit policy).
Authored on the strategy. At construction, candidate stops are filtered by a band; survivors are paired with targets; pairs that miss Min R:R are dropped.
Both stop bounds are × ATR so the band is a real interval on every coin. A percent ceiling next to an ATR floor used to invert silently when ATR rose.
Platform structural cap: no stop beyond 3 × ATR, at any setting.
Defaults on a new strategy: Floor 1.0 × ATR · Hard stop 2.0 × ATR · Min R:R 1.5.
Quick map
| Dial | One line | Default |
|---|---|---|
| Floor (Min stop loss) | Tightest stop accepted | 1.0 × ATR |
| Hard stop | Widest stop accepted and the exchange protective order distance | 2.0 × ATR |
| Min R:R | Target distance ÷ stop distance must clear this | 1.5 |
Floor (Min stop loss)
How tight is too tight for a stop on this strategy?
Candidates closer to entry than this are removed before targets are anchored. Raising the floor also moves the anchor: R-multiple targets are measured from the tightest surviving stop, so every target widens with it.
| Unit | × ATR (same axis as Hard stop) |
| Default | 1.0 — tightest deterministic ATR rung the constructor emits |
Example. Floor 0.5 × ATR — accepts half-ATR stops (cheap to be wrong, easy noise stops). Floor 2.0 × ATR — only wide stops; targets move out with the anchor.
Hard stop
How far can price go against a trade before the exchange closes it?
Two jobs:
- Exchange protective order — resting distance on the venue; evaluated on mark when touched
- Widest candidate admitted — stops further than this are removed
Position size is derived from it: notional is set so a fill at this distance costs the percent of headroom your size preset names. Widen hard stop → smaller size for the same risk dollars.
The setup’s own chosen stop is the soft stop (closed strategy-bar judgment). Hard stop is the touch case.
| Unit | × ATR |
| Default | 2.0 — admits every deterministic ATR rung |
| Structural max | 3.0 × ATR — nothing beyond; no dial recovers it |
Example. Hard stop 1.2 × ATR — only the closest ATR / tight structural stops survive. Hard stop 3.0 × ATR — band removes nothing by width; resting order sits three ATR out; same budget buys the smallest position.
Band ordering (required)
Floor must sit strictly below hard stop (floor < ceiling).
| Case | Result |
|---|---|
| Floor < hard stop | Legal band — candidates in between can survive |
| Floor = hard stop | Empty window — equal bounds admit no width; no candidate hits by construction |
| Floor > hard stop | Refused at authoring |
The editor clamps siblings until the predicate holds; the save path refuses a disordered band. Same rule at every layer that stores them.
Min risk : reward
How much reward must a setup offer for the risk it takes?
Last filter: every surviving stop is paired with every target; the pair keeps only if target distance ÷ stop distance clears this ratio. It removes pairs, not stops — so a coin can have valid stops and still build no setup.
| Unit | Bare ratio |
| Default | 1.5 — target at least half again the stop distance |
Example. Min R:R 1.0 — almost any target at least as far as the stop. Min R:R 3.0 — only the top of the target ladder clears; few setups.
Break-even win rate before fees rises with the ratio (higher bar → you need fewer wins to stay flat on expectancy — and fewer setups clear).
How construction reads the dials
- Build stop candidates (ATR rungs, swings, bands, …)
- Floor / hard stop filter the band (
too tight/too wide/ beyond3 × ATR) - Anchor targets from surviving stops
- Min R:R drops underpaying pairs
If every stop dies in the band, nothing reaches R:R. If stops survive but no target anchors, R:R never runs.