Premium & discount

Learning path 6/6 · Location decides whether a setup exists

Published 24 Jul 2026 · every rule box checked against mechanization v1 (frozen 16 Jul 2026) · parameter changes land in the changelog first

In three sentences

Split the H4 dealing range at its midpoint: upper half premium, lower half discount.
Bias needs three things at once: structure, range position, and untapped liquidity ahead.
Longs only from discount, shorts only from premium — 'no bias' is a position, and the engine holds it often.

What it is

Take the current H4 dealing range — the span between the last significant swing low and high — and split it at the midpoint. The upper half is premium (expensive), the lower half is discount (cheap). The idea is old market-making logic: buy where it's cheap relative to the range, sell where it's expensive.

The exact engine rule

Mechanization v1 Bias requires alignment of three things on closed H4 bars: structure (swing pattern), range position (longs only from discount, shorts only from premium) and untapped liquidity ahead as a target. If any leg of that triad is missing, bias is none and the engine hunts nothing — BTC produced zero signals in the launch smoke test for exactly that reason: bearish structure, discount position, no long allowed, no short target. Waiting is a feature.

Why "no bias" is a position

The most expensive habit in trading is manufacturing setups in the middle of a range. Encoding location as a hard filter means the engine simply goes quiet when conditions are mediocre — which is also why our signal counts look sparse next to services that alert hourly. Sparse is the point.

Common misuse

Redrawing the range until the trade you already wanted lands in the "right" half. The engine's range is computed from confirmed swings with fixed rules — it can be wrong, but it can't be negotiated with.

Common questions

Why does the engine go quiet for days?

Because location is a hard filter. When structure, range position and a target don't align, bias is none and nothing is hunted. Sparse is the point.

Can the dealing range be redrawn to fit a trade?

Not here — the range comes from confirmed swings with fixed rules. It can be wrong, but it can't be negotiated with.

Is premium/discount a prediction?

No. It's old market-making logic encoded as a filter: buy cheap relative to the range, sell expensive — and never trade against your side of it.

On real charts — marked vs. left alone

Below: the engine's own markings, captured from our TradingView indicator on live charts — screenshots, not illustrations. The second frame in each pair is the one nobody in this category publishes: a candidate that looks like the pattern and gets no mark, because it misses a threshold above. Rules are only testable if their failures are visible too.