Short-horizon state classification
Read Short-Term Momentum in 6B With Observable Inputs
An illustrative sequence of five completed closes runs 1.2790 → 1.2796 → 1.2793 → 1.2798 → 1.2800. Net displacement is +10 ticks and total close-to-close travel is 16 ticks, producing path efficiency of 10/16 = 0.625. The record separates direction from the path taken.
|+10 ticks|÷16 ticks traveledDirection: upPath: partly indirect
A compact market snapshot
A Momentum Record Combines Direction, Speed, Path and Participation
Momentum is not one oscillator reading. Define the observation window, then describe how far price moved, how directly it traveled, whether range and venue participation changed, and whether the move was executable.
Last eligible price minus first eligible price in the hypothetical window.
Sum of absolute close-to-close changes across the same completed bars.
Absolute net change divided by path length; bounded from zero to one.
Execution snapshot only; it can change before an order arrives.
No value above establishes that 0.625 is strong, weak or profitable. Thresholds require a causal baseline by session, event state and volatility regime, then untouched validation after costs.
The measurement vector
Define Five Inputs Before Naming the State
Use completed observations only. If the rule reads a five-minute bar, that bar's final high, low, close and volume are unavailable until its close.
| Input | Example calculation | What it describes | What it does not prove |
|---|---|---|---|
| Signed displacement | last - first, expressed in 0.0001 outright ticks | Direction and net distance over the window | Path quality or next return |
| Path efficiency | |net change| / sum(|step changes|) | Directness between zero and one | That a direct move will continue |
| Range ratio | Current completed-bar range divided by prior-only same-clock median range | Expansion or contraction versus a declared baseline | Directional intent |
| Participation ratio | CME contract volume rate divided by prior-only same-clock median rate | Venue-specific activity relative to comparable windows | Global OTC GBP volume or participant motive |
| Execution quality | Spread, entitled depth, slippage estimate, feed and order status | Whether the measured move is plausibly tradeable under the plan | That displayed liquidity will remain |
efficiency = |Pend - Pstart|÷Σ |Pi - Pi-1|For standard 6B, one current outright Globex tick is 0.0001 USD/GBP and $6.25. M6B uses the same price increment and a $0.625 tick. Verify full mechanics in the canonical contract guide.
Do not force a binary label
Classify Directional Expansion, Noisy Expansion, Drift and Balance
State rules should use a frozen vector of inputs. The qualitative descriptions below show the logic; they are not universal numeric cutoffs.
| State | Displacement / efficiency | Range / participation | Operational reading | Primary risk |
|---|---|---|---|---|
| Directional expansion | Material signed move with high relative directness | Range expanding; participation at or above declared baseline | Continuation candidate only at a valid location and price | Chasing after most of the move is already spent |
| Noisy expansion | Large net or gross movement with low efficiency | Range elevated; participation may be high | High activity without clean directional control | Wide stops, whipsaw and slippage |
| Quiet directional drift | Consistent signed move but modest speed/range | Participation normal or low | Direction visible; urgency unproven | Thin conditions or one print can distort the state |
| Balanced rotation | Small displacement and low efficiency | Range and participation near baseline | No momentum priority | Inventing a trend from alternating bars |
| Shock / unclassified | Extreme or discontinuous move | Spread, feed or event thresholds fail | Dedicated event rule or stand aside | Applying normal baselines to abnormal conditions |
Before any branch
Comparable Contract, Clock, Data and Event States Validate the Reading
A clean formula can still be wrong for the decision if the contract rolled, the comparison window contains a different session phase or the feed missed the fastest part of the move.
Exact dated 6B/M6B contract, roll status and liquidity migration are known.
Window and baseline share timezone, daylight-saving treatment and session phase.
Only completed bars and prior-only benchmark observations are used.
No sequence gap, stale best bid/offer, duplicate burst or unexplained correction.
Scheduled release and central-bank windows follow the predeclared event policy.
Spread, slippage allowance, order type and whole-contract size fit the plan.
Activity changes across the electronic day. Comparing a London/U.S. overlap window with a quiet late-session baseline can manufacture "relative momentum." Match the time-of-day bucket or model that difference explicitly.
Entry or no entry
Momentum Is a State Input, Not the Entry Price
A continuation decision still needs location, a causal trigger, structural invalidation and acceptable distance. Stronger momentum can make an entry worse if price has moved too far from the failure point.
Classify
All required inputs and gates produce one frozen state.
Check distance
Predeclared level, extension and room-to-invalidation rules.
Wait for price
Break, pullback or response rule completes causally.
Compute size
Zero whole-contract size or failed execution gate means no trade.
Continuation branch
- Directional-expansion state is valid.
- Price is not beyond the maximum extension from the reference.
- A predeclared pullback or acceptance trigger completes.
- Invalidation remains structural, not widened to fit size.
- Spread and slippage budget pass at decision time.
No-entry branch
- Momentum state is shock/unclassified or based on missing data.
- Price has outrun the permitted entry zone.
- Trigger did not complete before expiry.
- State conflicts with the higher-level scenario under a hard filter.
- Calculated size is zero or broker/exchange restrictions fail.
Failure is more than a red bar
Define Loss of Momentum Before the Position
Momentum can weaken without reversing, and price can reverse before a lagging metric changes. Exit or reduce only under the written rule; never reinterpret a losing trade as "still strong on the higher timeframe."
| Condition | Observable rule | Possible meaning | Required response |
|---|---|---|---|
| Structural invalidation | Eligible trade/close crosses the predeclared price | The specific setup is wrong | Execute the stop/exit protocol; no debate |
| Efficiency decay | Rolling prior-only efficiency falls below frozen threshold | Path becomes less direct | Only act if this was an exit rule, not an observation added later |
| No incremental progress | New participation fails to extend price by minimum ticks within time limit | Stall or two-way absorption hypothesis | Apply time stop or require a new setup |
| Opposing displacement | Counter-window closes beyond specified retracement fraction/level | Local state transition | Reduce/exit only as predeclared |
| Execution deterioration | Spread, data or order-status gate fails | Thesis may be unchanged but execution is unsafe | Cancel new orders; manage existing risk under contingency plan |
Hypothetical risk translation
A standard 6B long reference at 1.3027 with structural invalidation at 1.3015 has 12 ticks of price risk. At $6.25 per tick, that is $75 per contract before costs. With a $180 cap and $15 position-level cost/slippage allowance, floor(($180-$15)/$75)=2 contracts, for $165 planned total. A worse fill can exceed it.
- Distance
- 12 ticks
- Price risk
- $75 / contract
- Quantity
- 2
- Planned total
- $165
One-page operational workflow
Measure, Classify, Gate, Act and Review
This is the durable handoff. Every required blank defaults to stand aside, and every rejected candidate stays in the journal.
| Stage | Required inputs | Output | Fail-closed condition |
|---|---|---|---|
| 1. Identify | Dated contract, feed, session, timezone, window, bar type | Versioned observation context | Wrong/unknown month or clock |
| 2. Measure | Displacement, path length/efficiency, range ratio, participation ratio, spread | Numeric feature vector | Missing or future-dependent value |
| 3. Classify | Frozen thresholds and priority order | One state plus confidence/data flag | Overlapping rules without tie-break |
| 4. Locate | Reference level, extension limit, room to invalidation | Eligible or overextended | Momentum alone is the location |
| 5. Trigger | Causal entry rule and expiry | Timestamped entry candidate | No trigger before expiry |
| 6. Size | Invalidation ticks, tick value, cost/slippage allowance, risk cap | Whole-contract quantity | Zero size or funding/liquidity gate fails |
| 7. Manage | Price failure, time stop, momentum-loss and execution contingencies | Predeclared action | Rule invented after entry |
| 8. Review | Decision-time snapshot, orders/fills, excursions, costs, data flags | Complete candidate record | Only winners or filled trades retained |
The workflow makes a rule testable. It does not establish that directional expansion, high relative volume, efficiency or any combination predicts profitable 6B continuation.
Sources, method and editorial disclosure
- CME Group FX Product Guide 2026 and CME Rulebook Chapter 251 for current standard 6B unit, outright tick and contract rules.
- CME Time & Sales description for exchange trade price, time and size fields.
- CME Market Data Platform for event-based CME futures and options market data context.
- CME daily settlements and CME trading-hours and holiday resources for price-type and session controls.
Sources and contract terms were reviewed August 13, 2026. All prices, tick counts, ratios, state descriptions and risk examples are transparent hypothetical illustrations. They are not live market readings or study findings. Thresholds must be frozen and tested on licensed causal data with roll, session, event, fee, spread and slippage controls.