Execution decision workflow
6B Breakout Levels: A Testable Decision Process
Price trades two ticks above yesterday's high, but the bid-ask spread has widened, the quote feed just recovered from a gap and the next scheduled release is seconds away. That is not a breakout to upgrade. It is a candidate to reject.
A touch is not acceptanceNo performance claim
Define the claim
"High Probability" Is Not an Observable Level Property
A price can be visible and still have no demonstrated trading edge. The useful objective is narrower: define a breakout candidate before the test, determine whether current data and execution conditions make the observation trustworthy, specify what confirms or rejects it, and size the planned failure point.
Observable facts
- The dated 6B contract and exact price level.
- When and from which session the level was formed.
- Trades, quotes, volume and spread in the chosen CME feed.
- The entry, invalidation and order outcomes actually recorded.
Claims that require a study
- That prior highs break more often than they reject.
- That volume confirmation improves expectancy.
- That a retest entry is superior to a close-through entry.
- That any filter creates a stable win rate after costs.
The framework creates reproducible decisions and a research-ready event log. It does not establish that a named level, time window or confirmation rule is profitable.
Level provenance
A Level Needs a Timestamp, Contract and Construction Rule
Write the source while the level is still ahead of price. A line added after the move is an annotation, not decision-time evidence. Use the actual dated contract for execution and reconcile any continuous-chart level around a roll.
| Candidate source | Predeclared construction | Required check | Common failure |
|---|---|---|---|
| Prior session high or low | Named session clock, timezone and dated contract | Holiday/short-session status and data completeness | Mixing electronic-session and local-calendar definitions |
| Prior settlement area | Official settlement plus a fixed tolerance in outright ticks | Use CME settlement, not an arbitrary chart close | Treating settlement as a single guaranteed turning price |
| Multi-test auction boundary | Minimum touches, separation, tolerance and lookback fixed in advance | No future bar may create or relocate an earlier boundary | Drawing through failures until the line looks clean |
| Scheduled-event extreme | Exact release window and first causally available high/low | Clock synchronization and feed status | Using a bar whose final high was unknown at entry time |
| Weekly swing | Fixed pivot algorithm and confirmation delay | Dated-versus-continuous contract reconciliation | Repainting a pivot with later bars |
| Round price | Exact increment and maximum distance from price | Test separately from levels with auction provenance | Assuming visual salience equals participant interest |
Before the trigger
The Market-Quality Gate Can Cancel a Good Level
The gate is not a directional opinion. It answers whether the current observation can support the planned order and whether the planned risk is executable. Thresholds must be written before seeing the outcome.
Exact 6B or M6B month matches the order ticket; no continuous alias is being traded.
No unexplained timestamp reversal, missing interval, duplicate burst or stale best bid/offer around the test.
Inside spread is no wider than the predeclared limit for this setup and session.
The feed supports the intended observation; a top-of-book feed cannot validate full-depth claims.
The order is permitted, reduced or prohibited under the written scheduled-news window.
The contract is not inside a prohibited delivery/roll window and liquidity has not migrated elsewhere.
Fail closed on missing evidence
If quote timestamps cannot be trusted, spread cannot be measured or the contract identity is uncertain, classify the candidate as unobservable, not as a win, loss or discretionary pass. Missing data cannot be repaired with the later chart.
- Pass
- All required fields present
- Fail
- Threshold violated
- N/A
- Feature not in this rule
- Unknown
- Stand aside
State transition
Separate Probe, Break, Acceptance and Failure
A single trade beyond a line is different from a completed bar, sustained trade or a retest. Choose one causal trigger family; do not promote a failed trigger into another family after the outcome is visible.
Approach
Price is inside the predeclared distance. No breakout has occurred.
Probe
At least one eligible trade crosses the boundary; acceptance is unproven.
Declared trigger
The exact close, hold-time, volume or retest rule completes.
Invalid or expired
Price crosses failure level, quality gate fails or time limit ends.
| Trigger family | Information required | Tradeoff | Must be invalidated by |
|---|---|---|---|
| First eligible trade beyond zone | Reliable time and sales | Fastest, most exposed to brief probes and slippage | Predeclared return or structural failure |
| Bar close beyond zone | Fixed bar construction and close timestamp | More delay and potentially wider entry distance | Close-back or price level specified before entry |
| Acceptance for N seconds/minutes | Synchronized trades plus an exact duration rule | Filters brief probes but can enter after extension | Loss of boundary plus time expiry |
| Break and retest | Break state, return tolerance and rejection rule | May improve location; many moves never retest | Retest trades through the structural boundary |
Volume, delta or depth may be predeclared features, but they do not reveal motive and do not transform a trigger into proof. The 6B order-flow guide defines what CME futures feeds can and cannot observe.
Position math
Size the Failure Point, Then Decide Whether a Trade Exists
This hypothetical example demonstrates arithmetic only. It does not show that the breakout rule has positive expectancy.
Declared plan
- Instrument: dated standard 6B.
- Boundary: 1.2846 prior-session high.
- Trigger: eligible close at 1.2851.
- Structural invalidation: 1.2835.
- Distance: 0.0016 = 16 outright ticks.
- Per-contract price risk: 16 × $6.25 = $100.
Account constraint
- Maximum planned loss budget: $240.
- Cost/slippage allowance for the position: $18.
- Budget left for price risk: $222.
floor($222 / $100) = 2standard contracts.- Planned total: $200 + $18 = $218.
- Unused buffer: $22; it is not permission to widen the stop.
contracts = floor((risk cap - cost allowance) / (stop ticks × tick value))→floor((240 - 18) / (16 × 6.25)) = 2A stop can trigger into a gap or a thin order book and fill worse than 1.2835. If the plan has no explicit adverse-fill allowance and response to a rejected order, it is incomplete. See the canonical 6B/M6B contract math before sizing either product.
Pass or stand aside
The Breakout Decision Card
Complete the card before entry. A blank required field is a stand-aside result.
| Gate | Pass condition | Stand aside when | Journal field |
|---|---|---|---|
| Identity | Dated contract, venue and feed match | Symbol mapping or roll status is uncertain | contract_id |
| Provenance | Level existed before approach under frozen rule | Line was moved, backfilled or depends on future bars | level_rule, level_price |
| Quality | All required spread, feed, event and liquidity gates pass | Any required measure fails or is unknown | quality_vector |
| Trigger | One named causal trigger completes before expiry | Only a probe occurs or trigger arrives after expiry | trigger_ts, trigger_type |
| Invalidation | Objective level and order response are defined | Stop is selected from desired dollar risk rather than structure | invalid_price |
| Size | Whole-contract size fits risk cap after allowances | Calculated size is zero or margin/liquidity rules fail | risk_cap, qty |
| Decision | All required gates pass | One required gate fails | pass / stand_aside_reason |
After the event
Preserve Rejections, Non-Fills and Data Failures
A screenshot folder of successful breaks cannot answer whether the rule works. The event set must include every eligible approach and the reason each one passed, failed, expired or could not be observed.
Candidate level, timestamp, current contract, spread, event state and trigger status before outcome.
Submitted type and price, acknowledgment, modifications, partial fills, cancellations, rejections and timestamps.
Maximum favorable and adverse excursion from a frozen reference, without changing the exit rule.
Commissions, fees, spread paid and slippage relative to the decision-time executable quote.
Sequence gaps, stale quote flags, contract-map issues and any vendor correction.
Parameter version, exception used and whether any interpretation changed after seeing the result.
Research handoff
Only after the event log is complete should the rule be tested against a predeclared baseline, with costs, overlapping-event treatment, roll controls, session and news splits, untouched validation dates and parameter sensitivity. A neat decision process is useful operationally, but it is not itself evidence of expectancy.
- Finding
- Requires completed study
- Framework
- Provided here
- Win rate
- Unknown
- Promise
- None
Sources, method and editorial disclosure
- CME Group FX Product Guide 2026 for current 6B/M6B codes, units, increments and physical-settlement classification.
- CME Rulebook Chapter 251 for standard 6B contract mechanics and trading-termination rules.
- CME Time & Sales description for exchange trade price, time, trade size and conditionally represented bid/offer fields.
- CME daily settlements for authoritative settlement-data access and the distinction between settlement and an arbitrary chart close.
- CME holiday and trading-hours resource for current session and holiday verification.
- CME performance bonds and margins for the changing collateral framework.
Sources and exchange terms were reviewed August 13, 2026. The level taxonomy, gates, hypothetical prices and $240 risk example are editorial illustrations, not exchange rules or tested findings. This page presents no original event sample, win rate, expectancy or trading recommendation. Any performance claim requires a complete reproducible study on licensed data with realistic fills and costs.