Models · 5 live · more shipping

Five specs today. More every month.

Sentio ships with five structural pattern models for ES and NQ futures — and we’re adding new ones on a regular cadence. Every model is deterministic: read the rulebook, tune the parameters, replay any historical trigger. Pick one that matches how you already trade, build your own with the Model Builder, or wait for the next drop.

Interactive · try it

Turn a knob. See what changes.

Adjust the parameters below and the setup geometry redraws in real time — same code path as the app. This is what tuning a model actually looks like.

Forever · Liquidity reversal LIVE PREVIEW
Adjust the impulse knob — the displacement bar grows and the FVG zone widens with it.
Reversal Liquidity manipulation

Forever01

Forever is the full ICT-style delivery-state confirmation stack. Sweep of a key liquidity pool → inverse FVG on the reclaim → CISD closes past the last opposing candle. Nothing fires until every rail lines up, which is why it triggers less often but averages the highest R per fire of the five.

  • Sweep of PDH/PDL, session H/L, or Asia range extreme
  • Inverse FVG forms on the reclaim candle
  • CISD confirms with a close beyond the last opposing candle
  • Entry off the IFVG retest with the sweep wick as stop
Avg R+1.8
Fire rateLow
ComplexityAdvanced
See full spec → Best for: Trending NY AM after Asia range
Forever NQ · 5m · replay #100
Reversal Sweep

Unicorn02

Unicorn defines its entry as the exact zone where a breaker block overlaps a fair value gap on the same displacement leg. The overlap is small by design — natural tight-stop entries with a clean geometric invalidation. When the two zones agree, the pattern is unambiguous.

  • Sweep of recent liquidity on the execution timeframe
  • Displacement candle creates a fair value gap in the return direction
  • Breaker block from the failed order block overlaps that FVG
  • Entry inside the overlap; stop just beyond the sweep
Avg R+1.5
Fire rateMedium
ComplexityIntermediate
See full spec → Best for: Post-sweep reversals · NY AM & London
Unicorn NQ · 5m · replay #101
Reversal Failed sweep with rejection wick

Sweep Failure03

The simplest structural pattern in the roster. A large rejection wick pokes beyond a recent swing high or low, closes back inside range, and immediately forms an inversion FVG on the failure candle. Fires often — the trade-off is average R is lower than Forever or Unicorn.

  • Wick beyond the recent-high (or -low) swing
  • Body closes back inside the prior range (rejection)
  • Inversion FVG on the failure candle
  • Fade entry off the IFVG with the wick tip as stop
Avg R+1.1
Fire rateHigh
ComplexityIntro
See full spec → Best for: Range-bound sessions, cash-open & cash-close
Sweep Failure NQ · 5m · replay #102
Continuation Parent gap invalidated

Inverse FVG04

Inverse FVG is the only continuation model in the roster. A parent gap gets traded through by a displacement leg, flipping the gap's polarity from support to resistance (or vice versa). The retest of the flipped zone is where the continuation trade happens.

  • Parent FVG identified on the higher timeframe
  • Displacement leg trades cleanly through the parent
  • Polarity flips — parent becomes opposite-side liquidity
  • Retest of the flipped zone within N bars = entry
Avg R+1.4
Fire rateMedium
ComplexityIntermediate
See full spec → Best for: Strong trend days after news / open drives
Inverse FVG NQ · 5m · replay #103
Retracement Large impulse move

50 of 5005

Pure geometry — no structural overlays required. Sentio measures the largest recent impulse leg on the execution timeframe and marks the 50% retracement zone. When price returns to that zone and prints a rejection candle, the model fires. Works on any liquid instrument.

  • Largest impulse leg over the lookback window
  • 50% retracement zone marked at ½(high + low)
  • Return to the zone with a rejection candle
  • Entry off the zone; target the impulse high (or low)
Avg R+1.2
Fire rateMedium
ComplexityIntro
See full spec → Best for: Wide-range sessions, post-news, session-extreme drives
50 of 50 NQ · 5m · replay #104
Closed-loop learning

Every model gets scored. Losers auto-mute.

A model on Sentio isn't just a spec — it's a spec with a live scoreboard. Every triggered setup is walked bar-by-bar to real T1/T2/stop outcomes, aggregated per model / timeframe / symbol, and surfaced in your Analytics tab. When a model starts bleeding, the scanner shuts it up before you lose more.

01

Model-native R

Wins and losses are scored on the price you would have gotten — not on invented backtest numbers. A 50-of-50 T1 win scores 1.25R; a Forever T2 win might score 4R; a stop-out is always −1R. Geometry honored automatically.

02

Rolling win-rate guardrail

Three losses in a row, OR a win rate below 40% over the last six or more setups today — either one mutes the model until the next session boundary (18:00 ET). Only reduces exposure, never enables anything new.

03

Analytics that tell you why

Per-model, per-timeframe, per-symbol table with win rate, average R, best and worst outcomes, total R. Rows under 5 resolved setups render muted — no noise pretending to be signal.

04
Coming soon

AI parameter proposals

Weekly job hands the last month of triggered outcomes to Sentio AI and gets back tightening/loosening suggestions, each backed by evidence from your own trades. You review and click Accept or Reject — never auto-applied.

Roadmap · shipping regularly

Five today. More on the way.

The model roster grows every few weeks. Each new model goes through the same pipeline as the built-in five — spec published, parameters documented, replay-validated on live NQ/ES tape before it ever reaches your scanner. When a new one ships, it just shows up in your model list.

Get early access Or build your own
Custom & community templates

Or start from a template someone else already tuned.

Pick a template, inspect the compiled spec, clone it into your workspace. Every template is replay-validated before it lands in the library.

London Sweep → NY Continuation

Triggers14 Avg R+1.6 Timeframes15m · 5m

"When ES sweeps the London high on 15m and NQ fails to confirm, wait for a bearish displacement and enter on a 5m FVG retracement during New York."

  1. 1. smt_divergence bearish, ES→NQ
  2. 2. liquidity_sweep london_high, 15m, ES only
  3. 3. displacement down, 5m, min_impulse ≥ 1.2× ATR
  4. 4. fvg retest ≤ 6 bars
  5. 5. entry session ny

Post-CPI Reversal

Triggers8 Avg R+2.1 Timeframes1m · 5m

"On CPI release, fade the initial spike after the first 5m candle closes. Enter on inversion FVG retest with a stop above the spike high."

  1. 1. news_event impact: high, calendar: CPI
  2. 2. time_of_day 08:30 ET ±5m
  3. 3. displacement either direction, ≥ 2.0× ATR
  4. 4. inverse_fvg on the spike candle
  5. 5. entry fade the impulse

PDH/PDL Failure

Triggers19 Avg R+1.3 Timeframes5m

"Price tags the previous day's high or low, prints a rejection candle, and forms an inversion FVG. Fade the tag with a stop just beyond the wick."

  1. 1. key_liquidity reference: PDH, PDL
  2. 2. liquidity_sweep wick beyond reference
  3. 3. rejection_close close inside prior range
  4. 4. inverse_fvg on failure candle
  5. 5. entry fade

Silver Bullet 10–11am

Triggers11 Avg R+1.5 Timeframes1m · 5m

"Only fire inside the 10:00–11:00 ET killzone. Look for a sweep of a prior swing followed by a 1m FVG retracement in the opposite direction."

  1. 1. time_of_day 10:00–11:00 ET, mon–fri
  2. 2. liquidity_sweep prior swing, any timeframe
  3. 3. displacement opposite direction, 1m
  4. 4. fvg retest ≤ 3 bars, 1m
  5. 5. entry within killzone only

ES / NQ SMT Divergence

Triggers22 Avg R+1.4 Timeframes5m · 15m

"When ES makes a new high and NQ doesn't (or vice versa), that's SMT divergence. Wait for a displacement candle in the direction of the weaker index and enter on the retest."

  1. 1. smt_divergence ES ↔ NQ, 15m
  2. 2. market_structure CHoCH on weaker index, 5m
  3. 3. displacement weaker-index direction, ≥ 1.5× ATR
  4. 4. fvg retest ≤ 5 bars
  5. 5. entry weaker-index direction

Opening Range Break

Triggers17 Avg R+1.2 Timeframes5m · 15m

"Define the opening range as the first 30 minutes of the NY session. On a clean break + retest of the opening range boundary with a displacement candle, enter in the break direction."

  1. 1. session_window ny_open, first 30m
  2. 2. opening_range high/low of the window
  3. 3. displacement break of range, ≥ 1.3× ATR
  4. 4. retest range boundary ≤ 4 bars
  5. 5. entry break direction

None of these fit? Build your own.

Describe the setup in plain English. Sentio compiles it into a spec, replay-validates it, and drops it into the scanner alongside the built-in five — with more shipping every month.

See the model builder Get started