Trade Workflow · For Traders

From candle to trade decision,
in three honest gates.

This system finds momentum breakout candles on Binance altcoins and runs each one through a rigorous 3-step validation before it earns your capital. This doc walks you through exactly what happens to one signal — and why each gate exists.

Coins scanned
300+ USDT pairs
Validation gates
3 sequential
Methods tested
72 per signal
Bias
False-neg over false-pos
00EXPLAIN IT LIKE I'M NEW

If you've never traded before, start here.

Imagine you're at a casino, but it's one where you can study every past hand before you bet. You see a certain pattern keep appearing. Sometimes the hand wins, sometimes it loses. You want to know: over hundreds of past hands that looked just like this, did betting on it make money?

That's all this system does — but with cryptocurrency price candles instead of cards.

IN 3 LINES
1. The scanner watches 300+ crypto coins and notices big "breakout candles" the moment they happen.
2. For each one, it looks at every similar candle in that coin's past and simulates 72 different trade strategies — "if I had entered here, with this stop, aiming for this profit, would I have made money?"
3. If the past says yes and a trained AI model agrees and on-chain data isn't against you, it gives you a green light with exact entry, stop, and target prices.

What's a "momentum candle"?

A single price bar where a lot happened in one direction — the body (open-to-close) is much bigger than usual, and volume was much higher than average. It's the market shouting. Most of the time, after a shout, price either keeps going (trend continuation) or fakes out and reverses. This system tries to tell you which one you're looking at.

The four words you need to know

ENTRY
Where you buy
The price you actually enter the trade at.
STOP-LOSS
Where you quit
If price hits this, you exit with a small, accepted loss.
TAKE-PROFIT
Where you win
If price hits this, you exit with a profit. Often you take some off at TP1, rest at TP2.
R (risk unit)
Your loss size
"+2R" means you made twice what you were willing to lose. Math, not luck.
RULE #1
You decide how much you're willing to lose before the trade. Everything else — stop size, position size, profit target — is math from there. This system does that math for you.
01ANATOMY OF A TRADE

A breakout, a pullback, and three possible endings

Below is exactly what happens in every trade this system gives you. Click Play and watch. The animation loops through the three possible outcomes — WIN, LOSS, and BREAKEVEN.

SCENARIO: WIN — price runs to TP
SL ENTRY TP1 (+1R) TP2 (+2.5R) SIGNAL ENTRY FILLED
Price breaks out with a strong momentum candle. It pulls back to your entry zone (38.2% retrace), fills your limit order, then continues up to hit TP1 at +1R. You take 50% off. The remaining runner climbs to TP2 at +2.5R for a full win.

Reading the picture above

  • Signal candle (green) — the big breakout bar that triggers the scanner.
  • Blue dot — your entry price. On a Standard zone, price has to pull back to fill it.
  • Red dashed line (SL) — where you quit if price goes against you. Fixed loss.
  • Yellow dashed line (TP1) — first profit target at +1R. Often you close 50% here.
  • Green dashed line (TP2) — full profit target at +2.5R (or 2.0 / 3.0 depending on the method).
  • Moving white line — the actual price path through the trade.
NOTICE
In the BREAKEVEN scenario, price hits TP1, you close half in profit, then it reverses all the way back to your entry. Your stop was moved to entry when TP1 hit — so you exit the second half flat. Net result: small profit on the first half, zero on the second. This is what "NEUTRAL" means in the backtest.
02THREE REAL CASE STUDIES

What the whole workflow looks like end-to-end

Here are three illustrative walkthroughs showing how different signals end up with different verdicts. Numbers shown are representative of real output from the system.

CASE 01 · CLEAN WIN
SOLUSDT · 4H · LONG
TRADE — TOOK IT

Thursday afternoon. A 4-hour candle on SOL closes with 87% body and 5.1× volume on an ADX reading of 34. Regime score is 88/100 (strong GREEN). The scanner surfaces it.

Step 1 · Backtest

ANALOGS
142
STRICT 70% filter
CAND A
+0.64R
Standard · ATR · Partial-NoBE · 2.5R
CAND B
+0.64R
Same method — UNANIMOUS
WFO
PASS
OOS PF 1.8 · 5/5 windows

Step 2 · ML probability

78% (RF + isotonic, 142 analogs, all 11 features inside normal range for this setup).

Step 3 · AI verdict

"Unanimous candidate with 78% calibrated ML on a strong GREEN regime. OOS PF CI [1.4, 2.3] comfortably above 1. Pulse Intel is +9 (BULLISH: SOL CEX outflows + TVL tailwind). Take the trade."

Execution & outcome

→ Limit @ $148.20 (38.2% retrace) · Risk 1%
→ SL @ $144.90 (ATR-based)
→ TP1 @ $151.50 (+1R) · TP2 @ $156.45 (+2.5R)
# 6h later: TP1 hit, 50% closed, SL keeps original position
# 14h later: TP2 hit on runner
RESULT: +1.75R total · ~1.75% account gain
LESSON
Unanimous + strict filter + ML ≥ 70% + strong regime = highest-conviction setup. When all three gates agree strongly, size normally and let it run.
CASE 02 · SMART SKIP
REZUSDT · 4H · LONG
SKIP — DIDN'T TAKE

REZ looks amazing on paper. 93% body, 6.8× volume, parabolic trend. Emotionally it screams "buy now or miss it." Let's see what the system said.

Step 1 · Backtest — the trap appears

ANALOGS
38
LOOSE 20% — only 38 found
CAND A (raw)
WR 95%
Looks perfect...
HONEST PF
1.0
After NEUTRAL exclusion
FILL RATE
22%
Standard zone never filled
WHY IT FAILS
Raw backtest says WR=95% because on trending coins like REZ, Partial+BE management produces a flood of +0.498R "NEUTRAL" outcomes that look like wins on paper. Once those are excluded honestly, the real edge is gone. Fill rate of 22% means 78% of the time price just ran without pulling back — survivor bias.

Step 2 · ML

62% — but the filter is LOOSE and only 38 analogs exist. The probability is directional only.

Step 3 · AI verdict

"Despite raw headline numbers, three red flags: (1) LOOSE filter with n=38 is below statistical reliability, (2) honest PF = 1.0 means no actual edge after NEUTRAL correction, (3) fill rate 22% indicates severe survivor bias on the Standard zone. WFO verdict INSUFFICIENT (n_oos=3). Skip."

LESSON
The most valuable output of the system is the word SKIP. Parabolic moves that "can't miss" are exactly where retail traders blow accounts. The ratchet filter, NEUTRAL labeling, and fill rate honesty are the three guardrails that caught this one.
CASE 03 · BORDERLINE TRADE
AVAXUSDT · 1D · SHORT
TRADE · SMALL SIZE

AVAX prints a bearish daily candle. 78% body, 2.8× volume. Regime is mixed (YELLOW, 54/100). Candidates disagree. What do we do?

Step 1 · Backtest — A and B disagree

🟢 CAND A
Newest bucket
58%
METHOD
Aggressive · Fixed · Simple · 2.0R
EVw
+0.22R
🔵 CAND B
Weighted all-time
54%
METHOD
Standard · ATR · Trailing · 3.0R
EVw
+0.18R

WFO: BORDERLINE (n_oos = 6, OOS PF 1.18 · CI [0.85, 1.62]). Filter RELAXED 45%.

Step 2 · ML

ML(A) = 61%, ML(B) = 56%. Both above threshold but not high conviction.

Step 3 · AI verdict

"Candidate A has the edge on newest-bucket EVw and higher ML. WFO is BORDERLINE — edge plausible but noisy. Pulse Intel is NEUTRAL (+1). Take the trade at half size and exit on TP1 if reached; don't hold for TP2 without fresh confirmation."

Execution & outcome

→ Market short @ $24.10 · Risk 0.5% (half normal)
→ SL @ $24.46 (fixed 1.5%)
→ TP @ $23.38 (+2R)
# 2 days later: TP1 hit, closed full position per AI guidance
RESULT: +1R total · ~0.5% account gain
LESSON
Not every trade is A+. Borderline setups still have edge — but you take them smaller and respect the AI's exit guidance. Most accounts grow on these B-grade trades, not the A+ home-runs.
PATTERN
Across hundreds of signals, these three shapes repeat: the clean high-conviction win, the smart skip that saves you from a trap, and the borderline trade that you take smaller. The 3-gate system's job is to tell them apart reliably.
03WHAT THIS SYSTEM DOES

It helps you say "no" to bad trades

Most breakouts on altcoins look great in the moment and fail by the next candle. The problem isn't finding breakouts — it's knowing which ones have historical edge right now, in this regime, on this coin.

So the system does three things:

SCAN
Finds breakouts
Across 300+ Binance pairs on your chosen timeframe
VALIDATE
Three gates
Backtest → ML → AI, each stricter than the last
DECIDE
Trade or skip
With an exact entry, SL, and TP plan attached
PRINCIPLE
The goal is not to catch every winner. The goal is to filter out losers with confidence. A signal that fails any of the three gates gets skipped — and the system tells you exactly why.
04WHY YOU CAN TRUST IT

Six things that make this different from backtested hopium

01 · CAUSAL FEATURES
No lookahead
Every feature used to score a signal is strictly past-looking. Audited — 20/20 features clean.
02 · PURGED VALIDATION
No leakage
Walk-forward testing uses de Prado's purge + embargo method. Training samples can't secretly peek at test bars.
03 · 5 WINDOWS, NOT 1
Distribution
Results come from 5 rolling out-of-sample windows — a distribution, not a lucky cut.
04 · FILL RATE HONESTY
No survivor bias
If a "perfect" method only fills 20% of the time, you'll see that — not a fake 90% win rate on the subset.
05 · NEUTRAL BAND
Flat ≠ win
Trades that hit TP1 then reverse to breakeven are called what they are — flat — not counted as wins.
06 · REGIME AWARE
Now matters
Historical analogs from the current market regime pull harder than analogs from a different regime.
05A SIGNAL IS BORN

The scanner finds a breakout candle

EXAMPLE
Let's follow one: ETHUSDT · 4H · LONG — a candle closes with 85% body, 4.2× average volume, ADX rising into strong territory.

Every few minutes, the scanner runs _scanner_score_signal on the last 3 closed candles of every USDT pair. If a candle meets the momentum threshold (big body, high volume, directional ADX), it becomes a candidate signal and shows up in the Scanner tab with a proposed trade plan.

What you see immediately

COIN
ETHUSDT
4H timeframe
DIRECTION
LONG ↗
Bullish momentum
BODY
85%
of total range
VOLUME
4.2×
vs 7-bar average
ADX(14)
31.4
Strong trend
REGIME
83/100
GREEN
GOOD
Strong-looking signal. But "looks strong now" is not edge. It has to survive the three gates before it becomes a trade.
GATE 1 · STATISTICAL
Backtest + Walk-Forward
STEP 1 BUTTON
"Has this type of signal actually made money in the past — on this coin, in this kind of market?"

What happens when you click Step 1

The system scans the coin's full 1000-bar history, finds all past candles that look like this one (same body/volume profile, ratcheting down if needed to get enough analogs), and simulates 72 different trade strategies on each of them.

Those 72 come from every combination of:

  • 3 entry zones — Aggressive (market), Standard (38.2% retrace), Sniper (61.8% retrace)
  • 2 stop-loss styles — fixed 1.5% or ATR-based
  • 4 management styles — Simple / Partial / Partial-NoBE / Trailing
  • 3 take-profit levels — 2.0R / 2.5R / 3.0R

3 × 2 × 4 × 3 = 72 strategies, each tested against the same set of historical analogs.

Two candidates surface

Instead of picking one "best" method, the system surfaces two:

🟢 CANDIDATE A
Newest-bucket winner
72%

What's working right now on this coin. Picks the best method over the most recent time-decay bucket.

METHOD
Standard · ATR · Partial · 2.5R
WIN RATE
61%
EVw
+0.48R
PF
1.9
🔵 CANDIDATE B
Weighted all-time winner
68%

What has worked across the full history, weighted toward recent trades. Captures long-run robustness.

METHOD
Aggressive · Fixed · Trailing · 3.0R
WIN RATE
54%
EVw
+0.41R
PF
1.7
UNANIMOUS
When A and B are the same method, that's called unanimous — a very strong signal. The method dominates both the recent window and the weighted all-time view.

Walk-forward test

Right after the backtest, the system runs a rolling walk-forward: 5 different in-sample/out-of-sample cuts at 50/60/70/80/90% of the data. Each window trains on the "past" half and tests on the "future" half it hasn't seen.

WFO VERDICT
PASS
8+ OOS trades
OOS PF
1.42
95% CI [1.08 – 1.88]
EDGE HIT
4/5
Windows with PF>1
HONEST PF
1.35
Excl. NEUTRAL
Why walk-forward matters FOR CURIOUS

A regular backtest looks at the whole history at once and can accidentally cheat — it's been optimized on the same data it's tested on. Walk-forward simulates trading forward in time: you can only use data from before each trade to make the decision.

The "purged" part means training samples are dropped if their outcome would have resolved inside the test window — that would be a hidden leak of future information.

"4/5 windows with PF > 1" is much stronger evidence than "aggregate PF = 1.4" because it shows the edge holds in multiple time slices, not just one lucky cut.

GATE 2 · MACHINE LEARNING
Adaptive Classifier
STEP 2 BUTTON
"Does a trained model, looking at this specific candle's features, think it wins?"

How the ML model is trained

The system collects every historical candle on this coin that looks like today's signal (using the same ratchet filter), and labels each one by whether it would have won using the Candidate A method. Then it trains a model on those features — and uses the model to score today's candle.

The model type adapts to how much data is available:

AnalogsModelWhy
< 20Heuristic ruleNot enough to train anything honest
< 50Logistic RegressionSimple, hard to overfit on tiny data
50–149Random ForestGood middle ground
≥ 150Gradient BoostingBest at learning complex patterns
CALIBRATED
When there's enough data (n ≥ 60), the model's probabilities are isotonic calibrated. Without calibration, a raw "68%" from Random Forest might really mean 53%. With it, "68%" actually means 68%.

What comes out

ML · CANDIDATE A
Probability of WIN
71%
ANALOGS USED
94
FILTER
RELAXED 45%
CV FOLDS
5 · purged
MODEL
RF + isotonic
ML · CANDIDATE B
Probability of WIN
63%
ANALOGS USED
94
FILTER
RELAXED 45%
CV FOLDS
5 · purged
MODEL
RF + isotonic
What "RELAXED 45%" means FILTER BADGE

If your current signal is a 85%-body, 4× volume candle, the system ideally wants to train on historical analogs that are similarly explosive — say, at least 70% of that body and volume.

But very explosive signals often have few exact analogs. So the system ratchets the threshold down progressively — 70% → 55% → 45% → 35% → 25% → 20% — until it finds enough training samples. Every step is labeled:

STRICT 70% Tight analogs. Trust the probability number.
RELAXED 45% Looser match. Probability is still informative but wider.
LOOSE 20% Very broad analogs. Treat probability as directional only.

GATE 3 · AI REASONING
Dual-Candidate Verdict
STEP 3 BUTTON
"Given everything — backtest, WFO, ML, regime, on-chain — should we actually take this trade?"

The final gate sends a structured package to a reasoning model (Groq's gpt-oss-120b by default). It sees the full diagnostic picture and produces a verdict for both candidates, then picks a winner.

What the AI receives

  • Full signal characteristics (body, volume, ADX, regime)
  • Both candidates' backtest stats + fill rate + decay buckets
  • WFO verdict with OOS PF and confidence interval
  • Both ML models' calibrated probabilities
  • Canonical execution prices (it cannot make up numbers)
NO HALLUCINATION
Entry, SL, and TP prices are precomputed and injected into the prompt with a strict "copy verbatim" instruction. The AI reasons about whether to trade — not what the numbers are.

The verdict you get

FINAL VERDICT
ETHUSDT 4H LONG · Apr 18, 2026
TRADE · CANDIDATE A
Winner rationale:
Candidate A wins on three counts: (1) ML probability is 8pp higher (71% vs 63%), (2) OOS PF confidence interval stays above 1.0 across all 5 WFO windows, and (3) regime similarity to current conditions is 0.84 avg. Candidate B's Trailing management underperforms in high-ADX regimes like this one. Fill rate is a concern (Standard zone fills 47% of the time) but the ratchet-relaxed sample set has 94 analogs — enough to be directional. Pulse Intel adds +6 (bullish TVL delta on ETH).
When the AI says SKIP COMMON REASONS
  • Zones invalid — big-body candle pushed retrace entry below the structural stop-loss. No trade possible.
  • Fill rate too low — Standard/Sniper zones fill <40% on trending coins. Survivor bias warning.
  • OOS PF CI straddles 1.0 — edge isn't statistically distinguishable from random.
  • ML probability < 55% — model doesn't see this candle as a winner.
  • Regime mismatch — historical analogs came from very different regimes; signal weak.
09THREE ENTRY ZONES

Every signal gives you three ways in

Each zone is a different trade-off between fill rate and reward-to-risk.

🟢 Aggressive
0% retrace
Market entry at signal close. Fills almost always, but your stop is further from entry.
Fill rate~95%
R:R (avg)1.7
Best forStrong trends
🟡 Standard
38.2% retrace
Wait for a pullback to the 38.2% Fibonacci level of the signal candle body. Middle ground.
Fill rate~55%
R:R (avg)2.3
Best forChoppy trends
🔴 Sniper
61.8% retrace
Deep retracement entry. Best R:R but often never fills on trending coins.
Fill rate~30%
R:R (avg)3.1
Best forMean-reversion
ZONE INVALID
Sometimes a zone becomes structurally invalid — the retrace entry price falls below the stop-loss. The system rejects these zones automatically. It's not a bug. Don't try to "widen the stop" — that changes your strategy into something you haven't backtested.
10FOUR MANAGEMENT STYLES

How you manage the trade once filled

Simple
Full size, hold to TP2 or original SL. No partials, no breakeven. The "set and forget" option — cleanest but biggest drawdowns.
Partial
Take 50% off at +1R, move SL to breakeven on the rest. Lowest drawdown but many "flat" outcomes when TP1 hits then reverses.
Partial-NoBE
Take 50% off at +1R, keep original SL. Real downside stays. Often the strongest expected value on trending alts.
Trailing
Full size, move to breakeven at +1R, then trail 0.5×ATR from the close. Catches the big runners but gives back some gains.
PICKED FOR YOU
Don't guess. The 72-method backtest tells you which style has the highest EVw for this specific signal. That's what Candidate A and B represent — the backtest already picked the best management mode for you.
11EXECUTION CHECKLIST

Your five-point checklist before placing the order

01
Verdict is TRADE
AI picked a winning candidate. If it says SKIP, skip.
02
Filter badge is STRICT or RELAXED
If LOOSE, treat ML as directional only, size smaller.
03
WFO verdict PASS or BORDERLINE
INSUFFICIENT (n<5) means statistically you're flying blind.
04
Fill rate ≥ 40% on the chosen zone
Below that, you'll miss most of the signals you think you'll catch.
05
Pulse Intel not STRONGLY BEARISH
On-chain flow against your direction is a veto, especially for longs.

If all five check out: place the limit order at the zone price, SL at the computed stop, TP1 at +1R (if Partial or Partial-NoBE), TP2 at the method's target R multiple. Size for your risk budget — never more than 0.5–1% per trade.

12WHAT THE BADGES MEAN

Reading the honesty signals

BadgeWhat it meansYour response
STRICT 70%Training analogs closely match current signalTrust probability numerically
RELAXED 45%Analogs are looser matchesTreat as directional, size normal
LOOSE 20%Very broad analog setDirectional only, size smaller or skip
PASS WFO8+ OOS trades, clean edgeGreen light from walk-forward
BORDERLINE5–7 OOS tradesEdge plausible but noisy
INSUFFICIENT<5 OOS tradesStatistical blind spot — skip or wait
UNANIMOUSCandidate A = Candidate BStrongest signal the system gives
NEUTRAL 12%12% of past trades landed ±0.30R (flat)Expect more "flat" than wins would suggest
13ON-CHAIN PULSE

A free Nansen-lite layer over your signal

The Pulse tab blends three independent data sources into one composite score from −15 (strongly bearish) to +15 (strongly bullish):

FLOW · 40%
CEX net flow
Large transfers in/out of exchanges. Outflow = accumulation = bullish. Inflow = distribution = bearish. ERC-20 via Etherscan; SPL via Solscan.
TVL · 35%
DeFi lockup
Total Value Locked 24h + 7d delta from DefiLlama. Money flowing into a protocol is a real-world vote.
SOCIAL · 25%
LunarCrush
Galaxy Score + sentiment + alt rank. Retail attention and tone, weighted less than hard money.

A macro modifier (BTC.D delta + stablecoin supply trend) adds ±3 on top. Final score is ±15.

ScoreVerdictWhat to do
+10 to +15STRONGLY BULLISHConfirms long entries, cautious with shorts
+4 to +9BULLISHTailwind for longs
−3 to +3NEUTRALPulse is not a factor — trade the structure
−4 to −9BEARISHHeadwind for longs
−10 to −15STRONGLY BEARISHVeto for longs — wait or flip short
14CHEAT SHEET

The whole workflow on one screen

# 1 · SCANNER finds a breakout candidate
ETHUSDT 4H LONG · body 85% · vol 4.2x · regime GREEN
# 2 · STEP 1 · Run backtest + WFO
72 methods tested on 94 analogs (filter RELAXED 45%)
→ Cand A = Standard · ATR · Partial · 2.5R · EVw +0.48R
→ Cand B = Aggressive · Fixed · Trailing · 3.0R · EVw +0.41R
→ WFO = PASS · OOS PF 1.42 [1.08, 1.88] · Edge hit 4/5
# 3 · STEP 2 · Train ML on chosen method
RF + isotonic · n=94 · purged 5-fold CV
→ ML(A) = 71% · ML(B) = 63%
# 4 · STEP 3 · AI verdict
→ Winner = CANDIDATE A (TRADE)
→ Pulse = +6 (BULLISH, TVL tailwind)
# 5 · EXECUTE
Limit @ Standard zone · SL ATR · Partial 50% @ 1R · TP2 @ 2.5R
Size = 0.5–1% of capital
REMEMBER
The three gates exist to say "no" faster than you would emotionally. Most profitable trading careers end because of trades that should not have been taken. This system's highest-value output is the word SKIP.
Momentum Candle Backtester · Trade Workflow · v1.0