Candlestick Chart — Live
Multi-Chart
Stream Candles — Gold
| Candle (min) | B.Open | B.High | B.Low | B.Close | A.Close | Vol | Ticks |
|---|
Stream Candles — Silver
| Candle (min) | B.Open | B.High | B.Low | B.Close | A.Close | Vol | Ticks |
|---|
Stream Candles — US500
| Candle (min) | B.Open | B.High | B.Low | B.Close | A.Close | Vol | Ticks |
|---|
Backfill a Day
Pull a whole day of 1-minute candles from IG history. Each day is logged so it won't be pulled twice (it spends the weekly allowance). Tick Force to re-pull a logged day.
Fill Gaps
Detect missing minutes and pull only the gaps. Market-closed holes cost ~nothing; you spend allowance only on genuinely-missing candles.
Backfill Log
| When | Day | Epic | Res | Points | Upserted | Allowance Left | Status |
|---|
Strategy
Entry next bar; TP at reward × risk. Spread on: buy fills at ask, sell at bid; buffer adds extra slippage points.
Trend filters — gate entries (all selected must agree)
Tick any combination. A long only fires when every ticked filter reads up (or "trending" for strength gates); short mirrors. These are independent of the EMA10/20 filter above.
Date range scopes the test to a period — e.g. an up/down trend. Empty = last 30 days.
Trades
| # | Entry | Dir | Entry px | SL | TP | Exit | Exit px | Size | R | P&L $ | Result |
|---|
📌 Current take — research notes
updated 2026-07-06 · gold · ~27d in-sample⚠ Read the drawdown honestly. The Max DD% column below is realized (closed-trade only) and understates reality. The True DD in the Buckets table marks open positions to market and is ~1.5–2× higher for anything that stacks positions. Max concurrent positions is the real risk driver — a strategy showing 13% DD but holding 15 positions at once really draws ~25%.
⚠ Regime note (Jul 3–6): the last ~4 sessions were hostile to the decorrelators — stoch-er collapsed (29% win, +3%, Sharpe 0.04; bleeding >2 weeks) and zimp-3m-adx softened (+29%, true DD 10.3%). Every U-bucket's true DD roughly doubled vs the 07-03 numbers (U5+clv 14.2→26.6%). The clv family held up best (clv-5m positive in every trailing window).
Best standalone → clv-5m (+87%, true DD 8.8%, Sharpe 0.37, live). pbreak-er still the Sharpe leader (0.39, ~8% DD). Old anchor zimp-3m-adx is now mid-pack — keep for diversification, not as the star.
Best 100–150% return → mom-adapt-15m (+116%, Sharpe 0.38). But its true DD is ~22% (holds up to 14 positions). Big return in ~4 weeks structurally needs high concurrency — there is no 100%-return-at-low-DD.
Best buckets: the decorrelators are zimp (low-corr to all), stoch-er (negatively correlated with most) and pin-15m; the best trend-slot filler is now clv-5m (Round 22 CLV run — the highest Ret/DD single in the table; ~0 corr with stoch-er/pin-15m). Best low-DD trio → F1-clv: clv + zimp + emax-er (+217%, true DD 11.7%, 9 open — beats F1-stairs on both DD and efficiency). Best Ret/DD → U5+clv: all of U5 + clv (+531%, true DD 14.2%, Ret/TrueDD 37 — beats U6 on every axis, and its true DD is only ~2.5pts above the trio's). U5 (+273%, 13.7%) remains the calmer wide bucket. All U-, M-, F- and clv-bucket components are executor-wired — start each component's ex-* preset (hover a bucket row for the exact list).
Don't trust returns literally. These are compounded ~3-week in-sample figures, not forecasts — use them to compare, not predict. Sort by Real R/DD or Sharpe, not raw Return%.
| Goal | Pick | Return | True DD | Ret/DD | Open | Live |
|---|---|---|---|---|---|---|
| Best standalone | clv-5m (CLV run) | +87% | 8.8% | 9.9 | 8 | live |
| Best low-DD trio | F1-clv clv+zimp+emax-er | +213% | 14.5% | 14.7 | 9 | live |
| Best fully-live bucket | DIVERSE-3b mom-struct+zimp+pbreak | +341% | 18.1% | 18.8 | 15 ⚠ | live |
| Max efficiency (partial) | B5 Balanced-6 | +484% | 18.0% | 26.8 | 11 | engulf-er unwired |
| Current live bucket | CUR5 U5+clv minus stoch | +446% | 17.7% | 25.2 | 11 | live |
| Upgrade candidate 🆕 | CUR5 + pin-30m (htf hunt) | +586% | 13.2% | 44.2 | 12 | live ⚠ new |
| Best 100–150% single | mom-adapt-15m | +116% | ~22% | 5.3 | 14 | live |
Ret/DD = Return ÷ True DD (higher = better bang-per-drawdown). All picks above are fully executor-wired — start each component's ex-* preset from the dropdown.
Presets / Strategy library
Stats are at 1% capital risk per trade (Return %, Max DD % — compounded over the ~22-day in-sample window; ⚠ = under 30 trades, low confidence). Return%/Ret/DD = idealized (sequential); Real%/Real R/DD = overlap-aware — what you'd realistically get with concurrent positions, matching the Backtest page). Hover a preset name for its full setup. Archive hides it from the Backtest/Positions dropdowns; Load opens it in the Backtest tab.
| Preset | Strategy | TF | Win% | Return% | Real% | Max DD% | Ret/DD | Real R/DD | Sharpe | Comment |
|---|
Diversified Strategy Buckets
Running several different strategies on one account (each at 1% risk/trade), over the ~22-day in-sample window. Real DD = realized (closed-trade) drawdown — optimistic. True DD = honest mark-to-market drawdown including the floating P&L of every concurrently-open position — the number that matters. Max open = worst simultaneous position count (≈ % of the account at risk at once); a high value means real concurrent leverage. ⚠ = watch the concurrency. In-sample only — the components are all gold trend-followers, so their drawdowns are partly correlated (they can pile into losers together).
| Bucket | Components | Return% | Real DD% | True DD% | Ret/TrueDD | Max open | Live |
|---|
🏜️ Dust Data Backtest — full-history results
Regime-gated champions
Strategies behind the round-29 regime gate (1h ATR ≥ 30× spread AND daily |px−EMA20| ≥ 1.4 ATRs) — the only configurations that hold up across eras. Stats at 1% risk/trade, spread on.
| Strategy | TF | Gate | Results by window | Notes |
|---|
All presets × history windows (raw, ungated)
Overlap-aware Real% at 1% risk per trailing window. Verdict: survivor = positive on 1y+2y+3y · regime-only = positive in-sample, dead long-run · bust = negative everywhere. −100% = compounding wipeout. Click headers to sort.
Gold — order
Silver — order
Open Positions
| Instrument | Strategy | Dir | Size | Open | SL (pt) | TP (pt) | Current | P/L (pts) | P/L (est) | Status |
|---|
Position History
Closed deals (realized P/L) for the LIVE account, newest first.
| Closed | Instrument | Strategy | Dir | Size | Open | Close | P/L |
|---|
Strategy Executor
0 runningRun multiple strategies in parallel, server-side 24/7 — each keeps trading when you close the tab/log out and resumes after a deploy. Pick a preset (or bucket) + sizing and ▶ Start; each slot runs independently (forceOpen, so positions don't net) and is stopped individually below.