Σ
Futures Screener Top 10 by market cap · live
checking… no recent signals sound on ——:——:—— UTC · ——:—— BKK candle ——:—— refresh 60s
Portfolio total (informational only, pots never pooled) —
Paper Trading · Mean Reversion · $1,000 pot —
loading…
—
loading…
Entry — all four required, same 1h candle
  1. 1RSI(14) on the 1h chart is below 36 (oversold) — raised from 30 on 2026-08-27 after validating across three non-overlapping windows: win rate improved in all three (58→68%, 48→64%, 63→79%) with ~22% more signals
  2. 2Candle volume is at least 2× the trailing 20-candle average (not just the previous candle)
  3. 3The coin's move this candle is a statistical outlier vs. the other 9 coins (z-score ≤ -1.0) — it fell meaningfully more than "typical" this hour, not just in absolute terms
  4. 4Not a market-wide flush — if 4+ of the 10 coins are oversold at once, every entry is suppressed that cycle
Exit — whichever hits first
  1. 1Price moves 8% against the position (wide catastrophic backstop — deliberately not a tight stop, since tight stops were found to hurt this strategy family)
  2. 2Price moves 5% in favor of the position (take-profit — locks in the bounce instead of relying solely on RSI recovering)
  3. 348 hours held, regardless of price (time backstop)
  4. 4RSI recovers to 65 or higher (the bounce actually happened)
Signal Status —
—
—
Watchlist10
Avg RSI—
Oversold (<36)—
Feed Errors0
Closest to Signal ranked by RSI, lowest first
loading…
Paper Trading · Mean Reversion Short · $1,000 pot —
loading…
—
loading…
Not backtest-validated — wired in 2026-08-10 by explicit override after failing out-of-sample validation (recent 58d: +$9.86, OOS 58-116d ago: -$23.31, reversed sign). Kept running to keep collecting real data, same as this project's original two strategies once were.
Entry — all four required, same 1h candle (mirrors Long, inverted)
  1. 1RSI(14) on the 1h chart is above 70 (overbought)
  2. 2Candle volume is at least 2× the trailing 20-candle average (not just the previous candle)
  3. 3The coin's move this candle is a statistical outlier vs. the other 9 coins (z-score ≥ +1.0) — it rose meaningfully more than "typical" this hour, not just in absolute terms
  4. 4Not a market-wide melt-up — if 4+ of the 10 coins are overbought at once, every entry is suppressed that cycle
Exit — whichever hits first
  1. 1Price moves 8% against the position — i.e. price rises 8% (wide catastrophic backstop, same magnitude as Long)
  2. 2Price moves 5% in favor — i.e. price falls 5% (take-profit)
  3. 348 hours held, regardless of price (time backstop)
  4. 4RSI falls back to 40 or lower (the reversion actually happened)
Signal Status —
—
—
Watchlist10
Avg RSI—
Overbought (>70)—
Feed Errors0
Closest to Signal ranked by RSI, highest first
loading…
Paper Trading · Momentum · $1,000 pot —
loading…
—
loading…
Entry — all four required, same 1h candle
  1. 1Close is above the 50-period SMA (an established uptrend, not a blip)
  2. 2Close is a new 20-period high (breakout — the classic trend-following entry)
  3. 3Candle volume is at least 1.5× the trailing 20-candle average (a lower bar than mean-reversion's 2× — breakouts don't need as extreme a spike)
  4. 4Not 02:00, 08:00, or 22:00 UTC — backtested at 0.0% win rate in two separate non-overlapping windows (31 combined trades, zero wins), new entries paused during these hours
Exit — whichever hits first (trailing, not a fixed target — "let winners run")
  1. 1Price moves 10% against entry (wide catastrophic backstop)
  2. 2Price pulls back 4% from the highest close reached since entry (trailing stop)
  3. 3Close drops back below the 50-period SMA (trend broken)
  4. 47 days held, regardless of price (time backstop — longer than mean-reversion's 48h, since trend-following rides longer moves)
Signal Status —
—
—
Watchlist11
Trending (>SMA50)—
At New High—
Feed Errors0
Closest to Breakout ranked by distance below the 20-period high
loading…
Paper Trading · Momentum Short · $1,000 pot —
loading…
—
loading…
RETIRED 2026-08-17 — no new positions. Wired in 2026-08-10 despite failing out-of-sample validation (recent 58d: -$33.00, OOS 58-116d ago: +$28.32, reversed sign), kept running to collect real data. That data confirmed the failure: 18 closed trades, 22.2% win rate, -$3.55 realized. Retired with 0 open positions — nothing left to unwind. Figures below are the final frozen record.
Entry — all three required, same 1h candle (no entry-hour gate — never separately validated for this side)
  1. 1Close is below the 50-period SMA (an established downtrend, not a blip)
  2. 2Close is a new 20-period low (breakdown — the mirror of Long's breakout entry)
  3. 3Candle volume is at least 1.5× the trailing 20-candle average
Exit — whichever hits first (trailing, mirrors Long)
  1. 1Price moves 10% against entry — i.e. price rises 10% (wide catastrophic backstop, same magnitude as Long)
  2. 2Price pulls back 4% from the lowest close (trough) reached since entry — i.e. rises 4% off the trough (trailing stop)
  3. 3Close rises back above the 50-period SMA (trend broken)
  4. 47 days held, regardless of price (time backstop, same as Long)
Signal Status —
—
—
Watchlist11
Trending Down (<SMA50)—
At New Low—
Feed Errors0
Closest to Breakdown ranked by distance above the 20-period low
loading…
Paper Trading · BB Reversion · $1,000 pot —
loading…
—
loading…
Entry — all four required, same 1h candle
  1. 1Close is at or below its lower Bollinger Band (20-period mean minus 2.5× standard deviation)
  2. 2Candle volume is at least 2× the trailing 20-candle average
  3. 3The coin's move this candle is a statistical outlier vs. the other 9 coins (z-score ≤ -1.0)
  4. 4Not a market-wide flush — if 4+ of the 10 coins are BB-oversold at once, every entry is suppressed that cycle
Exit — whichever hits first
  1. 1Price moves 8% against the position (wide catastrophic backstop)
  2. 272 hours held, regardless of price (time backstop)
  3. 3RSI(14) recovers to 65 or higher (the bounce actually happened — the same recovery mechanism Mean Reversion uses, reused deliberately)
Paper Trading · BB Reversion Short · $1,000 pot —
loading…
—
loading…
RETIRED 2026-08-23 — no new positions. Wired in 2026-08-14 despite failing out-of-sample validation (recent 58d: -$0.06, OOS 58-116d ago: -$22.63, both net negative), kept running to collect real data. That data confirmed the failure: 7 closed trades, 0.0% win rate, -$35.61 realized, 71% of exits were stop-loss. Figures below are the final frozen record.
Entry — all four required, same 1h candle (mirrors Long, inverted)
  1. 1Close is at or above its upper Bollinger Band (20-period mean plus 2.5× standard deviation)
  2. 2Candle volume is at least 2× the trailing 20-candle average
  3. 3The coin's move this candle is a statistical outlier vs. the other 9 coins (z-score ≥ +1.0)
  4. 4Not a market-wide melt-up — if 4+ of the 10 coins are BB-overbought at once, every entry is suppressed that cycle
Exit — whichever hits first (mirrors Long)
  1. 1Price moves 8% against entry — i.e. price rises 8% (wide catastrophic backstop, same magnitude as Long)
  2. 272 hours held, regardless of price (time backstop, same as Long)
  3. 3RSI(14) falls back to 35 or lower (the reversion actually happened)
Signal Status —
—
—
Watchlist10
Overbought (touching band)—
Feed Errors0
Closest to Signal ranked by distance below the upper Bollinger Band
loading…
Shared Constraints applies to all six strategies
All six strategies (Mean Reversion, Mean Reversion Short, Momentum, Momentum Short, BB Reversion, BB Reversion Short): 5% risk per trade, max 1 open position per coin per strategy — six fully separate $1,000 pots, results never pooled, so which one actually has an edge stays visible. The three Short strategies are unvalidated overrides — see their own Strategy Rules panels for the exact backtest numbers that failed.
Traded pairs: Binance USDT-margined perpetual futures only — not USDC, not USD1, not coin-margined/inverse contracts
BTCUSDT · ETHUSDT · BNBUSDT · XRPUSDT · SOLUSDT · TRXUSDT · HYPEUSDT · DOGEUSDT · ZECUSDT · XMRUSDT
Momentum and Momentum Short: exclude BTCUSDT and DOGEUSDT — both backtested as consistent net-negative for Momentum specifically (not Mean Reversion) across two non-overlapping windows. Mean Reversion and Mean Reversion Short still trade all 10 pairs above.
Momentum and Momentum Short only, also trade SUIUSDT · UNIUSDT · LINKUSDT (added 2026-08-24) — backtested as consistent net-positive contributors specifically for Momentum across two non-overlapping windows, both when added individually and combined. Not tested or traded by any other strategy.