Strategy Arena — Full Fleet Performance
TOTAL STRATEGIES
73
LIVE ACTIVE
60
BACKTESTABLE
12
BACKTEST DATA
923,739
8.80y of klines
MC PATHS
4,000
COMPUTED
10d ago
STRATEGY FAMILIES (hover for description)
BTC-5m CoreBTC-15mBTC-MartingaleBTC-StreakBTC-QuantumBTC-HolderAGENT-BrainClaude-AgentClaude-ScannerCopy-BotTelegramAGENT-Sports
How to read this page
LIVE columns — what each bot has actually done in production, from its paper_account state (lifetime stats, no time window). N = bets, WR = win rate, ROI = % return on starting bankroll, PnL = realized $.
BACKTEST columns — for BTC-price-only strategies, simulated against 9 years of actual Bitcoin price history (923k Binance 5-minute bars). 9Y = win rate across the full history. 24M = last two years only (the regime we're actually in). PnL @ $0.48 assumes realistic market-maker fill prices (mid is $0.50, optimistic is $0.45).
Δ column — gap between backtest 24M WR and live WR. Large negative Δ means the bot is underperforming the math by a lot (probably execution / latency / wrong logic). Large positive Δ means it's beating the backtest (small sample variance or genuinely better logic).
⚠ Caveat: Copy bots and PM-data strategies can't be backtested here — only their LIVE performance is shown. The "MA-revert" hint applies only to the BTC family.
Master Table — every strategy in the fleet
| STRATEGY | STATUS | LIVE N | LIVE WR | LIVE ROI | LIVE PNL | BT 9Y | BT 24M | BT PNL | Δ vs BT |
|---|---|---|---|---|---|---|---|---|---|
| BTC-5m Core (2 strategies) | N: 166 | WR: 42.2% | Subtotal PnL: $-136 | ||||||
|
BTC_5M_BOT
Original BTC 5m up/down — signal-driven directional bet
|
— | 0 | 0.0% | +0.0% | $+0 | 55.9% | 53.1% | $+185,369 | — |
|
CLAUDE_BTC_5M
Claude-built BTC 5m variant
|
STALE | 166 | 42.2% | -2.7% | $-136 | 55.9% | 53.1% | $+185,369 | -10.9pp |
| BTC-Quantum (6 strategies) | N: 387 | WR: 47.8% | Subtotal PnL: $-2,770 | ||||||
|
BAYES_COLD
Cold-start Bayesian variant
|
— | 0 | 0.0% | +0.0% | $+0 | 56.9% | 54.3% | $+40,776 | — |
|
CONFLUENCE_FADE
Fade only on multi-signal confluence
|
— | 0 | 0.0% | +0.0% | $+0 | 56.9% | 54.3% | $+40,776 | — |
|
CRISIS_BAYES_SKEW
288-cell Bayesian posterior + dynamic fill-price EV gate + quarter-Kelly
|
— | 0 | 0.0% | +0.0% | $+0 | 56.9% | 54.3% | $+40,776 | — |
|
BAYES_SIZE
Bayesian size-aware variant
|
STALE | 46 | 45.7% | -2.1% | $-107 | 56.9% | 54.3% | $+40,776 | -8.6pp |
|
QUANTUM_BRAIN
Weighted ensemble of 6 sub-signals (streak + holder + price + time)
|
STALE | 282 | 49.6% | -9.6% | $-957 | 55.1% | 52.7% | $+91,958 | -3.0pp |
|
CRISIS_EXIT
Exit-discipline variant of CRISIS
|
STALE | 59 | 40.7% | -22.7% | $-1,706 | 56.9% | 54.3% | $+40,776 | -13.6pp |
| BTC-Holder (2 strategies) | N: 517 | WR: 51.8% | Subtotal PnL: $+505 | ||||||
|
HOLDER_CON
Fade holder skew — bet OPPOSITE smart money
|
STALE | 306 | 52.6% | +9.4% | $+468 | — | — | — | — |
|
HOLDER_PRO
Follow holder skew — bet UP if smart-money skew>+0.3
|
STALE | 211 | 50.7% | +0.7% | $+37.07 | — | — | — | — |
| BTC-Streak (3 strategies) | N: 147 | WR: 47.6% | Subtotal PnL: $-167 | ||||||
|
STREAK_FADE
Bet OPPOSITE the current streak (mean-revert)
|
STALE | 63 | 55.6% | +3.5% | $+176 | 54.1% | 52.6% | $+650,011 | +3.0pp |
|
STREAK_FADE_K4
Bet OPPOSITE the current streak (mean-revert)
|
STALE | 31 | 41.9% | -2.4% | $-118 | — | — | — | — |
|
STREAK_RIDE
Bet WITH the current streak (momentum)
|
STALE | 53 | 41.5% | -4.5% | $-225 | 47.1% | 48.4% | $-250,019 | -6.9pp |
| BTC-Martingale (2 strategies) | N: 608 | WR: 48.4% | Subtotal PnL: $+669 | ||||||
|
MART_PURE
Pure doubling — $5→$5120 on losses, reset on win
|
STALE | 473 | 47.8% | +7.1% | $+708 | 50.4% | 50.1% | $+1,093,396 | -2.4pp |
|
ANTI_MART
Same as Paroli — confirmed losers via backtest
|
STALE | 135 | 50.4% | -0.8% | $-38.81 | 48.5% | 49.7% | $+192,256 | +0.6pp |
| BTC-15m (2 strategies) | N: 0 | WR: 0.0% | Subtotal PnL: $+0 | ||||||
|
BTC_15M_ARB
Pair-arbitrage on the 15m markets
|
— | 0 | 0.0% | +0.0% | $+0 | — | — | — | — |
|
BTC_15M_BOT
BTC 15m up/down, same signal model
|
— | 0 | 0.0% | +0.0% | $+0 | — | — | — | — |
| AGENT-Brain (4 strategies) | N: 208 | WR: 21.2% | Subtotal PnL: $-788 | ||||||
|
AGENT_NOTHING_HAPPENS
Fade contrarian — bets nothing-happens markets
|
STALE | 70 | 80.6% | +0.0% | $+144 | — | — | — | — |
|
AGENT_HOLD_TO_SETTLE
Hold the position through to market resolution
|
STALE | 35 | 33.3% | -8.5% | $-68.64 | — | — | — | — |
|
AGENT_EDGE_NET_KELLY
Edge-net Kelly sizing — only fires when EV after costs is +ve
|
STALE | 27 | 46.2% | -1.8% | $-84.16 | — | — | — | — |
|
AGENT_LAMBDA_TAIL
Tail-risk pricer — fires on extreme price moves
|
STALE | 76 | 22.9% | +7.9% | $-780 | — | — | — | — |
| AGENT-Sports (1 strategies) | N: 21 | WR: 38.1% | Subtotal PnL: $+129 | ||||||
|
AGENT_SPORTS_FLB
Favorite-longshot bias on sports markets
|
STALE | 21 | 66.7% | +0.2% | $+129 | — | — | — | — |
| Claude-Scanner (5 strategies) | N: 90 | WR: 28.9% | Subtotal PnL: $-367 | ||||||
|
CLAUDE_RESOLUTION_SNIPER
Snipe near-certain resolutions <30 min out
|
STALE | 4 | 100.0% | +0.0% | $+20.58 | — | — | — | — |
|
CLAUDE_PMXT_ARB
Cross-venue arbitrage scanner (PM x other venues)
|
— | 0 | 0.0% | +0.0% | $+0 | — | — | — | — |
|
CLAUDE_SPORTS_MIRROR
Vegas-line vs PM divergence
|
— | 0 | 0.0% | +0.0% | $+0 | — | — | — | — |
|
CLAUDE_CONSENSUS_PROB
Bet vs PM consensus when probability diverges
|
STALE | 19 | 46.2% | -0.3% | $-156 | — | — | — | — |
|
CLAUDE_MAKER
Maker engine — quote both sides for spread
|
STALE | 67 | 47.5% | -0.5% | $-231 | — | — | — | — |
| Claude-Agent (16 strategies) | N: 1,724 | WR: 40.5% | Subtotal PnL: $-2,183 | ||||||
|
CLAUDE_ENSEMBLE
Ensemble of multiple Claude sub-strategies
|
PAUSED 17d | 84 | 80.6% | -3.3% | $+380 | — | — | — | — |
|
CLAUDE_SHARP_FOLLOW
Follow sharp wallet entries
|
PAUSED 17d | 156 | 72.2% | -0.4% | $+264 | — | — | — | — |
|
CLAUDE_PRE_RESOLVE
Pre-resolution arbitrage (mothballed)
|
— | 0 | 0.0% | +0.0% | $+0 | — | — | — | — |
|
CLAUDE_DEEP_TAIL
Buy very low-priced tails ($0.05-$0.15)
|
STALE | 17 | 38.5% | -0.0% | $-21.09 | — | — | — | — |
|
CLAUDE_MID_VOLUME
Mid-volume PM trader copies
|
PAUSED 17d | 190 | 67.5% | +0.4% | $-59.03 | — | — | — | — |
|
CLAUDE_QUALITY_KELLY
Kelly sizing on quality-filtered signals
|
PAUSED 17d | 123 | 66.7% | +0.2% | $-71.82 | — | — | — | — |
|
CLAUDE_VOL_FADE
Fade high-volatility moves
|
STALE | 23 | 61.5% | +1.4% | $-117 | — | — | — | — |
|
CLAUDE_WHALE_PRECISION
Tight-precision copy of whale activity
|
PAUSED 17d | 196 | 77.3% | -10.2% | $-145 | — | — | — | — |
|
CLAUDE_REVERSAL_HUNTER
Hunt PM sign-reversal patterns
|
STALE | 33 | 46.2% | -1.1% | $-148 | — | — | — | — |
|
CLAUDE_SHARP_PROVEN
Copy only the most-proven sharp wallets
|
PAUSED 17d | 101 | 63.3% | +0.9% | $-177 | — | — | — | — |
|
CLAUDE_WHALE_LOOSE
Loose-precision copy of whale activity
|
PAUSED 17d | 265 | 70.3% | -16.6% | $-208 | — | — | — | — |
|
CLAUDE_HIGH_TAIL
Aggressive tail pricer
|
STALE | 64 | 78.3% | +49.0% | $-253 | — | — | — | — |
|
CLAUDE_FRESH_FLIP
Buy after fresh price flips
|
STALE | 91 | 60.0% | +25.2% | $-262 | — | — | — | — |
|
CLAUDE_FAST_REACT
Low-latency reaction to PM signals
|
PAUSED 17d | 189 | 68.4% | -2.1% | $-385 | — | — | — | — |
|
CLAUDE_TAIL_DISCIPLINED
Tail-pricer with strict entry discipline
|
PAUSED 17d | 69 | 63.6% | -0.6% | $-470 | — | — | — | — |
|
CLAUDE_SHARP_CONVERGENCE
Multiple sharp wallets converging on same bet
|
PAUSED 17d | 123 | 69.0% | -2.6% | $-509 | — | — | — | — |
| Copy-Bot (15 strategies) | N: 1,695 | WR: 26.2% | Subtotal PnL: $-9,600 | ||||||
|
COPY_CONSENSUS
Copy when multiple sharp wallets agree
|
STALE | 103 | 67.9% | -3.0% | $+365 | — | — | — | — |
|
COPY_FADE
FADE the most-followed wallets (contrarian)
|
STALE | 33 | 75.0% | -11.1% | $-87.74 | — | — | — | — |
|
COPY_HIGH_PRICE
Copy buys at price > $0.80 (resolution near-certain)
|
STALE | 10 | 100.0% | -4.1% | $-164 | — | — | — | — |
|
COPY_SHARP
Copy the most-proven sharp wallets
|
STALE | 70 | 59.4% | -10.1% | $-170 | — | — | — | — |
|
COPY_WHALE
Copy high-notional whale trades only
|
STALE | 46 | 56.0% | -7.3% | $-188 | — | — | — | — |
|
COPY_HOTHAND_FAST
Copy currently-hot wallets
|
STALE | 24 | 23.1% | -10.0% | $-472 | — | — | — | — |
|
COPY_HOTHAND
Copy currently-hot wallets
|
STALE | 18 | 30.0% | +0.6% | $-495 | — | — | — | — |
|
COPY_VOLUME_1K
Copy trades over $1k notional
|
STALE | 43 | 36.4% | +2.5% | $-570 | — | — | — | — |
|
COPY_BIG_TICKET
Copy big-ticket whale entries >$5k
|
STALE | 83 | 51.1% | +1.9% | $-657 | — | — | — | — |
|
COPY_SPORTS_QUALITY
Filtered sports-market copies
|
STALE | 265 | 53.2% | +38.2% | $-700 | — | — | — | — |
|
COPY_LOW_PRICE
Copy buys at price < $0.15 (longshots)
|
STALE | 151 | 33.3% | +24.6% | $-1,180 | — | — | — | — |
|
COPY_NEW_LISTING
Copy on fresh market listings
|
STALE | 226 | 48.4% | +30.1% | $-1,180 | — | — | — | — |
|
COPY_CRYPTO_QUALITY
Filtered crypto-market copies
|
STALE | 141 | 43.4% | +5.1% | $-1,348 | — | — | — | — |
|
COPY_FAST_QUALITY
High-frequency quality copies
|
STALE | 272 | 45.1% | +31.7% | $-1,361 | — | — | — | — |
|
COPY_LATE_RESOLVE
Copy late-resolution-window entries
|
STALE | 210 | 50.0% | +38.1% | $-1,392 | — | — | — | — |
| Telegram (4 strategies) | N: 0 | WR: 0.0% | Subtotal PnL: $+0 | ||||||
|
TG_CONSENSUS
Telegram-signal-driven trader
|
— | 0 | 0.0% | +0.0% | $+0 | — | — | — | — |
|
TG_FADE
Telegram-signal-driven trader
|
— | 0 | 0.0% | +0.0% | $+0 | — | — | — | — |
|
TG_LOOSE
Telegram-signal-driven trader
|
— | 0 | 0.0% | +0.0% | $+0 | — | — | — | — |
|
TG_TIGHT
Telegram-signal-driven trader
|
— | 0 | 0.0% | +0.0% | $+0 | — | — | — | — |
| Misc (11 strategies) | N: 664 | WR: 51.2% | Subtotal PnL: $+734 | ||||||
|
FADE_KELLY
Uncategorized strategy
|
STALE | 182 | 54.9% | +19.0% | $+949 | — | — | — | — |
|
HOLDER_FADE_EXTREME
Uncategorized strategy
|
STALE | 25 | 68.0% | +8.5% | $+423 | — | — | — | — |
|
MEGA_K8
Uncategorized strategy
|
STALE | 2 | 100.0% | +13.1% | $+392 | — | — | — | — |
|
FADE_4
Uncategorized strategy
|
STALE | 29 | 62.1% | +5.9% | $+295 | — | — | — | — |
|
FADE_3
Uncategorized strategy
|
STALE | 65 | 55.4% | +3.2% | $+160 | — | — | — | — |
|
ULT5_MEGA_SNIPER
Uncategorized strategy
|
STALE | 9 | 55.6% | +1.8% | $+90.10 | — | — | — | — |
|
QUANTUM_GATE
Uncategorized strategy
|
STALE | 38 | 52.6% | +1.3% | $+64.99 | — | — | — | — |
|
default
Uncategorized strategy
|
STALE | 120 | 49.2% | -0.3% | $-14.84 | — | — | — | — |
|
ULT1_TIERED_PAROLI
Uncategorized strategy
|
STALE | 53 | 43.4% | -2.7% | $-79.73 | — | — | — | — |
|
ULT4_REGIME_ADAPT
Uncategorized strategy
|
STALE | 59 | 40.7% | -5.7% | $-284 | — | — | — | — |
|
ULT2_EDGE_WEIGHTED
Uncategorized strategy
|
STALE | 82 | 43.9% | -12.6% | $-1,261 | — | — | — | — |
| FLEET TOTAL (73 strategies) | 6,227 | 39.31% | $-13,973 | ||||||
Top 6 Backtestable — Deep Dive
MART_PURE
BTC-Martingale — Pure doubling — $5→$5120 on losses, reset on win
By year (WR / n bets / pnl) — 9y backtest @ $0.48 fill
2017
55.3%
39,304 · $+146,852
2018
50.3%
104,327 · $+116,026
2019
50.3%
104,768 · $+116,014
2020
50.6%
105,155 · $+133,767
2021
49.7%
104,923 · $+84,035
2022
50.0%
105,120 · $+101,655
2023
50.2%
105,104 · $+114,990
2024
50.5%
105,408 · $+127,390
2025
50.1%
105,120 · $+107,993
2026
50.1%
44,509 · $+44,675
| FILL | HORIZON | TRADES | P5 | P25 | MEDIAN | P75 | P95 | P(profit) |
|---|---|---|---|---|---|---|---|---|
| $0.45 | 3mo | 25,920 | $+64,147 | $+68,801 | $+72,014 | $+75,339 | $+79,660 | 100.0% |
| $0.45 | 6mo | 51,840 | $+132,284 | $+139,154 | $+144,140 | $+148,960 | $+155,996 | 100.0% |
| $0.45 | 1yr | 105,120 | $+274,108 | $+284,690 | $+292,059 | $+299,427 | $+310,286 | 100.0% |
| $0.48 | 3mo | 25,920 | $+19,842 | $+24,154 | $+27,167 | $+30,076 | $+34,336 | 100.0% |
| $0.48 | 6mo | 51,840 | $+43,113 | $+49,451 | $+53,919 | $+58,490 | $+64,620 | 100.0% |
| $0.48 | 1yr | 105,120 | $+92,253 | $+103,058 | $+109,915 | $+117,032 | $+127,421 | 100.0% |
| $0.5 | 3mo | 25,920 | $-6,857 | $-2,767 | $+75.75 | $+3,068 | $+7,058 | 50.6% |
| $0.5 | 6mo | 51,840 | $-10,272 | $-4,437 | $+102 | $+4,590 | $+10,575 | 50.7% |
| $0.5 | 1yr | 105,120 | $-15,647 | $-6,171 | $+512 | $+7,295 | $+16,871 | 52.3% |
LIVE comparison:
473 bets · WR 47.8% · ROI +7.1% · Realized $+708
vs backtest 24M WR 50.1%
STREAK_FADE
BTC-Streak — Bet OPPOSITE the current streak (mean-revert)
By year (WR / n bets / pnl) — 9y backtest @ $0.48 fill
2017
44.4%
11,296 · $-21,777
2018
55.3%
23,296 · $+86,585
2019
56.2%
22,686 · $+94,770
2020
56.8%
22,060 · $+99,148
2021
54.6%
23,740 · $+79,433
2022
55.7%
23,089 · $+91,085
2023
53.9%
23,807 · $+71,680
2024
53.3%
24,935 · $+67,532
2025
52.2%
25,451 · $+54,373
2026
53.1%
10,547 · $+27,181
| FILL | HORIZON | TRADES | P5 | P25 | MEDIAN | P75 | P95 | P(profit) |
|---|---|---|---|---|---|---|---|---|
| $0.45 | 3mo | 6,243 | $+21,934 | $+24,206 | $+25,702 | $+27,308 | $+29,524 | 100.0% |
| $0.45 | 6mo | 12,487 | $+45,782 | $+49,051 | $+51,489 | $+53,705 | $+57,140 | 100.0% |
| $0.45 | 1yr | 25,321 | $+95,355 | $+100,563 | $+104,275 | $+107,987 | $+113,417 | 100.0% |
| $0.48 | 3mo | 6,243 | $+10,834 | $+12,964 | $+14,366 | $+15,925 | $+18,054 | 100.0% |
| $0.48 | 6mo | 12,487 | $+23,409 | $+26,473 | $+28,707 | $+30,889 | $+34,214 | 100.0% |
| $0.48 | 1yr | 25,321 | $+49,986 | $+54,973 | $+58,402 | $+61,674 | $+66,765 | 100.0% |
| $0.5 | 3mo | 6,243 | $+4,173 | $+6,118 | $+7,565 | $+8,961 | $+10,807 | 100.0% |
| $0.5 | 6mo | 12,487 | $+10,067 | $+12,910 | $+15,105 | $+17,249 | $+20,392 | 100.0% |
| $0.5 | 1yr | 25,321 | $+22,732 | $+27,470 | $+30,712 | $+34,203 | $+38,841 | 100.0% |
LIVE comparison:
63 bets · WR 55.6% · ROI +3.5% · Realized $+176
vs backtest 24M WR 52.6%
ANTI_MART
BTC-Martingale — Same as Paroli — confirmed losers via backtest
By year (WR / n bets / pnl) — 9y backtest @ $0.48 fill
2017
52.7%
39,304 · $+94,332
2018
48.4%
104,327 · $+16,805
2019
47.7%
104,768 · $-24,090
2020
46.8%
105,155 · $-70,232
2021
48.2%
104,923 · $+4,710
2022
48.0%
105,120 · $-6,137
2023
47.9%
105,104 · $-14,152
2024
49.5%
105,408 · $+73,572
2025
49.9%
105,120 · $+94,590
2026
49.1%
44,509 · $+22,856
LIVE comparison:
135 bets · WR 50.4% · ROI -0.8% · Realized $-38.81
vs backtest 24M WR 49.7%
CLAUDE_BTC_5M
BTC-5m Core — Claude-built BTC 5m variant
By year (WR / n bets / pnl) — 9y backtest @ $0.48 fill
2017
52.8%
1,888 · $+4,540
2018
55.9%
5,063 · $+20,542
2019
59.0%
4,724 · $+26,731
2020
60.5%
4,382 · $+28,112
2021
56.4%
5,594 · $+24,046
2022
56.1%
5,204 · $+21,640
2023
55.4%
5,317 · $+20,166
2024
54.6%
5,516 · $+18,515
2025
54.0%
5,924 · $+18,029
2026
50.5%
2,490 · $+3,049
| FILL | HORIZON | TRADES | P5 | P25 | MEDIAN | P75 | P95 | P(profit) |
|---|---|---|---|---|---|---|---|---|
| $0.45 | 3mo | 1,449 | $+4,551 | $+5,660 | $+6,380 | $+7,155 | $+8,263 | 100.0% |
| $0.45 | 6mo | 2,899 | $+10,131 | $+11,682 | $+12,790 | $+14,009 | $+15,560 | 100.0% |
| $0.45 | 1yr | 5,880 | $+21,701 | $+24,195 | $+25,967 | $+27,796 | $+30,400 | 100.0% |
| $0.48 | 3mo | 1,449 | $+2,061 | $+3,048 | $+3,775 | $+4,450 | $+5,489 | 100.0% |
| $0.48 | 6mo | 2,899 | $+4,875 | $+6,434 | $+7,473 | $+8,564 | $+10,070 | 100.0% |
| $0.48 | 1yr | 5,880 | $+11,337 | $+13,675 | $+15,285 | $+16,948 | $+19,233 | 100.0% |
| $0.5 | 3mo | 1,449 | $+483 | $+1,480 | $+2,129 | $+2,877 | $+3,874 | 98.0% |
| $0.5 | 6mo | 2,899 | $+1,889 | $+3,335 | $+4,332 | $+5,380 | $+6,826 | 99.7% |
| $0.5 | 1yr | 5,880 | $+5,119 | $+7,313 | $+8,760 | $+10,406 | $+12,650 | 100.0% |
LIVE comparison:
166 bets · WR 42.2% · ROI -2.7% · Realized $-136
vs backtest 24M WR 53.1%
BTC_5M_BOT
BTC-5m Core — Original BTC 5m up/down — signal-driven directional bet
By year (WR / n bets / pnl) — 9y backtest @ $0.48 fill
2017
52.8%
1,888 · $+4,540
2018
55.9%
5,063 · $+20,542
2019
59.0%
4,724 · $+26,731
2020
60.5%
4,382 · $+28,112
2021
56.4%
5,594 · $+24,046
2022
56.1%
5,204 · $+21,640
2023
55.4%
5,317 · $+20,166
2024
54.6%
5,516 · $+18,515
2025
54.0%
5,924 · $+18,029
2026
50.5%
2,490 · $+3,049
| FILL | HORIZON | TRADES | P5 | P25 | MEDIAN | P75 | P95 | P(profit) |
|---|---|---|---|---|---|---|---|---|
| $0.45 | 3mo | 1,449 | $+4,551 | $+5,715 | $+6,491 | $+7,211 | $+8,319 | 100.0% |
| $0.45 | 6mo | 2,899 | $+9,964 | $+11,682 | $+12,901 | $+13,953 | $+15,560 | 100.0% |
| $0.45 | 1yr | 5,880 | $+21,812 | $+24,250 | $+26,023 | $+27,685 | $+30,455 | 100.0% |
| $0.48 | 3mo | 1,449 | $+2,009 | $+3,048 | $+3,775 | $+4,450 | $+5,437 | 100.0% |
| $0.48 | 6mo | 2,899 | $+5,031 | $+6,486 | $+7,473 | $+8,512 | $+10,018 | 100.0% |
| $0.48 | 1yr | 5,880 | $+11,233 | $+13,675 | $+15,285 | $+16,844 | $+19,233 | 100.0% |
| $0.5 | 3mo | 1,449 | $+483 | $+1,480 | $+2,179 | $+2,827 | $+3,775 | 98.3% |
| $0.5 | 6mo | 2,899 | $+1,889 | $+3,335 | $+4,283 | $+5,330 | $+6,776 | 99.8% |
| $0.5 | 1yr | 5,880 | $+4,869 | $+7,214 | $+8,810 | $+10,356 | $+12,600 | 100.0% |
QUANTUM_BRAIN
BTC-Quantum — Weighted ensemble of 6 sub-signals (streak + holder + price + time)
By year (WR / n bets / pnl) — 9y backtest @ $0.48 fill
2017
40.5%
2,034 · $-8,097
2018
57.5%
2,684 · $+13,056
2019
58.5%
2,337 · $+12,588
2020
60.7%
2,122 · $+13,911
2021
58.5%
2,819 · $+15,239
2022
56.3%
2,685 · $+11,420
2023
56.1%
2,853 · $+11,792
2024
55.7%
3,108 · $+12,222
2025
52.8%
3,366 · $+8,110
2026
50.3%
1,498 · $+1,719
| FILL | HORIZON | TRADES | P5 | P25 | MEDIAN | P75 | P95 | P(profit) |
|---|---|---|---|---|---|---|---|---|
| $0.45 | 3mo | 836 | $+2,092 | $+2,923 | $+3,533 | $+4,087 | $+4,918 | 100.0% |
| $0.45 | 6mo | 1,672 | $+4,849 | $+6,179 | $+7,010 | $+7,841 | $+9,115 | 100.0% |
| $0.45 | 1yr | 3,392 | $+10,825 | $+12,764 | $+14,094 | $+15,479 | $+17,363 | 100.0% |
| $0.48 | 3mo | 836 | $+710 | $+1,490 | $+2,009 | $+2,529 | $+3,256 | 99.3% |
| $0.48 | 6mo | 1,672 | $+1,940 | $+3,135 | $+3,914 | $+4,693 | $+5,940 | 100.0% |
| $0.48 | 1yr | 3,392 | $+4,966 | $+6,784 | $+8,083 | $+9,330 | $+11,148 | 100.0% |
| $0.5 | 3mo | 836 | $-202 | $+546 | $+1,045 | $+1,544 | $+2,292 | 91.8% |
| $0.5 | 6mo | 1,672 | $+294 | $+1,342 | $+2,090 | $+2,888 | $+3,985 | 96.7% |
| $0.5 | 1yr | 3,392 | $+1,434 | $+3,180 | $+4,277 | $+5,524 | $+7,219 | 99.2% |
LIVE comparison:
282 bets · WR 49.6% · ROI -9.6% · Realized $-957
vs backtest 24M WR 52.7%
Methodology & Caveats
The backtest model. Each 5-minute Binance bar is treated as one Polymarket BTC up/down market. Outcome = "Up" if close ≥ open (matching Polymarket's >= tiebreaker rule). Each strategy emits Up/Down/skip based on streak structure visible BEFORE bar i+1. Win = bet matched outcome.
Fill-price scenarios. $0.50 = theoretical mid (no edge); $0.48 = realistic market-maker fill once you've earned queue position; $0.45 = optimistic, only achievable if you're consistently first to react before PM mid moves.
Monte Carlo projections. Beta-binomial posterior over recent (24-month) WR with effective prior of 30 observations centered on 55%. 4,000 paths per (strategy × fill × horizon). Bands show 5% / 25% / median / 75% / 95% final-$PnL percentiles.
The live-vs-backtest gap. Our regime monitor today found STREAK_FADE at 36.8% WR live vs 55.87% backtested — a ~19pp gap. Sources of slippage: (1) adverse selection — fills land on the side OTHER bots are eager to take; (2) queue position — at $0.45 only 10-30% of orders fill; (3) the live bot may use slightly different logic than the backtest decider. Apply at least a 2-3pp haircut to every backtest WR. Worse, sometimes, in low-volume regimes.
What can't be backtested here. Copy bots (need wallet trade history), sports markets (need Vegas-line history), Claude scanner agents (need PM-specific data). For those strategies only LIVE columns are populated — back-test columns show "—".