Honest tiers. Real capital, paper, backtest — labeled.
Systems beat gut feeling.
$178,000 on the screen and I didn't sell. That's when I knew: I need bots that don't panic when I would — and don't FOMO when I shouldn't.
Live-vs-Backtest experiment: Four momentum bots (Tactician 2.0, Rotator, Tri-Rotator, Hedge Hopper) run in parallel for 45 days to verify backtest predictions live — see live ≠ backtest article. Scout retired today — post-mortem.
Every backtest on this site is a bet on regime continuity. We test on the past — the future is a regime draw we haven't fully observed. Walk-forward catches what we've seen; it cannot catch what hasn't happened yet. Why this matters →
The Watchdog
The only bot trading real money. Conservative: sits in cash about 60% of the time, buys BTC only when multiple safety filters agree we're in a confirmed bull phase.
- ✓Beat HODLHODL means simply buying and holding — no trading. Someone who HODLs BTC buys once and never sells. Our strategies must beat this baseline, otherwise why bother with the complexity? in 7 out of 10 years
- ✓Beat HODL in all 3 historical market regimesA regime (or 'market phase') is a longer stretch of time with similar market behavior — e.g. bull market (rising), bear market (falling), sideways. A good strategy works in multiple regimes, not just one.
- ✓Worst drawdownThe worst loss from a peak to the subsequent trough. If the bot falls from $20,000 to $11,600 that's a -42% drawdown. Note: not the loss at the end, but how deep it went in between.: only -53% (vs. HODL: -77%)
- ✓About 1 trade per year — low frequency, high conviction
Waiting for buy trigger. Conditions not yet aligned.
Not the highest backtest return on this page — but the most consistent across all market regimes. That's why this is the bot trusted with real money. Exact size of the real allocation is kept private; the return % shown is the actual live performance.
The Surfer
A grid bot with a brain. Surfs the small chop waves, but paddles out before the tsunami — switches to just-holding-BTC when a confirmed bull rally starts, then comes back in when the chop returns.
- ✓Reads market regime with one simple rule: 20-day average vs 100-day average
- ✓In confirmed bull markets → just holds BTC for the full ride (no grid bot caps the upside)
- ✓In chop and bear → runs the grid bot with 30% BTC / 70% cash (more buying power for dips)
- ✓Backtest on 2.7y of BTC: +228% vs just-holding +159% — and with smaller worst-case drop (-33% vs -50%)
- ✓Cross-validated on 4.8y of ETH (incl. the 2022 bear): same edge holds. Out-of-sample test passed.
GRID mode (30% BTC / 70% USDT, tight15 range around $75,693). SMA20 ($71,253) is -4.3% vs SMA100 ($74,485). Bull regime requires SMA20 > SMA100 × 1.02; currently below threshold. Surfer is paddling, harvesting chop, no tsunami in sight.
Live virtual tracking starts today on $10,000. After 3 months of clean live equity curve, we'll consider a small real-money pilot. The first hybrid bot in our lineup that earned its slot via simulated alpha — see the full Surfer review.
The Sentry
Sister bot to The Genius — uses two simple ideas at once: bet against the crowd when too many traders are on one side, and follow the money when it suddenly flows into a DeFi protocol.
- ✓Bets opposite of the crowd when way too many traders pile into one direction — and gets paid for it
- ✓Notices when money flows fast into a DeFi protocol — buys the token before the rest of the market catches on
- ✓Watches 12 well-known DeFi tokens every 15 minutes — finds 1-5 chances per day
- ✓First test on 3 years of data: one of the two ideas works very well, the other works modestly. Both run together.
Watching 12 DeFi protocols for funding extremes + TVL flow signals. No active position. Portfolio $10,000 (+0.0% since launch).
Brand new. Runs in parallel to The Genius for 60-90 days. Whichever bot performs better in live data graduates to real capital — the other gets retired.
The Alpha Hunter
Steps outside crypto into US stocks. Once a month, picks the 10 US stocks that gained the most over the last year and holds them. Income stream that doesn't care what BTC is doing.
- ✓Completely outside crypto — protects you when BTC has a bad year
- ✓Once a month: keep the 10 best-performing US stocks of the last 12 months
- ✓8-year test: +742% vs. just holding the S&P 500 +212%
- ✓Realistic test — includes the actual fees and slippage you'd pay in real life
Holding 9 US stocks (equal-weight): AMD, AVGO, C, CAT, FCX, GOOGL, INTC, NEM, ROKU. Portfolio $9,996 (-0.0% since launch). Next rebalance ~30 days from last.
Paper-traded, no real capital. The method rests on 30+ years of academic evidence (Jegadeesh & Titman).
The Rotator
Switches daily between BTC and ETH — holds whichever coin has the stronger 21-day trend. Sits in cash when both are falling.
- ✓One question, every day: is BTC or ETH stronger? → hold the winner
- ✓If both are trending down → cash (protects capital in crypto winter)
- ✓Beat HODL in all 3 historical market regimesA regime (or 'market phase') is a longer stretch of time with similar market behavior — e.g. bull market (rising), bear market (falling), sideways. A good strategy works in multiple regimes, not just one.
- ✓Works on many different time-window settings — not just one lucky tweak
Holding ETH. 21-day ETH +17.8%, vs BTC +14.0%.
8-year test: +3,020% vs. just holding BTC +621%. Held up in 3 out of 3 historical sub-periods — no single lucky window propping up the result.
The Tactician 2.0
Buys BTC when the price has been trending up for 30 days, sits in cash otherwise. Shrinks the position when the market gets wild.
- ✓Simple rule: BTC up over 30 days? → buy. Down? → cash
- ✓When price swings get wild: smaller position, less risk
- ✓8-year test: +1,126% vs. just holding BTC +603% — worst dip only -58%
- ✓Running head-to-head against Rotator, Tri-Rotator, and Hopper as a live comparison
Long BTC since 30-day momentum turned positive (currently +8.3%).
Backtest says: Tactician underperforms its rotation siblings. We're keeping it live for 45 days anyway — because backtest ≠ live.
The Hedge Hopper
Switches between BTC and a Gold Miners ETF (GDX). Takes advantage of the fact that crypto and gold cycles rarely fall together.
- ✓BTC is running → hold BTC. Gold is running → switch to Gold Miners
- ✓If both are falling → cash
- ✓6-year test: +5,532% vs. just holding BTC +610%, worst dip only -45%
- ✓Held up in 3 out of 3 historical sub-periods — beats holding BTC in all of them
Holding BTC. 40-day BTC +15.0%, vs GDX -3.7%.
Launched yesterday. The 2020-2024 gold run was historically unusual — future cycles may differ. We're watching 45 days of live data before drawing conclusions.
The Contrarian
Does exactly the opposite of what most trading books say. Buys when other traders are most excited, sells when they panic.
- ✓The classic textbook rule loses money on BTC (see our RSI-lies article) — so this bot just flips it.
- ✓8-year backtest: +711% vs. just holding BTC +473%
- ✓Worst drop along the way: -64% (vs. holding BTC: -77%)
- ✓Only ~11 trades in 8 years — almost no fees, very patient
In cash since RSI(14) dropped below 30.
Works because most traders follow the textbook — and lose. If the crypto market ever calms down and starts behaving like a normal market again, the textbook rule could start working and this bot would stop. 45-day live test in progress.
The Tri-Rotator
Like Rotator but with 3 coins: BTC, ETH, and SOL. Each day, holds the one with the strongest 30-day momentum.
- ✓3 coins instead of 2 → more chances to ride the current winner
- ✓SOL added massive extra upside during 2023-2024
- ✓Works across many different time-window settings — not just one lucky tweak
- ✓Honest catch: only 5.7 years of data, one market cycle. SOL can crash harder than BTC
Holding ETH. 30-day ranking: ETH +9.9%, BTC +8.3%, SOL -3.1%.
6-year test: +34,929% vs. just holding BTC ~+901%. But that's only one full market cycle of data — eye-popping numbers from a single window are weaker evidence than steady ones from many windows. Live verification in progress.
The Genius
Monitors 22 DeFi protocols for governance events (votes, upgrades) and trades on them. Completely independent of the BTC price.
- ✓Not tied to BTC — a second, uncorrelated source of return
- ✓Runs on Bybit Testnet: real execution logic, virtual money
- ✓Only graduates to real capital after 50+ testnet trades match the backtest
- ✓Scans DeFi events automatically, trades without human intervention
Watching 24 DeFi protocols on Snapshot + Tally for governance events. No active position right now.
No meaningful live track record yet. Clearly labeled Virtual/Testnet for full transparency.
10 Bots, 1 System
The ten bots aren't random. Each covers a different asset, signal, or risk level. The Surfer (regime-aware grid bot) just joined as bot #2. Four momentum bots run in parallel — backtest predicts Tactician 2.0 underperforms; live data verifies. Scout retired earlier.
Its cycle filter determines the base position. This one decision is responsible for 80% of overall performance. Few decisions, but the right ones.
Regime-aware grid bot. SMA20 vs SMA100 +2% rule: confirmed bull → just hold BTC; chop or bear → run grid (30% BTC / 70% USDT). Backtest 2.7y BTC: +228% vs HODL +159% with smaller drawdown (-33% vs -50%). Cross-validated on 4.8y ETH (incl. 2022 bear). First hybrid bot in the lineup — earned its slot via simulated alpha. Live paper-tracking starts today.
M-30 momentum + vol-scaled position sizing (target 60% portfolio vol). Backtest 8y: +1,126% vs HODL +603%, MaxDD -58%. Multi-metric panel ranks it last among the three momentum bots — kept live as a 45-day Live-vs-Backtest verification experiment.
Cross-asset momentum: every day compares 21-day momentum on BTC vs ETH and holds the stronger one (cash if both negative). About 47 rotations per year. Backtest 2018-2026: +3,020% vs BTC HODL +621%. Walk-Forward 3/3 windows beat HODL across a robust parameter plateau.
Three-asset momentum: BTC vs ETH vs SOL on 30-day lookback. Holds whichever has strongest positive momentum, cash if all three are negative. About 50 rotations per year. Backtest 2020-2026 (5.7 years, SOL data limit): +34,929% vs BTC HODL ~+901%. Walk-Forward 2/3, plateau across all lookbacks 10-30. Paper-traded. Caveat: single-cycle window + SOL outage tail-risk — monitored particularly closely.
Cross-asset momentum outside crypto: BTC vs GDX (Gold Miners ETF) on 40-day lookback. Captures the mismatch between crypto cycles and gold cycles (2022 crypto winter = 2022-2024 gold run). Backtest 2018-2026 (5.7 years): +5,532% vs BTC HODL +610%, MaxDD only -45%. Walk-Forward 3/3 windows beat HODL, robust plateau across lookbacks 10-90. Caveat: only one rotation cycle tested — monitored as 45-day Live-vs-Backtest experiment.
Textbook RSI (buy<30, sell>70) loses money on BTC — we documented it. The Contrarian flips it entirely: BUY above RSI 70 (the crowd's exhaustion = our momentum), SELL below 30 (the crowd's dip = our falling knife). 8-year backtest: +711% vs BTC HODL +473%, drawdown -64% vs -77%. 11 trades in 8 years. Paper-traded since today.
Completely uncorrelated to Bitcoin. While the BTC bots cover the Bitcoin market, The Genius scans 22 DeFi protocols for governance events. Currently running on Bybit testnet — virtual money, real execution. Graduates to real capital after 50+ testnet trades confirm the backtest.
The Sentry runs DeFi crowd-fade plus governance frontrun
Sister bot to The Genius. Runs two ideas in parallel: bet against the crowd when too many traders pile on one side, and follow the money when it suddenly flows into a DeFi protocol. Watches 12 well-known DeFi tokens every 15 minutes. First test on 3 years of data: one of the two ideas works very well, the other modestly — both run together. Paper-traded for 60-90 days alongside The Genius; whichever performs better in live data graduates.
The Alpha Hunter extends the system into US stocks
A completely independent return stream. Every month it rebalances into the 10 US stocks with the strongest 12-month momentum. In backtesting (2019-2026, out-of-sample, costs included): +742% vs S&P 500 +212%. Paper-traded portfolio, not real capital.
The result: Ten independent return sources across crypto, precious metals, and US stocks. The Surfer (regime-aware grid bot) is the latest addition — first hybrid bot that earned its slot via simulated alpha. Four momentum bots (Tactician 2.0, Rotator, Tri-Rotator, Hedge Hopper) run in parallel for 45-day Live-vs-Backtest verification — backtest predicts Tactician underperforms its rotation siblings; live data confirms or contradicts. Scout retired earlier. See methodology and post-mortems.
What does The Watchdog say right now?
Live signal, updated daily
Our radar currently says:
Stay away from Bitcoin.
For 189 days in cash
BTC now
$84,250
2 of 3 conditions not met. All 3 must show Yes for the bot to buy.
3 independent filters · 6 years of backtest data · No gut feeling
Frequently Asked Questions
What is a trading bot?
A program that automatically buys and sells based on fixed rules. No human presses the button. The bot decides based on data and executes the trade itself.
How exactly do the bots work?
Tier rule: PAPER and BACKTEST bots have fully open rules on their article pages (read and replicate). REAL CAPITAL bots (currently only The Watchdog) keep exact parameters in the newsletter to preserve the edge. You see the WHAT and WHY for every bot — the exact HOW is gated only where real money is at risk.
Can I follow the signals?
Yes. Newsletter subscribers get every signal change across all 10 running bots by email (BTC cycle filter, regime-aware grid harvester, BTC momentum + risk-parity sizing, BTC/ETH rotation, BTC/ETH/SOL rotation, BTC/GDX cross-asset rotation, US stock momentum, DeFi governance, contrarian crowd-fade). You decide what to do with it.
How much money is in the bots?
Only The Watchdog runs on real capital (exact amount is private; newsletter subscribers get the scale). The Surfer (just launched paper today, $10,000 virtual), Tactician 2.0, Rotator, Tri-Rotator, Hedge Hopper, Contrarian, Genius, Sentry and Alpha Hunter are PAPER (virtual money, real execution logic). Total: 10 live bots (1 real-capital, 9 paper). Scout was retired 2026-04-17 after multi-metric analysis — see /post-mortems.
Are the backtest numbers realistic?
Every number is based on 6 years of real historical data. Including fees, slippage, and realistic execution. Still: backtests don't guarantee future results.
What about bots that failed?
All documented on our post-mortems page. Scout was downgraded from live to BACKTEST-only when full-period numbers showed it loses to HODL. Alpha Hunter Light was retired the day it launched. The Watchdog+Tactician combinations (Options B and C) both failed walk-forward validation. See the full list →
We currently run 10 live bots.
1 on real capital (The Watchdog) and 9 paper-tracked with $10,000 virtual capital each. The Surfer 🏄 just joined today — a regime-aware grid bot that beat just-holding-BTC by +1.14% per month with lower drawdown in our backtests. Live tracking starts now. Real-money pilot considered after 3 months of clean live equity curve.
Every bot's status, equity curve, and signals are visible above. Newsletter subscribers get every signal change by email.
Don't miss a thing
Want to know when the bot buys?
Newsletter subscribers get notified at every signal change. Plus: weekly bot updates and performance reports. Free.
TheBot-Letter
Get the same signals our 10 live trading bots send our internal team — before they show up on the website.
- The 12 Trading Lies PDF38 pages · every common myth busted with 6 years of real BTC data · arrives instantly
- Bot signal accessEvery BUY/SELL across all 10 live bots · in your inbox the moment it happens
- Weekly performance reviewEquity curves · what worked · what didn't · zero marketing spin