I owe you an update on this one.
The old Scout was a breakout pattern hunter. It worked in backtesting on a narrow 2024-2025 window. Then I did what every honest researcher should do: I extended the backtest to 6 years. And the old Scout collapsed. A different market regime broke it.
So I retired the old breakout logic and rebuilt The Scout from scratch. This time with a different signal, and with the discipline of a proper 6-year test from day one.
The new Scout is called The Scout (Hash Rate Momentum). It is currently running in backtest-only mode. That status is visible on its card, because I don't want to lie to you about what it is.
Why Hash Rate Momentum
Bitcoin is secured by miners. Miners run ASICs, which cost money, electricity, and depreciation. When miners expand — bringing more hashrate online — it's because they believe mining is profitable. Profitability correlates with forward BTC price expectations.
Hash rate doesn't lead price perfectly. But it's a real-world signal from people with real capital at stake. Much less noisy than the usual TA indicators.
The Scout asks one question, once per day:
"Is the Bitcoin network hash rate trending up faster than price?"
If yes → miners are bullish → take a long BTC position. If no → step back. Let The Watchdog handle BTC allocation.
It's a momentum bot with an on-chain signal instead of a price signal.
What The 6-Year Backtest Shows
Here's the part I don't love, and won't hide.
The Scout Hash Rate Momentum over 2020-2026:
| Metric | Value |
|---|---|
| Total return | +575% |
| Win rate | 55% |
| Total trades | 173 (~29/year) |
| Max drawdown | -46% |
| Holding period | ~10 days average |
| Benchmark: BTC HODL | +954% |
Two uncomfortable truths:
1. The Scout loses to HODL over the full period. +575% is good, but HODL was +954%. In a strong Bitcoin bull, stepping aside costs you. Over 6 years of what was a very bullish BTC regime, The Scout underperformed pure HODL.
2. The drawdown is real. -46% max drawdown is meaningful. On a $10k paper portfolio, that means seeing it drop to $5,400 before recovering. Not fun.
Why am I keeping it on the site anyway?
Because the walk-forward analysis shows The Scout wins in the windows where HODL is weak and loses in the windows where HODL is strong. That matches the theory: hash rate momentum captures regime transitions, not bull trend extensions.
If the next 6 years look like the 2018-2022 window (where HODL only returned +63%), The Scout would likely outperform. If they look like 2020-2021 (rocket), The Scout will underperform. That's the trade-off.
The Scout is a regime-insurance bet, not a HODL-beater. And I label it that way.
How The Scout Decides
Every trading day:
- Pull Bitcoin hashrate (30-day moving average) from Blockchain.com
- Compute 30-day change in hashrate vs 30-day change in price
- If hashrate momentum > price momentum by a threshold → long BTC for up to 14 days
- Exit on target (+8%), stop-loss (-5%), or timeout (14 days)
That's it. Three parameters: the threshold, the target, the stop. All fixed since the walk-forward test concluded.
No Claude. No sentiment. No news. Just on-chain data from a free API.
Current Status
The Scout is labeled BACKTEST on the site. Here's what that means:
- The logic is fully coded and tested
- It runs in simulation on real-time hash rate data
- No broker connection, no paper portfolio tracked yet
- Equity curve shown is the 6-year backtest, not live tracking
Why not paper-trade it like The Alpha Hunter? Honestly: I'm still debating whether it earns a place in the line-up given the HODL underperformance. Paper trading commits me to tracking and reporting a bot I might retire. I'd rather be upfront about the label than pretend.
If I move it to Paper status, you'll see the badge change.
The Honesty Tradeoff
Here's what I learned building and rebuilding The Scout:
The old version (breakout patterns, Mars-Bot lineage) scored +19.7% on 16 months with 71.4% win rate. Amazing numbers. When extended to 6 years with honest sizing, it collapsed to -37%.
The new version (Hash Rate Momentum) has uglier headline numbers. +575% over 6 years. 55% WR. -46% drawdown. But it survives the regime tests that killed its predecessor.
I publish the ugly honest number. Not the pretty cherry-picked one.
The Bigger Point
The Scout's story is the Trading Bias series in action.
Overfitting — the old Scout was overfit to a short window. Real test on longer data exposed it.
Regime Bias — the new Scout is honest about WHICH regimes it wins in and which it loses. That's the right framing.
Benchmark Mismatch — The Scout's job is not to beat HODL in a rocket bull. It's to add uncorrelated returns across regimes. Different benchmark.
Cherry Picking — I could show only the winning walk-forward windows. I don't. Here's the losing ones too: 1 of 3 windows the bot underperformed HODL.
The Scout is the bot that keeps me honest. It reminds me that "beats HODL" is not always the goal. Sometimes the goal is "works when HODL doesn't."
-> Back to all bots -> The Watchdog: 184 Days of Doing Nothing — my main BTC bot -> Trading Biases series — the methodology behind these honest tests -> Our Best Bot Was a Lie — what the old Scout taught me
Disclaimer: Not financial advice. The Scout is in backtest status. No real capital is deployed. Past performance does not guarantee future results.
This bot in the post-mortem ledger: See /post-mortems →
Every retired strategy and failed walk-forward — documented publicly.
