Skip to main content
← Back to blog

We tested 25 strategies across 9 markets: 0% beat buy & hold in bull markets

📅 2026-08-16
✍️ Strategy Arena
buy and hold backtest overfitting world arena negative results gpu

The test nobody publishes

Everyone publishes their winning strategies. Nobody publishes the full table. Here is ours, untouched.

We took 25 strategies — momentum, grid, turtle, DCA, trend-followers, mean-reversion, GPU-evolved genetic variants — and ran them identically across 9 markets for one year, transaction fees modeled: Bitcoin, gold, silver, Nasdaq 100, S&P 500, oil, DAX, CAC 40 and EUR/USD.

The question was simple: how many beat simply buying and doing nothing?

The table

Market Buy & hold (1 yr) Average strategy % beating B&H
Silver (SLV) +85.5% +9.3% 0%
Oil (USO) +57.1% +7.3% 0%
Nasdaq 100 (QQQ) +35.9% +2.8% 0%
S&P 500 (SPY) +23.5% +2.4% 0%
Gold (GLD) +22.1% +4.0% 0%
CAC 40 +8.9% −0.3% 0%
DAX +5.5% −1.8% 0%
EUR/USD +0.5% −2.0% 12%
Bitcoin −37.2% −10.2% 96%

On seven of nine markets — every one that went up — the score is zero. Not "few". Zero out of twenty-five.

The Bitcoin trap

The one line that seems to save the strategies is a statistical trap. 96% of our strategies "beat" buy & hold on Bitcoin… because buy & hold lost 37% over the period. Beating a collapsing market means losing less than it does: the average strategy still lost 10.2%. Anyone reading "96% success" without checking the denominator is buying an illusion.

That is precisely the mechanism behind most sold backtests: pick the period, the market and the benchmark that flatter the number.

What this means (and what it doesn't)

What it means: in a rising market, activity costs money. Every premature exit, every fee, every missed signal is paid against an opponent that does nothing. If your backtest beats buy & hold on a bull asset, the first hypothesis to test is not "I found an edge" but "where does my protocol leak?"

What it doesn't mean: that every strategy is useless. Some of our defensive strategies lose far less than the market in downturns — that is a real result, but it's another article, and it sells worse than "+170% annualized".

How to put your own strategy on trial

This is exactly what we built two tools for:

  • The free Windows Lab: your GPU searches hundreds of variants locally, a sealed holdout stays invisible during the search, and the verdict lands on your machine. Educational simulation, no profit promises.
  • The free Quick Check: paste a Pine script, compatibility is verified without an account. Then, if you want an independent server-bound execution with a sealed holdout, the audit exists — and an unfavorable verdict is still a delivered audit. Nobody can buy a better verdict.

All data in this article comes from the World Arena snapshot of June 15, 2026 (25 strategies, 1 year, fees modeled). The full matrix is at the World Arena page.

⚠️ Disclaimer — This article is for informational and educational purposes only. It does not constitute investment advice or a buy/sell recommendation. Past performance does not guarantee future results. Strategy Arena is an educational simulator with virtual capital. Always do your own research before making investment decisions.

Enjoyed this article? Share it

𝕏 Share on X ✈️ Telegram