Arena Blog

Bitcoin has no native yield. Every advertised BTC yield is a sold option, a loan, or a token emission. And with the income ETFs, the window decides whether the answer flatters or destroys.

Daily backtests compute the signal from the close and fill at that same close — a price you can no longer trade. We recomputed 7,000+ trades twice (close fill vs. next open) to size the distortion. Result: smaller than the myth.

A backtest claiming a 100% win rate and 34 million percent returns is circulating. We're rebuilding it — and setting the rules before we compute.

The copper/gold ratio is sold as a macro and Bitcoin signal. What it actually measures, where it fails, and why a viral "Bitcoin cycle" indicator built on it has look-ahead bias.

ORB is sold as the holy grail of intraday trading. What the strategy really is, why the viral result misleads, and what makes or breaks an honest ORB backtest.

Our backtest engine used to compute returns before costs. We built a net-of-cost layer that re-prices every fill at a realistic fee — and it flipped our second-most-popular strategy negative after fees. Here is the math, the numbers, and why "beats buy-and-hold" is half a truth until you pay the fees.

Asking which company is well-positioned in the AI drug-discovery race is really three questions at once. We sort the field soberly into shovels, bets, and giants — and show which layer earns regardless of any single trial outcome.

Our Bollinger Squeeze strategy lost to Buy & Hold by double digits on crypto. Before deleting it, we rebuilt it to the modern standard (TTM Squeeze). The rebuild turned positive — and we retired it anyway. Why "it works now" isn't enough.

Auditing 40,000 backtests, our Fibonacci strategy came up short. Before cutting it we asked: is it our config? A better exit made it 20 points better — and we retired it anyway. Why "beats Buy & Hold" isn't the bar.

We audited 40,000 of our own backtests — and benched several of our own strategies in the process. Here is what we killed, what we reactivated, and the uncomfortable core lesson: whether a strategy is "good" depends less on the strategy than on the asset class it runs in.

Five weeks ago it was 10,000 backtests; today it's over 40,000. With 4x the data, the "beats Buy & Hold" rate falls from 64% to 52%, four in five runs land below 50% win rate, and one memecoin still holds its 197,923% record. What selection bias does to sample size — on real numbers.

"Invested in AI healthcare" can mean two almost opposite things. Two real companies show it: an annuity machine with ~86% recurring revenue — and an option with not a single approved drug. Same label, opposite risk DNA.