Options flow analysis has more believers than evidence. That said, the evidence that exists generally supports the core thesis - informed options activity precedes underlying moves. The interesting question is which parts of the thesis survive scrutiny.

What the research consistently shows

Stocks with abnormally high options volume tend to outperform over the following 5-20 trading days.
The effect is stronger in small and mid-cap names than in large caps.
OTM activity carries more signal than ATM, controlling for everything else.
Concentrated single-strike activity outperforms broadly distributed volume.

What is more controversial

Whether the effect persists net of transaction costs for retail-sized positions.
Whether sweep vs block distinction adds incremental signal or just correlates with volume.
How quickly the signal decays - most studies find 5-day signal stronger than 20-day.

Where naive backtests fail

Most retail-published 'UOA backtests' suffer from look-ahead bias (filtering on data not available at the print), survivorship bias (only including stocks that still exist), and overfit thresholds. The cleanest studies that survive are point-in-time, include delisted names, and avoid parameter optimization.

The realistic edge

When you filter for Vol/OI above 5, total premium above $250k, sweep structure, and a 5-10 day holding window, the historical average excess return is modest - somewhere in the 0.5-2% range above market, with high variance. That is not a casino - it is a positive-expectancy signal worth combining with your own analysis.

Treat unusual options activity as a screen, not a strategy. It tells you where to look. What you do with what you find is still your job.