A serious newspaper about an unserious idea.

For the last fifteen years, two genres of financial writing have dominated the internet. One is rigorous, expensive, and read by hedge funds. The other is free, viral, and almost entirely wrong. Standard Poorly is an attempt at a third thing: rigorous, free, and openly admitting it is at most 30% useful.

Every indicator we publish is real. The data is sourced from public APIs and scrapes. The p-values are not invented. The walk-forward validation, when we run it, is honest. We just happen to also publish the multiple-testing burden, the OOS misses, and the indicators with great backstories and terrible statistics — because those are the most informative thing we know how to print.

With the McRib. Specifically: the bitcoiner folklore that every time the McRib comes back, Bitcoin has a substantial near-term run. Half-meme, half-folk-finance, and exactly the kind of claim that sounds too dumb to be real and survives long enough that someone should probably check. We went looking for the data. The pattern showed up often enough across the limited number of McRib reappearances to be worth reporting — with the usual multiple-testing caveat that any pattern fit to a handful of events is one coincidence away from being noise.

A colleague then surfaced a second folklore variant of the same sandwich: was the McRib's release schedule itself timed to dips in the spot price of pork? McDonald's drops it when bellies are cheap, pulls it when they aren't. That one's on the backlog. Two folklore correlations, one sandwich, both worth running through the engine.

Then we did it again with weather — the question of whether the predominant weather pattern across the population-weighted U.S. is indicative of where the market goes on a given day. Two competing theories made it interesting: maybe good weather thins out trading (everyone's outside) and pushes volatility up, or maybe good weather lifts mood, lifts risk appetite, and pushes equities up. Both stories are plausible; both can't be true at once. The data is the only way to tell which folklore is doing the work.

Then it was podcast hosts. Specifically the All-In crew — Chamath, Sacks, Friedberg, Calacanis — and the question of whether their on-air conviction calls, tracked at the day-of-broadcast and held in a transparent rules-based way, beat SPY. (Several do. Several don't. The methodology matters more than the ticker.) That work grew into Pundit Poorly / Bestie Buys and a full backtest engine with five published methodology presets.

The pattern repeated enough times — a goofy cultural artifact turning out to contain a real, measurable, statistically modest signal — that it stopped feeling like a series of one-offs and started feeling like a publication. The project you're reading is that publication.

To make statistical literacy entertaining enough that a non-quant reader leaves the site with a more accurate understanding of what “this signal predicts the market 72% of the time” actually implies — which is, almost always, “less than you'd think.”

The product is a public good with a power-user tier. The free Almanac is the publication. The paid Analyst tier adds custom horizons, the data chat, BYOB, and alerts — for people who want to do their own research. The paid Trader tier adds the API, MCP server, real-time quotes, and compound alert rules — for people who want to automate the workflow. The Press tier is the same product, free, for journalists who agree not to launder our indicators into clickbait without the multiple-testing caveat.

We publish anonymously, as an editorial collective. The reason is boring and institutional: the people who run this also have day jobs in adjacent industries (finance, data, consulting), and a public byline would create conflicts on stuff that has nothing to do with Standard Poorly. The site itself takes no positions, accepts no advertising, and earns no commission on anything we describe. The ledger is the byline.

For corrections, press inquiries, or to tip us on a signal you think we should look at — including the next McRib — the addresses are below. Real humans read them, usually within a day. Press passes get same-day turnaround if you're on deadline.