Every few years a chart goes viral claiming the McRib predicts recessions, Taylor Swift album drops move the Nasdaq, or the market sells off during Mercury retrograde. It is great content. It is terrible investing.
We know, because we built those indicators. On purpose. Then we did the thing nobody who tweets the charts ever does: we measured them against the actual S&P 500, out of sample, and wrote down the number.
Here is the punchline up front. The fun indicators carry almost zero predictive signal. The boring market-plumbing ones carry the edge. And we will tell you exactly which is which, because that honesty is the entire point of this site.
Why weird indicators are so seductive
Weird indicators feel like a secret. If the McRib's return window "calls" a market top, you have found an edge the suits missed — a cheat code hiding in plain sight at the drive-thru. It is a better story than "high-yield credit spreads widened," and stories spread.
They are also easy to cherry-pick. Run any quirky time series against the market over a few years and a handful will line up by pure chance. Screenshot the winners, ignore the losers, and you have a viral thread. The losers never get posted.
So we ran the experiment properly instead.
What we tested
We track 658 indicators on StandardPoorly. For every one, we compute its correlation against forward S&P 500 (SPY) returns across six horizons — 1 day, 5 days, 21 days, 63 days, 126 days, and 252 days — and validate out of sample so we are scoring the signal on data it was not fit to.
We sorted the indicators into two honest buckets:
- Novelty ("poorly") categories — the deliberately absurd stuff: chaos (McRib, lunar phases, celebrity events), weather, lifestyle search trends, sports, culture, crypto vibes, political-trade noise. 256 indicators.
- Plumbing ("standard") categories — the unglamorous machinery of markets: macro, credit, currencies and rates, market internals, sentiment, energy activity, travel and leisure. 155 indicators.
Then we compared the average absolute correlation of each bucket. No spin, just the arithmetic.
The honest results
| Bucket | Example indicators | Indicators | Avg \ | correlation\ | |
|---|---|---|---|---|---|
| Novelty (the jokes) | McRib, Taylor Swift, Mercury retrograde, Full Moon | 256 | 0.021 | ||
| Plumbing (the edge) | credit spreads, CPI, rates, market internals | 155 | 0.150 |
The plumbing carries roughly seven times the correlation of the jokes — and that is the conservative read, because the novelty average is propped up by a few flukes. The median novelty indicator has an absolute correlation of 0.009. That is not a small edge. That is a coin flip wearing a costume.
Now the specifics, because naming names is the fun part. Here are real indicators from our database with their real, measured correlations to SPY:
| Joke indicator | Category | Correlation vs SPY |
|---|---|---|
| McRib available (national return window) | chaos | 0.071 |
| Days since most recent McRib return | chaos | 0.035 |
| McRib comeback day (first day of return) | chaos | 0.015 |
| Taylor Swift album release day | chaos | 0.024 |
| Days since most recent Taylor Swift album | chaos | -0.069 |
| Mercury retrograde period | chaos | -0.029 |
| Full Moon | chaos | -0.003 |
| Groundhog Day (Feb 2) | chaos | 0.002 |
| Super Bowl Sunday | chaos | 0.000 |
| Search: "Mercury retrograde" | lifestyle_search | 0.000 |
A correlation of 0.07 is the McRib's high score, and it is still statistically a rounding error. Super Bowl Sunday and the "Mercury retrograde" search trend land at a clean 0.000. The famous Super Bowl Indicator, measured honestly, predicts exactly nothing.
For contrast, here is the boring stuff that actually moves with the market over a 252-day horizon:
| Plumbing indicator | Category | Correlation vs SPY |
|---|---|---|
| High-Yield Credit Spread (BAML OAS) | macro | 0.723 |
| Investment-Grade Corporate OAS | currencies_rates | 0.696 |
| Case-Shiller Home Price Index | macro | -0.525 |
| CPI — All Items | macro | -0.485 |
| HYG/LQD ratio (junk vs investment grade) | market_internals | -0.459 |
| Revolving Consumer Credit | credit | -0.423 |
Credit spreads correlate with forward equity returns at 0.72. The McRib correlates at 0.07. That is the whole article in two numbers.
Why smart people still fall for the McRib
This is not a story about gullible people. It is a story about statistics doing exactly what statistics do.
Multiple testing. Test 256 silly indicators against the market and a few will look impressive by chance alone. Our flashiest novelty result — "days since last Liam Neeson movie" at -0.36 — is a textbook example. It is not a market signal; it is what randomness looks like when you run hundreds of trials and screenshot the winner. Run enough coin flips and one streak looks like a prophecy.
Data mining without a holdout. If you fit an indicator and grade it on the same data, of course it fits. The discipline is testing it on data it has never seen. Most viral indicators have never met a holdout set.
Survivorship in your feed. You see the McRib chart that "worked." You never see the thousands of fast-food, sports, and celebrity series that went nowhere. Your timeline is a highlight reel of flukes, and a highlight reel feels like a pattern.
Narrative gravity. "The consumer is splurging on novelty sandwiches, a sign of confidence" sounds smart. It is a story wrapped around noise. The plausible mechanism is what makes the spurious correlation dangerous — it gives you a reason to believe the coin flip.
What actually has edge
The signal lives in the plumbing. Credit markets price risk before equities do — when high-yield and investment-grade spreads widen, stress is usually already in the system. Market internals like the HYG/LQD ratio (junk bonds versus investment-grade) show whether a rally has participation or is running on fumes. Macro series — CPI, home prices, consumer credit — set the regime that equities trade inside of.
None of it will go viral. There is no drive-thru tie-in. But it is where the measurable relationships actually are, and it is what we surface first.
The StandardPoorly deal
We keep the joke indicators because markets are absurd and pretending otherwise is its own kind of dishonesty. The McRib indicator is real, it is live, and you can watch it do nothing. That is the bit.
But every indicator on the site — silly or serious — gets the same honest correlation score, the same out-of-sample test, the same horizon breakdown. We will never dress up a 0.07 as an edge. The fun ones are brand and color; the boring ones are the signal; and we label which is which, every time.
Browse all 658 — McRib included — at /indicators. When you want the ones that actually move with the market, the live reads are at /signals.
Order the McRib because it is delicious. Don't order it because of your portfolio.
Not investment advice.