Out of SampleA weekly newsletter · Free · Every Sunday
One claim about markets.
Tested. Code included.
The finance internet runs on confident claims that nobody checks. I check them against real data, show my method, and publish the result. Even when the answer is boring. Even when it makes me look wrong.
First issue lands Sunday, October 4. No picks, no alerts, no spam. Unsubscribe in one click.
What the first two issues found
What $1 became holding stocks, versus selling whenever they looked expensive. 1881 to 2023.
Issue 01How long the "stocks are expensive" signal kept you in cash starting April 1995. You'd have missed the entire late 90s run.
Issue 01The famous VIX buy signal cites 207 data points. Only 9 are independent. Its 97% hit rate falls below doing nothing.
Issue 02How every issue works
Same five sections. Every week.
- The claimSomething people actually say, quoted from a real source.
- How I tested itThe data, the dates, the exact rules. No hindsight allowed.
- What the data saysThe table. No spin.
- What it meansMy read on it, in plain English.
- What would change my mindThe result that would prove me wrong, stated up front.
The scoreboard
Claims tested so far
| No. | Date | Claim | Verdict |
|---|---|---|---|
| 01 | Oct 4, 2026 | Sell stocks when they look expensive $1 became $327, not $1,265 |
Lost money in every threshold tested |
| 02 | Oct 11, 2026 | A VIX above 40 is a buy signal that's right 9 out of 10 times 207 data points are really 9 |
True, and nearly meaningless |
The rules
What I'll never do
I test claims and methods. I don't make calls on stocks, and I won't review your portfolio.
Sunday morning, every week, no matter what the market did. Nothing in between.
Every issue says how it could be wrong. Corrections get their own issue with my name on it.
Worth your Sunday morning?
One tested claim a week. Free.