MaineblockA statistician's corner of the internet
Maineblock started years ago as a blockchain project, back when blockchain was the niche worth explaining. The technology changed; the mission didn't: take something technical, do the math honestly, and explain it in plain English.
Today Maineblock is a one-person shop run by a working statistician — the kind who has cleaned survey data at midnight, argued about weighting schemes, and written up results for people who needed the answer, not the jargon. Every tool on this site is built the way analyses get built in practice: state the assumptions, check the ones you can, flag the ones you can't, and translate the output into a sentence a human can act on.
No accounts, no tracking pixels, no “premium insights” upsell on page three. The calculators run entirely in your browser — your numbers never leave your device — and the guides say what the math actually supports, including when the answer is “you can't tell from this data.”
If you use statistics to make decisions with money, time, or reputation on the line, you're in the right place.
How every tool on this site works
Three steps, applied the same way whether the question is a clinical trial or a stock position. Each one is illustrated with the kind of output the tools actually produce.
State assumptions
Assumption check — live
- n·p = 177.4 ≥ 10 — normal approximation OK
- n·(1−p) = 1,596.6 ≥ 10
- Dropout is random — assumed, not verified
Computed live from your inputs — green, amber, or red — never a checkbox you tick yourself.
Flag limitations
Drawdown warning
Even at half-Kelly, losing streaks will cut deep — a 50% bankroll drawdown at some point is normal, not a sign the edge vanished.
Every tool states where its math stops working, right next to the result.
Interpret in plain English
Interpretation
You need about 1,774 per group to reliably detect a move from 10% to 13%. Allowing 10% dropout, enroll 1,972 per group.
Every result ships with a plain-English translation — no jargon without a decoder ring.
The Investor Stats Cheat Sheet
A free two-page PDF: expected value, Kelly sizing, Sharpe, sample-size math, and how to read a p-value without fooling yourself. Plus new tools as they ship.