Incomplete statistics are a disease.
And a nice parallel showed itself today between bitcoin and the world cup.
Where in the recent world cups match Yesterday, when Belgium equalized 1-1, the commentator went:
"Belgium hasn't scored twice against spain in 46 years, (so a win for them is unlikely - he didn't say it but thag's where his intentions went)"
As an argument for Belgium to win 2-1 or better is unlikely. "They'll have to create some real magic."
All power to the commentator btw, it's a great catch phrase, it creates engagement and keeps the viewer attached to the screen: "what if"... "if belgium wins that'd be legendary"...
Exactly his job.
But it is indeed primarily aimed towards creating engagement, not really making a proper call one can place a bet on.
"Not done in 46 years" sounds almost like a certainty Spain wouldn't lose. "All in on Spain".
But going deeper reveals all, how many matches did they play last 46 years?
Only 13 since. And since football's (soccer) dynamics change drastically after 1-1 (a match going 1-1 -> 2-1 is drastically different than a match going 2-0 -> 2-1. So conditional statistics apply and are important: "how many times did Spain-Belgium go 1-1 first? Then 2-1 (or better)?"
Only once in those 46 years.
So the sample size reduces to just one.
And just because something happened once before, implies almost no edge or significance it will happen again.
World cup is a fantastic example.
And the best part? Spain won, so a large amount of confirmation bias is activated to that commentator and he now seems like a genius with that comment, he thinks that and the gambling continues with his next dubious statement.
The link to bitcoin?
You tell me. But how many have been saying $btc bottomed in June/early July, made a rally in July and therefore the bottom "will" be in Q4 since this has happened "the past 10+ years"? Where in reality, it only happened once/twice before (bitcoin bottoms macro about once every 4 years).
But over the entire history this sequence (bottom June, rally July, bottom Q4) only happened twice and is therefore an incomplete statistic.
In 2014, bitcoin made a low in April (not June), then in October (Q4) 2014 but only actually bottomed in Q1 2025.
Anyone remember the "new low in April" callers in April 2026?
Yeah, they're nice and quiet about it now too.
So truly they were only "right" 1 out of 2 times so far.
Why is this so inconsistent? Because sample size is limited, and the user is using limited statistics.
What is often pushed as absolutes "hah, I told you", is merely a tradeable or predictive statistic.
And the core of the issue is indeed an incomplete statistic and sample size.
It's indeed great for engagement "hah, I told you, follow me"
But it's not tradeable. We need at least 4/5 data points at the bare minimum, not 1 or 2.
Eliminating incomplete statistics and only incorporating thorough statistics, are the easiest way to skyrocket your directional success.
Case closed.