Note
How to check whether someone actually predicted what they say they did
Four checks, in order of how often they catch something. Most disputes die at the second one.
Somebody says they called it. Here is how to find out, in the order that catches the most.
1. Find the original wording
Not the retelling. The post, the clip, the paragraph. Paraphrase is where most of the drift happens and it is usually innocent — people remember the gist and the gist is always sharper than what was said.
Watch for hedges that vanish in the retelling. “I think it can be done in thirty years, provided there is an exponential increase in tonnage” becomes “he predicted thirty years.” The condition was doing real work and it is gone.
2. Establish when it was published
This is where most claims actually die, and it is the check people skip.
Clip channels repackage old footage under new titles constantly. While building this archive I found a video titled “Michael Saylor — I’ve COMPLETELY Changed My 2026 Prediction” whose narrator states, in the clip, a target “by the end of 2025”. The upload was recent. The prediction was not.
Use the platform’s own publish date, not the upload date of whatever version you found. If the two disagree, you are looking at a repost.
3. Read what was actually claimed
Three things quietly change the meaning:
By, or on. “Hits $200k by December” is satisfied by a single tick. “Is above $200k in December” needs the price there on the day. People switch between these depending on what happened.
The conditional. “If X, then possibly Y” is not a prediction of Y. It is a prediction about the relationship, and it is only wrong if X happened and Y did not.
The unstated source. “We’ll triple our compute next year” sounds checkable and is not, because nobody publishes the figure. A claim with no possible referee cannot be won or lost, only argued about.
4. Check whether the standard moved
The last one, and the hardest, because it needs something written down at the time. Was there a stated threshold then, and is the threshold being used now the same one?
Usually there is nothing to compare against — the test only ever existed in someone’s head, which is what makes this failure so easy and so common. That absence is the specific gap this registry was built to close: the test is written and hashed alongside the quote, before the outcome, so the comparison is possible at all.
If you have a link and want the first three done for you, paste it in — the source is fetched, the wording is checked against it, and the publish date comes from the platform rather than from whoever posted it.
Every figure above is a count from the record on August 29, 2026, and the archive is exportable if you want to check the arithmetic yourself.
Also here: People who write predictions get scored. People who say them don't. · What makes a prediction testable · The goalposts problem · What a 96% accuracy score actually measures · How a track record lies without a single false statement · Sometimes the honest verdict is that nobody can tell · Everyone predicts in January and finds out in December