By the Kunkafa team
AI Stock Prediction Accuracy Explained
A 70% next to a price move is a rate taken from finished forecasts, checked against what happened and corrected. What it promises, and what it never does.
You open a market in Kunkafa. Next to a 1.4% rise over the next four hours sits one number: 70%. Before you do anything with it, you want to know what it promises.
The finding. It is a rate. Among finished forecasts like this one, the price reached that level, that way, inside that period, about seven times in ten.
Where Kunkafa stands. Kunkafa counts its rates against real outcomes every morning, corrects them when they drift, and publishes the record.
What to do. Read the level, the direction and the duration with the number. Then check the record before you lean on it.
What the 70% is counting
A forecast in Kunkafa Predictions is four things at once. How far the price could move, as a percentage. Which way. How long it has. And Kunkafa’s confidence that it gets there.
The 70% belongs to that one line. Among finished forecasts that looked like this one, the price reached that level, in that direction, inside that duration, about 70% of the time.
Take away the level or the duration and the percentage loses its meaning. That is why every percentage on screen is labelled with what it measures.
| Change | Duration | Kunkafa’s confidence |
|---|---|---|
| +1.4% | next 4 hours | 70% |
| +2.8% | next 4 hours | 38% |
| -1.2% | next 4 hours | 64% |
| -3.1% | next 4 hours | 27% |
Four rows, four values, all true together. A small move is easier to reach than a large one. The up side and the down side of the same market are rarely mirror images. The side you asked about is shown first, the other below it, so neither gets read alone.
Reaching a level is not the same as being right
Kunkafa’s confidence concerns whether the price touched the level at any point inside the duration. Not where the price finished. The app keeps those two questions apart.
A move that touched +1.4% two hours in and then slid back has still reached the level. The record measures whether that level was reached within the duration.
A rate that is checked and corrected
The 70% is only worth reading because someone is counting. Kunkafa’s confidence values are compared with what actually happened and corrected when they drift. If forecasts that said 70% reached their level far less often than that, the number was wrong. It gets pulled back until it is not.
That check is built into how the experts are made. Seventy experts, not one, trained on about 10 billion data points of price history across stocks, indices, currencies, commodities and crypto, and tested on about 2 billion they had never seen. When they disagree, the probability drops, and you see that too.
Rain is the useful comparison. A 70% chance of rain says nothing provable about one day. It is a claim about every day that looked like tomorrow, tested by keeping score over many of them. Market probabilities are tested the same way. Understanding market uncertainty goes further into what such a claim can support.
The score is kept in public. Every morning Kunkafa counts how past forecasts ended, on live markets, and publishes the result at Kunkafa forecast performance results, with the raw figures at /api/stats.json.
When the app refuses to give you a number
Under 100 finished examples the record prints no rate at all, and says why: “Not enough past examples to estimate performance reliably.” A rate below 1% prints as “under 1%”, because a decimal would promise a precision the sample cannot support.
Those refusals are the part most worth trusting. Any tool can produce a percentage for every situation. A tool that sometimes declines is telling you where its evidence runs out.
What 70% does not promise
It does not promise the level is reached. About three times in ten, on a number like that, it is not. Those occasions are not malfunctions. They are the forecast working as described.
- It is not a return. What you risk, what it costs to be there, and when you leave are yours; the forecast says nothing about them.
- It is not a schedule. Reaching a level early, late, or after a move against you all count the same in the record. They feel nothing alike in practice.
- It is not the whole picture. A high value on a small move and a low value on a large one can be the same market saying one thing quietly. Read the rows together.
Most of the time the numbers will not be dramatic. Whatever is known is already in the price, so most forecasts sit near 50/50, and a 52% is shown as 52%. The few that lean are the ones worth stopping on. Why Kunkafa says that out loud is set out in Probability, Not Prophecy.
Every term above is defined in the Kunkafa market forecasting glossary, and the Kunkafa probability calibration and Brier score calculator lets you test a set of probabilities against outcomes yourself. Then read a live number in context in the Kunkafa market forecast app.
Keep reading
Twelve guides that go with this one, by the question you came with.