By the Kunkafa team
Can AI Predict the Stock Market?
You want a yes or a no. The honest answer is a probability with a record beside it, and Kunkafa publishes that record every morning.
You want a yes or a no. Can AI predict the stock market, or can it not? Neither answer is honest. The honest answer is a probability, with a record you can check.
The finding. No tool sees the future. A forecast names a level, a direction, a period and a probability, and then gets counted. That count is the only answer worth having.
Where Kunkafa stands. Kunkafa puts a probability on a stated move, shows both directions, and publishes every morning how past forecasts ended.
What to do. Ask any tool for its record. If there is none, treat the number as marketing.
Why “can AI predict the stock market” has no yes or no
A screenshot of a 94% win rate is greed talking. It sells certainty, the one thing markets never hand out.
“Nobody can forecast anything” is fear talking. It costs just as much, because it leaves you no way to compare one path against another.
Neither belief survives contact with a record. So the useful question is not whether a machine sees the future. It is whether a number printed today can be checked against what happens next, by you.
“Will the market go up?” cannot be graded. It never said how far, by when, or how sure. Rewrite it: of forecasts like this one, how often did the price reach that level, that way, inside that period? Now it is a counting problem.
The trap: accuracy with no stated question
A 90% accuracy claim is a warning sign. Accuracy with no stated question is a free parameter, and there are four easy ways to make a big one.
- Narrow the filter until a handful of examples remain. Any rate becomes reachable.
- Count only the forecasts that closed well, and leave the open ones out.
- Pick the window that flatters.
- Grade an easy question, such as “did the price move at all?”, and report the rate.
None of that requires dishonesty. It is what happens when a number is chosen for a landing page instead of produced by a rule.
What Kunkafa shows instead
Kunkafa is an AI financial markets app for stocks, indices, currencies, commodities and crypto. A forecast has four parts: how far, as a percentage; which way; how long it has; and Kunkafa’s confidence that it gets there.
The side you asked about is shown first. The other side sits below it. The homepage example: could gold go up 1% within 12 days? Kunkafa’s confidence: 84%. Below it, down 0.6% in the same period, 70%. Either, both or neither could happen. The figures are illustrative.
Seventy experts, not one. When they disagree, the probability drops, and you see that too. They were trained on about 10 billion data points of price history and tested on about 2 billion they had never seen.
A 52% is shown as 52%. Markets are efficient. Whatever is known is already in the price, so most forecasts sit near 50/50. That is the finding, not a hedge. A tool with a strong lean on every market at every duration is reporting noise in a firm voice. Why exact price forecasting is impossible goes further.
The record, published every morning
Every morning Kunkafa counts how past forecasts ended, on live markets. The record is public at Kunkafa forecast performance results, with the raw figures at /api/stats.json.
It shows three counts. Forecasts evaluated. How many were confident enough for the published filter. How many reached the level. You can read them over the last day, 7 days, 30 days or all time.
Two rules do most of the work. Under 100 finished examples, no rate is quoted at all; the panel says the sample is too small. A rate below 1% prints as “under 1%”. A forecast still open is counted as still running, not parked until it looks better.
A worked example, with invented round numbers. In a month, 1,000 forecasts are evaluated. 200 are confident enough for the filter. 116 finish in the forecast direction, which is 58%. 80 reach the level, which is 40%. The 58% belongs to the 200, not the 1,000. Read the “confident enough” column first. How to read Kunkafa forecast results explains each column.
You can also grade Kunkafa yourself. Open the app, pick a market, and move the confidence filter: the counts move with it. A saved outlook is a frozen copy of a forecast as it stood when you saved it. Keep a few, wait, and compare them with what the price did. The counting rules are written out in how Kunkafa calculates market probabilities.
What it still cannot do
- It cannot say what happens next. A 70% leaves a 30%, and that side arrives often.
- It describes finished forecasts, not the next one. Your market today is one draw from that population.
- It cannot rescue a market it has too little history on. Where the sample is thin, it says so.
So, can AI predict the stock market? It can put a number on a stated move, publish the rule, and let you count. Everything after that is your judgement. See a live forecast and its record in the Kunkafa market forecast app.
Keep reading
Twelve guides that go with this one, by the question you came with.
