By the Kunkafa team
Danelfin vs Kunkafa: AI Scores and Price Probabilities
A stock score and a probability of reaching a price level answer different questions. Compare the event being measured before comparing the numbers.
You have a Danelfin score of 8 on a stock. You have a Kunkafa forecast on the same stock. One is a rank. The other is a probability. Comparing the two numbers directly is the first mistake.
The finding. No. Danelfin ranks stocks on beating the market over three months, or gives a range of returns. Kunkafa gives the probability of a stated price move within a stated period, in both directions.
Where Kunkafa stands. Kunkafa answers how far, which way, by when and how sure, then counts how past forecasts ended every morning.
What to do. Write down the event each number describes before you compare any accuracy claim.
We publish Kunkafa, so this is a product-owner comparison rather than an independent review. It is based on product descriptions checked on 7 September 2026, not a controlled test of either product's investment performance. No winner or return advantage is claimed.
What Kunkafa's price-level probability describes
Kunkafa is an AI financial markets app for stocks, indices, currencies, commodities and crypto. A forecast has four parts. How far, as a price level. 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. Both directions, always.
The homepage example reads like this. Could gold go up 1% within 12 days? Kunkafa’s confidence: 84%. Below it, the other side: down 0.6% in the same period, 70%. Either, both or neither could happen. The figures are illustrative.
Seventy experts sit behind each forecast, not one. When they disagree, the probability drops, and you see that too. A 52% is shown as 52%.
Every morning Kunkafa counts how past forecasts ended, on live markets, and publishes the record. It separates finishing in the forecast direction from reaching the level. How Kunkafa calculates market probabilities defines the terms. Kunkafa forecast performance results holds the daily count.
An upward level and a downward level can both be reached in the same period. So the two figures for Kunkafa’s confidence need not add to 100%. A level-touch probability is not a stock ranking or an expected return.
What Danelfin's AI Score describes
Danelfin rates stocks and ETFs from 1 to 10 on their probability of beating the market over the next three months.
The score is a rank. An 8 is not an 80% probability of reaching a price level. It says a stock sits high on the list relative to others. Danelfin's explanation of the AI Score sets out the definition.
Danelfin also offers AI price forecasts
The score is only part of Danelfin. Its AI Price Forecast documentation describes a central expected return with uncertainty bands for 1, 3, 6 and 12 months.
Its forecast-band explanation describes a band meant to cover the middle 68% of likely outcomes. A range of returns is a different object from the probability of touching a level on the way.
| Question | Kunkafa | Danelfin |
|---|---|---|
| What is described? | Kunkafa’s confidence in a market reaching a price level | AI Score: relative performance; AI Price Forecast: a range of returns |
| How is it shown? | A level, a direction, a duration and a probability, both ways | A 1–10 score, plus a separate central forecast and band |
| Which period? | The duration on the forecast, five minutes to seven years | Three months for the score; 1, 3, 6 and 12 months for the forecast |
| What to check? | The level-touch definition and the daily record | The benchmark for a score; duration and band coverage for a forecast |
Why the percentages are not interchangeable
Take an invented share priced at 100. Over the period it rises to 105, falls to 95 and finishes at 101.
It touched both an upward level and a downward level. Whether it beat the market depends on the benchmark’s return over the same period. Whether anyone made money depends on entry, exit, size and costs.
Those are three separate events. A forecast about touching 105 cannot be judged on the closing price alone. A rank against a benchmark cannot be judged by asking whether any rise occurred. And a 68% band describes coverage of a range, not a 68% chance of touching its upper edge.
How to evaluate alternatives fairly
- Compare the same assets and dates, using outputs recorded before the outcomes.
- Ask how unresolved outcomes and missing data are handled.
- Separate a provider’s own historical analysis from an independently reproduced result.
- Pick the output that answers your question, not the largest-looking number.
Need a rank of US stocks against the market? That is Danelfin’s job. Need to know how far a price could move, by when, how sure, and what the other side looks like? That is Kunkafa’s. What a Kunkafa confidence of 70% means goes deeper. Read a live forecast in the Kunkafa market forecast app.
Keep reading
Twelve guides that go with this one, by the question you came with.
