Guide

What the confidence levels on a CS2 prediction mean

Why a CS2 prediction carries a confidence level as well as a probability, where the boundaries between levels sit, and what each level has actually delivered.

A confidence level is a band the probability falls into

Every prediction has a win probability. The confidence level is the band that probability lands in — a label over a range, not a second opinion arrived at separately. It can never disagree with the number, because it is derived from it.

Why the levels exist at all

A page of raw percentages is hard to scan and easy to over-read. Sixty-two and sixty-four are not meaningfully different picks, and shown side by side as numbers they invite a comparison that the underlying estimate cannot support. A level says "these two are the same kind of pick", which is the truthful reading.

The levels are a setting, not a constant

How many levels there are, and where one ends and the next begins, is configured rather than hard-coded, and the scale has already changed as the record grew. That matters for reading the archive: when the scale changes, every past pick is placed on the current scale again by its win probability. The per-level breakdown on the accuracy page therefore always describes the scale you see on the picks, and the overall accuracy does not move.

A higher level is not a stronger claim about one match

It is a claim about a larger share of similar matches. The top level is expected to be wrong sometimes, and when it is, that is a normal outcome rather than a failure of the level. How the record is measured explains what "expected" means here.

The lowest level is a real answer

A pick at the lowest level is saying the match is close to even. That is worth publishing: knowing a match is a coin flip is more useful than a confident-sounding sentence about it, and pretending otherwise is how a record gets padded with claims nobody checked.

Levels are shown as shape and text, never colour alone

The level appears as a rating and its name together. Colour carries no meaning on its own anywhere on the site, so the level is readable without seeing colour at all.

The record is broken down by level

The interesting question about the levels is whether the higher ones actually land more often. The accuracy page answers it directly, per level, from the same reviews that produce the overall number — so a level that is not earning its label has nowhere to hide.

What a level is not

It is not a stake size, a recommendation, or a rating of the teams. It describes how far from even one estimate sits, and nothing else. What a prediction is and is not covers the distinction in full.

Questions

How many confidence levels are there?: Three. The set and its thresholds are configurable, so this number is read from the scale in force — the same one the stars on every pick are drawn against — and the figure beside each level on the accuracy page is always current.
Why is the exact percentage not shown everywhere?: Because on a list of matches a two-point difference reads as a distinction that the estimate cannot support. The full probability is on the prediction itself.
Do the levels change what the prediction says?: No. The level is derived from the probability. If the thresholds move, the label can change while the underlying call does not.
What does it mean if a top-level pick loses?: That the expected minority of them happened. One result at any level says nothing; the per-level record over many matches is the thing to read.
Are low-confidence picks published?: Yes, and they count towards the record like every other. Publishing only the comfortable ones would make the accuracy figure meaningless.
Does a level tell me how much to risk?: No. Nothing on this site is staking advice — see what a prediction is and is not.

Back