Elo vs Ree

If you have used Rankade, you will have seen "ree" scores on your group's rankings. Ree is Rankade's own rating system, and it is a reasonable thing to be curious about — particularly if you have only ever met Elo before.

This page explains what each system is and how they differ in kind. It is not a product comparison.

What Elo is

Elo is an open, published formula. Two ratings go in; the gap between them produces an expected score; the result moves both ratings by the difference between what was expected and what happened. The K-factor decides how far.

Because it is published, every property of it is inspectable. You can calculate any rating change yourself with a pocket calculator, confirm that points gained by one player equal points lost by the other, and satisfy yourself that nothing else is going on. Sixty years of chess, sports and esports use have given it a very well-understood set of strengths and weaknesses — including the ones that count against it, like having no notion of how confident it is in a number. The full explanation is here.

What ree is

Ree is a proprietary algorithm belonging to Rankade. The formula is not published, so what can be said about it honestly is limited to what Rankade documents.

From that, the shape is recognisable: a player or a "faction" — Rankade's word for one side of a match, whether that is a single player or a team — gains score for a good result, and gains more when the opposition was stronger. So the central intuition is the same one Elo has. Rankade presents ree as tuned to work across a very wide range of activities under a single system, from board games to sports to video games, and to handle team and multiplayer results natively.

What cannot be said is how it does any of that, because that is the part which is not public. That is not a criticism — plenty of good rating systems are proprietary, and a closed formula is harder to game deliberately. It is simply a difference in what you can verify for yourself.

The difference that actually matters

For most casual groups the practical difference in the rankings is smaller than it sounds. Both systems reward beating stronger opponents more than weaker ones, and both converge on a sensible order after enough matches. If all you want is a fair leaderboard, either will give you one.

The difference shows up when you want to understand a number rather than accept it. If someone in your group asks "why did I only gain 4 points for that?", an open formula has an answer you can work through together. A closed one has an answer you have to take on trust.

That is the reason Custom Elo Games is built on Elo: the groups using it are small enough that arguments about the leaderboard are part of the fun, and those arguments are better when the maths is checkable. You can check any single head-to-head result on the Elo calculator, which shows the expected score it worked from as well as the points exchanged.

If you are weighing up the two platforms rather than the two rating systems, that is a different question and we have written about it separately — see Custom Elo Games and Rankade compared.