Absolute vs relative pool ratings

What a relative rating is

A relative rating — Elo, Glicko, Fargo Rate and the rest of the family — has no notion of an absolute standard. It only knows results: you beat them, they beat someone else. From a large enough web of those results, the system finds the set of numbers that best explains the outcomes. Nobody ever measures a player directly; every rating is a position in a network of other ratings.

That is an elegant design and it works remarkably well when the network is dense. It also carries two consequences that no amount of clever maths removes.

Consequence one: it needs a lot of games

A result is one bit of evidence, and one bit is very little. So the system needs volume before it can separate you from luck — which is why FargoRate uses a robustness measure and treats 200 games as the minimum for calling a rating established, blending a starter rating into the number until you get there. Until you have paid that price in games, your rating is partly a guess about you.

Consequence two: it is anchored to your neighbours

Because every rating is defined against other ratings, a group of players only lines up correctly with the rest of the world if enough games connect them to it. Where that connection is thin — an isolated region, a new league, a scene whose players rarely travel to open events — the group can settle at a level that does not match the same numbers elsewhere. FargoRate describes precisely this case, two nearly isolated groups with one rated too high relative to the other, as a vexing problem, and notes that games against established opponents are worth more for exactly this reason.

The practical version for a player: if your city is stacked with strong players, or barely connected to the wider rated population, your number is partly a statement about your surroundings. Two players of the same standard in different scenes need not read the same, and neither of them can do anything about it except play more games against outsiders.

What an absolute rating is

An absolute rating measures performance against a fixed standard rather than against people. Golf handicaps work this way against par. Athletics works this way against the clock. A stopwatch does not care who else is in the race, and 10.4 seconds in Manila is 10.4 seconds in Manchester.

Pool has not traditionally had one, because pool lacks an obvious clock. Runout Rank supplies the equivalent: a set of defined table layouts and one question — can you run this out? Ten tables at a level, one attempt each, no retries and no skips. The number that comes out is computed entirely from your own results against the layouts.

So there is no opponent pool to be strong or weak, nothing to drift against, and no minimum number of games before the measurement is valid. You have your rating at the end of the first session, and it means the same thing anywhere.

How a fixed standard avoids becoming a memory test

The obvious objection: a fixed set of layouts stops measuring skill as soon as you have played it a few times, because you are then recalling solutions rather than finding them.

Runout Rank avoids that by fixing the difficulty rather than the tables. A level is a set of published constants — object ball count, ball in hand or not, minimum spacing, blockers — and layouts are generated fresh inside those rules every time. You never see the same table twice, and every table asks the same question. Ten of them in a row average out what luck remains.

What an absolute rating cannot do

It is not a handicap system, and it should not be used as one. A relative rating exists to predict a match between two specific people, and it is far better at that than any absolute measure — because match outcomes are what it is built from.

An absolute rating also has its own variable to keep honest: the equipment. Pocket cut, table size and cloth speed all change how hard a run-out is, so a rating taken on a nine-foot table with tight pockets is a different measurement from one taken on a bar box. Fix your conditions, take the test on the table you compete on, and compare your own numbers over time.

Which one do you want?

They answer different questions, and a serious player can reasonably carry both.

Runout Rank is independent and is not affiliated with, endorsed by or connected to FargoRate. Everything said here about Fargo Rate is drawn from FargoRate’s own published material and is described as fairly as we know how; it is a good system, and this page is about where its design does and does not fit a particular kind of player.