ELO vs rank?

[celtic_neonom_]

I am confused between ELO and rank. They seem to behave very differently. Players with similar ELO have very different rank. Can someone please explain these rating systeems and what is the better method?

[pecelot]

http://www.agecommunity.com/faq_mp.aspx#ratings

http://aoe3.jpcommunity.com/rating2/ — at the bottom there's the Elo calcuation method.

They usually differ that much because PR decays over time due to inactivity of players, whereas Elo ranking does not.

[macacoalbino]

Also there's the thing that pr counts team games and 1v1 in one rank and ELO separates them

[roiarthurbis]

ELO is really accurate way of gauging a players skill as it's harder to obtain, while PR is just as accurate but can be inflated very easily. Therefore being less accurate in some cases.

[deleted_user0]

Pr is really not accurate because you lose it when inactive and gain ot alot quicker when just active, be it a new acc or one that became active again after decaying to the minimum.

Also macacoalbino made a good point about separating 1v1 and team, especially since its easier to gain pr through team, as there are more teamgames.

[roiarthurbis]

Pr is really not accurate because you lose it when inactive and gain ot alot quicker when just active, be it a new acc or one that became active again after decaying to the minimum.

Also macacoalbino made a good point about separating 1v1 and team, especially since its easier to gain pr through team, as there are more teamgames.

PR is very accurate for players who are very active.

[deleted_user0]

Not really either, since the pr you gain or lose depends on the pr of the people you play. You cant just take it as an isolated incident. Your words are quite contradictionary

[Gendarme]

depends on the pr of the people you play.

Not to mention the difference in homecity levels and the fact that some civs gain more each win than others (not entirely sure about this one).

[ssaraf]

since the pr you gain or lose depends on the pr of the people you play.

It is supposed to be that way. Else all high prs will bash low prs all day and gain loads of pr.

[Gendarme]

The point (that I think Umeu tried to make) is that skilled players have low PR as well, due to decay. If you are, let's say, a lieutenant colonel (PR 33), you can easily lose against Umeu (PR 25), which will cause you to lose a lot of points, even though you really shouldn't. Even if you might not yourself ever play against someone with a decayed PR, you will play against someone who has and thus has an inaccurate rating, and this will cause your rating to be inaccurate after the game as well.

Of course, the magnitude of the inaccuracy in the system depends on the activity of the decayed accounts, and it might not be large enough to be a real issue. However, the point is that the system is flawed.

[macacoalbino]

Yeah but dont forget that both can be cheated: point trading, insta qs resigns, accounts made for noob bashing, etc..

[Gendarme]

insta qs resigns

?

[macacoalbino]

It lowers the account pr/elo artificially and raises random ppl as well, which leads to the same effect to what umeu said about pr decayed accounts

[lemmings121]

insta qs resigns

?

"oh shit that chinese lagger dude again and on great lakes? menu>resign"

[lemmings121]

also, there are those team-noob-bashers, who only plays team games against subpr20 players, and from time to time they resign 50qs games to keep their pr "noob-bashing-friendly"

people are weird...

[tedere12]

insta qs resigns

?

"oh shit that brazil lagger dude again and on great lakes? menu>resign"

[Gendarme]

Zoi approves.

[lemmings121]

"oh shit that brazil lagger dude again and on great lakes? menu>resign"

You should stop playing kickass and play one game with me, so you can stop this bs

[deleted_user0]

The point (that I think Umeu tried to make) is that skilled players have low PR as well, due to decay. If you are, let's say, a lieutenant colonel (PR 33), you can easily lose against Umeu (PR 25), which will cause you to lose a lot of points, even though you really shouldn't. Even if you might not yourself ever play against someone with a decayed PR, you will play against someone who has and thus has an inaccurate rating, and this will cause your rating to be inaccurate after the game as well.

Of course, the magnitude of the inaccuracy in the system depends on the activity of the decayed accounts, and it might not be large enough to be a real issue. However, the point is that the system is flawed.

Exactly, but its more than just that. Basically, in order for the system to be healthy, it requires the big fish to eat the little fish. But thats no longer really happening. Many top players are not really inactive, but they play alot unrated games. So there are now fish that they would eat, which arent, and these fish grow bigger. The few top players of reasonable rank that do still play alot of rated get inflated ranks due to this because they are eating fish bigger than they should be because theyre not being eaten. And then there are tiny blue piranhas floating around the waters that can somehow eat an entire shark, and these tiny blue fish complicate matters even further. The entire system is corrupted now and far from accurate, but this is true for elo as well as pr. But pr has more things that make it inacurate, such as homecity levels and activity affecting points gained or lost, as well as there being no difference made in team or 1v1

Also elo punishes you far more for losing to lower ranked than pr does, so its much easier to maintain a high pr than high elo, which is why the lack of overal activity has actually deflated elo ranks a lot, while it hasnt deflated pr as much. But in both cases its impossible to reach levels of a few years ago without cheating, there simply arent enough games to be played for that

[__Uhlan__]

there both inaccurate lol. Any top player, pr36+ can sit in qs for a few days and get top elo/pr fairly easy.

[deleted_user0]

Theyre both inacurate, but pr much more so lol.

Also its not realky true, many players would suck at qs, and its more than just a matter of civ choices or map screws. @princeofkabul can explain why

[princeofkabul]

because too many players follow "meta" no matter what, which makes them rather easy to beat.

[Forjustice]

In my opinion,Elo is very accurate for the players who is new to this game and active.PR is very accurate for the player who is active and get the higher civ.

[pecelot]

Yeah but dont forget that both can be cheated: point trading, insta qs resigns, accounts made for noob bashing, etc..

Elo is much less affected by that

[momuuu]

Tbf elo has a decay problem too in the sense that people that have been inactive are never worth their high elo rating and that at this point people sit on the high elos more than anything. I feel like elo does swing pretty heavily compared to pr, a few good games make you go from captain to high major elo sometimes. At lower ranks both seem rather accurate tbh but both also have their pros and cons.

I think if elo wouldnt be so extreme itd be better, at least for aoe.