Way to make ELO matchups more similar?

[FrankJones]

The player I was referring to is not what I would refer to as "holding one's own". This player rarely wins. 23% win rate, so, slightly higher than 1% probability of winning 3 in a row. To do that 4 times in one arena season and get to Elite... not easy.

It's a bit absurd to take as example games they have played less than 20 times. In Ticket to Ride Europe, they did qualify for Elite league and they've been holding their Arena score across many games, which means them qualifying was not a fluke.

Mathematically, their chances against top players are not very high, but they are not negligible. At a gap of 400 Elo, it's still over a 5% chance chance that the lower rated player wins. If it was me, I'd try to capitalize on that and do anything to make it happen.

I think we have gotten ourselves talking about two different things here.

I was referring to the lowest ranked Elite arena player, who has played hundreds of times. When you said the player was holding their own, I misunderstood, I thought you mean that player, the one who is ranked lowest in Elite, not the person posting in this thread. My mistake. The player I was referring to is the one with the 23% win rate, across hundreds of games.

With regard to that, I am surprised that player managed to make it up 4 levels from bronze to Elite. I mean, advancing 4 levels to Elite and still having the bare minimum 100 Elo is a bit odd.

[FrankJones]

Beating players better than you doesn't mean you have to aim right at the top though. It can start with small gaps first and climb from there. Nothing wrong with the occasional massive gap, but I can see why players wouldn't want to regularly keep running into it.

Exactly.

A sub-200 Elo player is not goi to beat me at Earth. 0% chance. It won't even be a learning experience, because a sub-200 player isn't good enough yet to even understand why the outcome of the game is so lopsided or have the foundation upon which to learn from seeing what I'm doing.

But, in arena games for a game with lots of Elite players, there's no way to avoid running into a player 500 Elo higher. It's just going to happen sometimes.

[tbhp]

As an aside, you said Elite League, not 300s, and making 300s are stronger than making Elite League. Since the example game seems to be Ticket to Ride Europe, we'll use the OP as an example. They are an Elite League player, but not a 300. In fact, they even made Elite League before 200 rating. Getting 200 rating is probably stronger than making Elite League as well. So within the Elite League we're talking 100s vs said 700s. Elite's upper and lower bounds are absurdly far apart.

In every new game you play you start at 0 Elo and you need to finish games in order to climb. So the Elo you have initially doesn't reflect your actual level and the win probability your Elo would mathematically suggest. Only once you've played enough games and you start to struggle to climb higher does it reflect that you've finally reached a number that more or less reflects your current level.

Well OP is in the 200s now and seems to have been there for a while so we can assume it is mathematically correct at this point, and it means they have a 5% chance of winning against people whose Elo is in the 700s (I checked after my previous post and it is actually a 5% chance when there is a 500 Elo gap).

I wouldn't say those prospects are absurdly far apart. Of course, it's possible that among the 200s Elo players who beat 700s Elo players, a big percentage of that data comes from players who are actually better than their Elo would suggest. But how can OP know they are not one of those players if they don't even try and they think they have no chance? Like I said, that's a defeatist attitude. When you look at the numbers and you see 5% of the people in your situation managed to do it, you should think "well, why not me?"

Beating players better than you doesn't mean you have to aim right at the top though. It can start with small gaps first and climb from there. Nothing wrong with the occasional massive gap, but I can see why players wouldn't want to regularly keep running into it.

I would agree if we are talking about a game like Chess where the skill ceiling is absurdly high. I probably wouldn't choose to play against a Grandmaster if my goal was to improve at the game (though I would like to try it at least once and see how it goes).

But with games that top out at 800 Elo, I feel like the window of skill is much narrower. It doesn't seem crazy that someone who is relatively new to the game could outplay a top player. It regularly happens in Harmonies that players who are in the 300s manage to get consistent results against top players, because they want to win and they are willing to do the necessary thinking to be great at the game.

In that sort of game, there is no reason to limit yourself I think. You can aim to beat anyone and aim to win tournaments right away. When I started playing competitively, I finished in the top 3 more often than not and never had a point where it felt like an opponent was out of reach. The problem is that most people who subscribe to tournaments don't seem to put much thought into their games. It's like they show up for 30 seconds, make their move and that's it. It's like they think there is some kind of magical barrier that differentiates average players from top players. But it's not magic. It comes down to how much thought and investment you are willing to put into the game.

[FrankJones]

I think the question of whether a 200 Elo can beat a 700 Elo depends on the luck factor in the game.

[tbhp]

I was referring to the lowest ranked Elite arena player, who has played hundreds of times. When you said the player was holding their own, I misunderstood, I thought you mean that player, the one who is ranked lowest in Elite, not the person posting in this thread. My mistake. The player I was referring to is the one with the 23% win rate, across hundreds of games.

Oh, I see. I reframed the discussion towards OP and I said they were holding their own, in contrast to the lowest ranked player who is not.

It seems a bit crazy to think that a winrate of 23% would mean you have only a 1% chance of getting three wins in a row. But the math checks out.

With regard to that, I am surprised that player managed to make it up 4 levels from bronze to Elite. I mean, advancing 4 levels to Elite and still having the bare minimum 100 Elo is a bit odd.

Sure, but out of the thousands of players with this sort of winrate, some of them are going to get lucky and advance to the next league, especially if they play lots of games. So, you found maybe the one player who managed to get lucky four times this season.

This really shows that like some people said before, advancing to Elite league is really just a numbers game. You'll eventually advance since you can't go down (unless your winrate is so low that you can't realistically luck into it in a span of three months).

[tbhp]

A sub-200 Elo player is not goi to beat me at Earth. 0% chance. It won't even be a learning experience, because a sub-200 player isn't good enough yet to even understand why the outcome of the game is so lopsided or have the foundation upon which to learn from seeing what I'm doing.

I think the question of whether a 200 Elo can beat a 700 Elo depends on the luck factor in the game.

That's an interesting discussion because it illustrates that Elo is just a mathematical model. It doesn't perfectly reflect win probabilities in a linear manner, because in reality every game will have its own curve.

It's possible that there are games where beginner players just don't have the necessary baggage to have a chance against more experienced players, while simultaneously the Elo ceiling is not particularly high because among people who are more experienced at the game, the game mechanics don't afford much possibility for skill variance.

So, an experienced player would be able to win against beginners at a frequency much higher than their Elo would suggest. Or putting it another way, if they only ever played against beginners, their Elo would end up going very high, higher than it would be when playing in a more diverse pool of players. Whereas in some other games, the player may lose against the beginners at the frequency their Elo says they should, so it would come down to the same result in both cases.

So, Elo is not a fixed number even if a player's skill remains the same. Depending on the mechanics of the game in question, it also depends on who is playing who. Different pools of players, as well as different match ups between those players, will lead to different Elo distributions.

It's kinda like the saying that goes "easy to learn, hard to master", which for some games can be "hard to learn, easy to master". Every game has its own unique curve in that way.

Regarding the game you play, I believe Harmonies is a different type of game in that here, it is actually realistic for a 200 Elo player to beat a top player. They have read the same rules, they can see the same board, they may know all the cards that exist. So, I believe that if they really put the thought into it, they have the potential to come up with a great game that can beat anyone.

So even though the games we play have similar Elo ceilings, the win probabilities for each match up may not be the same. And this is not a matter of how luck based the game is either, more a matter of is it experience or just pure thinking that plays more of a role.

[nik592]

OP is referring to arena games.

I thought they were talking about both:

I've noticed lately I'm often matched up with someone whose ELO I don't come close to. For example, in Parks I have an ELO around 30. I don't even qualify to play arena, which is fine, I need to get better first. But my last couple matchups have been players who have close to 400 ELO. I try to learn from what they're doing, but it also is kind of a downer; I can very quickly see I don't have any hope of coming close.

This seems to be not Arena, filters can be applied.

In Ticket to Ride Europe, I just ranked 342 in Arena. I played two games today against someone who annihilated me, and okay, I respect that, but I just got done with the second game and happened to see the rankings on the main page, and this person is in the top 10.

This part is about Arena, which yes, I missed responding to because I kind of read their first bit and went, why don't you just filter?

I get the issue though - for some games, a high skill player can be almost playing a completely different game to a low skill player and the low skill player can barely develop their skills (sometimes because the game is over quite quickly). I usually don't play with Elo filters unless I start encountering those kind of scenarios and then I limit to closer to my own (though usually still a band or two higher than mine, because yes, challenge is good).

[FrankJones]

This really shows that like some people said before, advancing to Elite league is really just a numbers game. You'll eventually advance since you can't go down (unless your winrate is so low that you can't realistically luck into it in a span of three months).

True, especially for a game that can be played in 6 minutes, apparently. (I've never played it.)

[Jellby]

I think the question of whether a 200 Elo can beat a 700 Elo depends on the luck factor in the game.

If you accept the elo model, the probability is 5%, no matter the type of game. Call it luck factor, call it probability of having a bad day, or a phone call interrupting you.

The luck factor comes in what 200 and 700 mean. In a high-luck game, 700 is probably the most you can ever hope to reach, and even then it would be overestimated. In a low-luck game, 700 is probably just high-average.

[FrankJones]

I think the question of whether a 200 Elo can beat a 700 Elo depends on the luck factor in the game.

If you accept the elo model, the probability is 5%, no matter the type of game. Call it luck factor, call it probability of having a bad day, or a phone call interrupting you.

The luck factor comes in what 200 and 700 mean. In a high-luck game, 700 is probably the most you can ever hope to reach, and even then it would be overestimated. In a low-luck game, 700 is probably just high-average.

Well, I think you got me here. Dang.

I have vociferously defender the Elo model in the past.

But I see your point, so I now must (partly, anyway) reject the Elo model.

Because, a 200 player at Earth does not have a 5% probability of beating me. 1%? Maybe. 5%? Definitely not. In 250 games with the abundance expansion, I haven't lost once to a 200 Elo player. I don't think I've lost to a 300 or lower player. 300+? Yes, I've had some losses.

Terraforming Mars, on the other hand, apparently has such vast disparities in corporations that a 200 Elo player can get Point Luna and beat a 700 player who has to choose between UNMI with a terrible hand or Phobolog with a terrible hand.

In Earth Abundance, there is no such starting disparity.

In Pente, I am never losing to a 200 Elo player. (Pente is a zero-luck game, which adds some further discussion I suppose).

So, perhaps this goes back to what ChiefPointThief (and others) were saying in a 20-page thread awhile back: The Elo model is imperfect when games have varying amount of luck.

In another game, I believe myself to be 700 level, though I have not reached it yet, haven't played enough games), but in that game, I believe three are enough variance factors in the game mechanics that I could lose to a 200 Elo player.
(It's also possible that game has too much potential variance and luck within a single game for anyone to maintain a 700 Elo, which also ties in to that previous long thread.

Maybe I'm changing my mind somewhat on some firm stances I expressed in that other thread, though I'm not sure.

Regardless, that's part of the risk and gamble inherent in arena. We may get matchups we're not happy with for one reason or another. So, unless the arena filter settings are changed (which is unlikely), there is no way to guarantee more similar Elo pairings.

(Which was the original subject of this thread, I think.)