Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[timchen]

I think Colony games are just more rewarding to engines, and it's harder to put together an engine well than it is to play big money.

In addition, in colony games there is almost never a situation when one can edge out a win when the other player spent too much time building too much. (I mean sure if you insist it is possible, but it is a mistake that is a lot harder to make.)

The reason behind this phenomenon is just that in lots of colony games the platinum is such a great card.
And it make drastic difference to your buying power, which makes the colony buying decision a lot more obvious.

Compared to a money oriented province game,  here the winning play is to start buying provinces with uncertainty.

[Geronimoo]

Province vs Colony or how Silver is broken...

Silver is maybe the most efficient card in the game: you only pay $3 for a card that will give you $2 each time you play it AND there is no restriction to playing it because it's a Treasure. A hand like 3 Silver, 2 Copper is super easy to aquire very fast and = Province .The power of Jack is due mostly to the power of Silver. When we get to Colony games there is no all Silver hand that'll allow you to buy Colony, so a deck can not rely on Silver anymore and needs to branch out.

[paddyodoors]

Certainly not everyone plays with veto mode. I understand where rrenaud is coming from, and I also get where you're coming from. I think you're both making valid points.

Not that this doesn't relate to OWB, but at this point, could we get this forked over to isotropic discussion? I'd like to make some more points on the topic, but it does feel a little weird to post them here.

WW, are you still looking to make more points on the topic, as you said?  I am interested to hear your thoughts.

[WanderingWinder]

Certainly not everyone plays with veto mode. I understand where rrenaud is coming from, and I also get where you're coming from. I think you're both making valid points.

Not that this doesn't relate to OWB, but at this point, could we get this forked over to isotropic discussion? I'd like to make some more points on the topic, but it does feel a little weird to post them here.

WW, are you still looking to make more points on the topic, as you said?  I am interested to hear your thoughts.

Yep. Busy couple days, long post. Will try to do when I get home.

[WanderingWinder]

So, the thing with the ranking system - what is it that you want it to do? You want it to tell you the relative strengths of the different players at the game. But what is the game? This is the issue. To a certain extent, every single different board is a different game. So, somebody's probably the best in the world at playing 'the first game' somebody else at set Y, etc. etc. etc. But those don't come up that often, so you want to have a ranking that's somewhat independent of the board. I mean, I think that ideally, you have different ratings for different system. I have a prosperity-bias rating, a full random rating, maybe a preset-kingdom rating, a veto mode rating, etc. etc. I should certainly have different ratings for different numbers of players in the game - somebody who's really good at 2-player might not be so good at 4-player. I think you have different ratings to reflect this. Now, the isotropic system tries to combine all these things into some kind of overall rating. You could look at having the averaged out ratings of all possible boards, or something like that, as an overall rating anyway. But the question is what to include here, and how much to weight different things. Now, some people, let's say many people, have the idea that the game is really about how to react to any board, with no restrictions, yadda yadda yadda, and that should be totally random. That if you're not playing with attacks, you're missing a big part of the game. That if you're manipulating the kingdom in any way at all, you're shifting off of the overall, average-out skill that they think rating should reflect. I agree with them, but it's not that man, they're obviously right. It's that that's what I want out of my rating system. Different people want different things, and that's fine. So anyway, I think there ought to be different rating systems, and then people are just wanting different things out of the same couple numbers, and you just can't make 2 numbers be 12 different things. Ok.

[WanderingWinder]

Now the specific issue of OWB/biasing prosperity. First off, biasing prosperity does more than just add more colony games. It also gets you those cards more often, so I'd guess he knows them more than other people, better than other people, and that gives him a bit of an advantage. It also helps 2- and 3-card combos to hit more, which helps too; you recognize them, know how to play them, know their strengths. So like, I could bias intrigue all the time, and I'd know how to use the intrigue cards better, yadda yadda yadda yadda. So anyway, it inflates the winrate even when colonies aren't present.
Ok.
Now, the idea that colonies don't make a big difference is pretty laughable to me. The idea that they just help the engine player is a little off, but largely right. The idea that they make games longer and that just helps the better player, well, that's pretty messed up. It promotes different things. It promotes a different style of gameplay. It promotes different strategies and combinations. Not always ones that are harder to do, not always does it reduce variance, not always does it help the stronger player. It just doesn't. I mean, platinum and colony have consistently consistently been in my bottom 5 win rate given availability cards for a very long time. Now, I guess this might just mean that I suck, and I get really lucky all the time in the non-colony games. But I don't think so. It's just a different skillset.
Ok, now for OWB-specific stuff. Even if we assume that the only difference is playing with colony and platinum, which as I point out above, isn't true, then let's look at these numbers. Pulling the data up today... He has 4393 games, overall winrate of 1.27. 2566 games with platinum and colony, winrate when those are available is 1.32. Okay, this means that there are 1827 games without platinum and colony, and his winrate in these games would be approximately (can't calculate exactly because of rounding) 1.20. If we scale that down to 19% (I can't tell what the exact percentage should be, because for some reason the availabilities for your basic cards aren't the same in the overall data), we'd get a winrate of 1.22. Which is a pretty big difference. I can't tell exactly how that affects his rating, but it's going to knock it back down a couple levels, at least, I expect. But as we all know, he's a top player anyway. Just not the clear #1 that the board makes it look like (unless you think that the current rating system take everything into account totally perfectly yadda yadda yadda see my last post).

[WanderingWinder]

Oh, also, I don't think this is going to be a big deal for iso. I mean, what are you going to do with all these games that we have? It's a big issue without a real solution. But I hope that the new official server has different policies here, as in having more rating pools, having games where you can't see the cards before you accept, having something for full random, etc. etc.
I hope it also has a spectator mode, where you can watch other people playing. These are my two biggest wishlist items for the official app in comparison to what we have on iso.

[ecq]

Province games are shorter and simple strategies tend to be more competitive.  Big Money + X wins a lot of the time, and a level 0 who goes first and picks the correct X can give a top-ranked player a run for his money on a healthy percentage of boards.

Colony games have a lower percentage of boards where simple Big Money + X wins.  Further, Prosperity contains several cards that would make you favor at least Big Money + X + Y, if not a full-blown engine, so simpler strategies aren't as competitive.  Weaker players have fewer chances to cause an upset, so stronger players would have fewer rank-damaging losses.

That's my guess, at least.

More importantly, though, I think skill comparisons based on level are pretty meaningless among the top players.

[paddyodoors]

... More importantly, though, I think skill comparisons based on level are pretty meaningless among the top players.

Good point... I never thought of that, but it seems true.

[olneyce]

man, i like that you are friendly and civil but i disagree with a lot of what you have to say. 

a few things:

• you are right, we don't know exactly how much influence biasing and veto will have on levels. but i'm not sure we can dismiss it as being negligible either. by my math, OWB wins 6% more colony games than he does province only games and he makes it so that about 60% of his games have been colony games. we can argue on whether a 6% change is a "small amount of noise" or a large amount, but it is something.  and i'm also not sure i would call TrueSkill a "very noisy metric". 
• i still don't think the fun aspect should matter as much for competitive play.  but if you want to bring it up, just remember that it can cut both ways. i prefer playing with a fully random set. no automatic vetoing of attack cards.  no bias towards sets. i prefer the random mix of games, and i think that the current IsoDom challenge helps demonstrate this.  how many match reports in the first round were basically hey there's witch/BM again.  ditto masquerade for intrigue.  should some people enjoying a bias give them preference over my preference for a fully random set?  and that is before we even consider if that bias gives them any advantage at all.  winning games is going to be more fun than losing most of the time, so any change which cuts into that is not something i will be eager to embrace.
• it is a game and it is supposed to be fun.  but that doesn't mean that we can just cast aside the competitive aspect of it.  chess and baseball are just games too. fun should always be a factor, but you can't enter in to a competitive environment and expect that people aren't going to care about advantages.  if you play in a softball beer league yeah everyone there is trying to have a good time but they are still gonna get mad if they catch you trying to sneak a lead from first base.

Yeah, I don't really agree with this, but I totally get where you're coming from.

The main point I'd make is that (for the most part) you already get the type of games that you want, and you get a ranking system that reflects performance over time.  The lack of a truly-random option is unfortunate, but if you set it to no-veto and just bias for ALL of the sets you'll get something fairly close to random.

A world where bias or veto games stop counting for ranking would impose a far more strict constraint on people who feel otherwise. 

As for competition vs. fun, I clearly agree that there needs to be a (fairly high) baseline of competitive balance.  And I absolutely would reject anything that would interfere with the internal structure of the game (like a handicap for low-ranked players or something).  But the unofficial/official rankings on Isotropic?  Sure, it's important that they genuinely reflect a general level of skill and performance.  But I just can't get too worked up about them missing the 'true' level of someone by a few slots.

And I do think it is a pretty noisy metric.  Am I the third best player or the 35th?  Within the course of about two weeks I was in both slots.  I don't think my skill actually changed much there.  I mean, in the grand scheme of 10,000 players, that's still fairly precise.  But we're having a conversation about the very top and how people compare against each other there, right?

[timchen]

Is "yadaa" the same as "you know" or "blah" or "etc"?

[rrenaud]

I'd say yadda yadda yadda ~= blah blah blah, maybe it slightly connotates more bored and less agitated.

[WanderingWinder]

Is "yadaa" the same as "you know" or "blah" or "etc"?

Basically yes.

[WanderingWinder]

And I do think it is a pretty noisy metric.  Am I the third best player or the 35th?  Within the course of about two weeks I was in both slots.  I don't think my skill actually changed much there.  I mean, in the grand scheme of 10,000 players, that's still fairly precise.  But we're having a conversation about the very top and how people compare against each other there, right?

There's two major things going on here that cause these massive fluctuations. First, you're looking at the ranking which is based on level. Really, TrueSkill gives you a ranking that ought to be based off of the rating. And because you're uncertainty rises every day, whether you play or not... this has a big impact. Also, if you look at this range of the list, and in fact, lots of the list, everybody is pretty well bunched up. That people are relatively similarly-skilled does not necessarily mean that the skill rankings are wrong. They could well be right - if the people are relatively similarly-skilled. Which, to me, it seems they are.

[WanderingWinder]

you are right, we don't know exactly how much influence biasing and veto will have on levels. but i'm not sure we can dismiss it as being negligible either. by my math, OWB wins 6% more colony games than he does province only games and he makes it so that about 60% of his games have been colony games. we can argue on whether a 6% change is a "small amount of noise" or a large amount, but it is something.  and i'm also not sure i would call TrueSkill a "very noisy metric".

Actually, you're underestimating things here. See post #80 for the maths, yo.

[greatexpectations]

Actually, you're underestimating things here. See post #80 for the maths, yo.

not underestimating, just different.  i think we have the same numbers we just used different methods and wording.

 i went and converted the win rates to a number of wins and then calculated a % of games won instead of using the 1.xx win rates.  i got 60% for province and 66% for colony.  that is a 6% difference as i tried to describe it.  convert that to win-rates and you get 1.20 and 1.32, same as you got.