Does it? I think you don't really understand what you are talking about. Even a 100% chance doesn't offer anything significant.
Crude estimate with a lot of bad assumptions: A) cost of loss to @ -400 ELO: -25 B) cost of loss to @ -50 ELO: -10 C) marginal gain from successful selection (B-A): 15 D) estimated successful selection rate: 0.50 E) estimated loss rate: 0.20 F) minimum # of matches in season: 130 G) marginal gain over course of season (C*D*E*F): +195 Note that (G) is larger than the spread between 1st and 8th place. Also note that there are publicly-known techniques (primarily queue-canceling) which offer a 100% successful selection from a certain band of ELO, if not of a particular partner. The marginal gain is probably overestimated slightly, but you can analyze real data if you care to provide a better number. The loss rate is likely underestimated, given the weakness of top-tier builds to firestorm which is commonly found at the ~1300 ELO band.
I didn't "blame" the pvp season, as I clarified that selective queuing happened before it. I just said that the season encourages speculation in a very visible manner, beeing selective queuing a part of this problem. If you do selective queuing all the time, it means that you are playing most of the time with high-ranked players. That also means that you can build your deck knowing that you are probably only going to face 1 or 2 kind of builds (the ones that are used at the top rank), discarding the things and items you don't need to face those builds but that you would need to play against other type of builds. That way, you can increase the effectiveness against the builds you're going to play with and nullify the fact that you would do poorly against other builds because with selective queuing you can, most of the times, choose your opponents. So in your case, losing to B is almost equally likely, and far more risky. By playing only against high ranked players you guarantee yourself to keep at the top rank, because even when you lose you don't lose much, and you can easily win those points back in the next match. On the other hand, no selective queuing means that most of the times when you lose you lose a lot, and when you win you win very little. So you don't have to be a visionary to see the benefits of selective queuing, and I think you already know that. Edit: I removed an example I used with -REDACTED- about his win rate because he got offended with it, when in fact I was just making an example with his case of how selective queuing can affect rating. Anyway, it's done.
I want to let something clear here, because maybe my previous opinions were misleading. First of all, I don't have any intention of calling anyone a cheater here. As far as I know, it's not really against the rules, at least for now, to do selective queuing, so it wouldn't be cheating when someone does that. Secondly, my opinion of selective queuing being unfair is in this case directly related with the pvp season: if there was no such competition, I wouldn't mind who tries to select their opponents. I still wouldn't do it, but that is a matter of opinion. But, given that the pvp season gives a reward at the end of it, and that selective queuing affects rating (easier to stay at the top if you are already there, or easier to climb up the rank if you are below the top rank), I wanted everyone to have the same chances of winning, so that is when selective queuing becomes a real issue. Because players that do it have better chances than the rest. This is the crucial point of the debate in my opinion. Should selective queuing be allowed, as it happens now? Or should it be forbidden? That is the question.
The thing is that in a game with no luck selective queuing shouldn't matter -- someone at 1000 Elo should be twice as easy to beat as someone at 2000 Elo (if you're at 1500 Elo). However, when you're playing a game that introduces luck into the mix then selective queuing can be used to minimize losses and maximize gains. Someone who is 2000+ Elo can lose to someone who is < 800 Elo in theory if they just never draw attack cards. It is EXTREMELY improbable, but it happens. That's the problem with selective queuing EVEN if you're not building a deck just to counter someone else (which in itself is hugely problematic -- I mean look in a tournament someone selectively built around an enemy deck for an easy shut out).
And that's exactly why selective queuing is beneficial even for a high ranked player. They can eliminate the risk of losing a lot of points for an unlucky game, not to mention that also they are avoiding some builds that they are not prepared for.
Oh yes, I've been against selective queuing for a LONG time (all the way back to SLG's tournament). The only reason I started trying to do it was to combat it -- if you can't beat them, join them (and then beat them at their own game).
That's also assuming the lower ELO's have played enough games to have settled at that rating. Meaning that they should be rated 1000 or whatever, due to their expected win/loss record. ELO works with a well established field, where everyone has played enough games to settle at their appropriate ratings. Cardhunter has plenty of players dropping in and out all the time. I've played a handful of MP games since release, and am currently sitting at around a 700 ELO. You would falsely assume that it has some measurement on my win ratio, when in fact I've won every MP match I've played. My ELO is more a measurement of my MP activity than anything else (if we had degrading ELO you would see even more of this). So much of your ELO is tied up into the items you have, the deck you use, and how that deck is positioned in the metagame (as well as how much you play MP) that I would be careful about extrapolating the probabilities of winning based strictly on ELO in the current system. Experience does matter as well, but ELO is an abstract measurement at best when used in this type of setting.
But waiting to queue against a certain opponent means you can see what they are using right before your battle. It is like knowing what someone is throwing before they do in Rock-Paper-Scissors...
Oh I'm not arguing the value of this, it protects your ELO. And ELO currently has potential value due to Flax's league/season. Gamers are going to game.
Yep, I'm really feeling this now. I can't get a game for love nor money now I've dropped rating, and it's going to make it impossible for me to finish the league.
So can anyone tell me why this isn't simply colluding? Because we can already report players for that.
Because it is only technically colluding if you arrange matches to win or lose. Selectively queuing isn't colluding by definition.
Well, I agree to some extent. While I wouldn't call selective queuing "colluding" myself, I still think it makes the whole competition unfair. The fact the game is influenced by luck is entirely irrelevant, and a weak argument. Everything in this world is luck-based at least to some extent. While luck might occasionally be a deciding factor in some forms of competition (say, various tournament formats), it's definitely less of a problem in leagues. The fact you get the chance to play lots and lots of games over a long period of time makes luck entirely irrelevant, since over a large enough number of games luck evens itself out (and every competitor is affected by luck in the same way) so, in the long run, the better player will have won more games. But, leaving luck aside, purposefully trying to influence the system in order to gain an advantage... well, I wouldn't call that fair. And I'll leave it at that. Now, it's obvious the main problem with the system as a whole is the fact the game lacks a large enough community of active MP players. If we had dozens of players queuing for a match at any given time, selective queuing would be way less effective. It's not like Elo-like systems haven't been known to have problems, or to be inducing players into all kinds of fishy practices either (http://en.wikipedia.org/wiki/Elo_rating_system#Practical_issues) As a very practical counter-measure, I'd suggest modifying the RMM in a way the very top players (say, the top 50, or something) are more likely to be matched against non top players opponents. A league is not a brackets tournament, so the top players aren't intended to only be playing among themselves anyways, so I think that would be entirely fine.
I don't know if this would be a good solution. Even elo games are supposed to be interesting because there shouldn't be too much item advantage for anyone, as opposed to what happens in uneven elo matches (of course this isn't always the case). And I think top ranked players like playing against other top players, you can't just add a rule only for them because that wouldn't be fair, so the RMM modification would have to be for all players in that case. I agree with all the rest though. Imo, it's unethical to do selective queuing when you have a competition that rewards elo. The simplest solution would be if BM declared selective queuing as a form of colluding, so when we see that 2 players are according to queue at the same time in the lobby chat, we could report them.
Why? That's the weakest form of selective queuing. The better is spectating a game, seeing what the enemy is running, making a build to counter it, and then re-spectating the game until it ends ... THEN queuing. It doesn't mean you'll match up against them, but it does significantly increase your odds (most players re-queue shortly after a match).
Yeah, you're right there. But would you like the possibility of being unable to play against other top players? I mean, I like playing all kind of players with different ratings, but I also like, without any type of selective queueing, being paired with another high elo player once in a while . That's why I don't think Brandreus' suggestion would be the best solution for this.
