Introducing an Elo-type system for alliance wars, like the trophy system for raids, would be the best way to address the matching issue.
(Of course, then we’d get the same complaining we do about raid matchmaking: “Why were we matched against an alliance so much stronger than us?” Because you’re good at wars, Nimrod.)
I don’t like the idea of summing players’ top 30 heroes for matching. It removes the incentive to level those heroes, because levelling would just bring you more difficult matches.
true, but it would match you against appropriately benched alliances. And people are going to level their heroes anyway, it’s in the nature of the game as it will make titans, raiding, etc better.
While ELO is good in concept, not sure it’s practical for AW which is only twice a week. You get balanced under ELO for Raids on every atttack (and every defense), so that is 6 or more a day just on attacks and potentially a lot more than that on defenses. This gives lots of input for the ELO system to prove reasonably reactive.
With only 2 encounters a week, however, the balancing will take ages, and you’d get a lot of potentially ‘unfair’ encounters in that period while balance is achieved. Even so, substantial changes to your alliance would then unbalance the ELO sytem, and it will take a while for you to find your new level.
I think @Dante2377 / @Revelate’s suggestion of assessing top 30 chars would work. If people choose to not level their heroes for the purpose of their AW position, then that doesn’t really harm anyone… and really they would be doing so to their detriment in other aspects of the game, so realistically I don’t see it happening much, so don’t see it as a problem.
(Edit:
Don’t know why I type ELO in caps. It’s more correctly the ‘Elo’ System. Explained Here.
Edit again:
Yes, @Brobb, that’s probably why I typed it in caps, but the part I don’t understand is that I had already read up on it, and knew it wasn’t an acronym. Old habits die hard. (Gee… found a new way of communicating in a thread without adding more posts… yay us ). )
Your criticisms of an Elo system in this context are correct. It was a dumb suggestion by me.
I’m still uncomfortable about the idea of summing top 30 heroes, but maybe I’m being irrational. I certainly don’t have a better idea. I’ll go away and think about it for a while, suitably chastened.
(Edit: you probably type “ELO” in caps for the same reason I often do - it feels like it ought to be an acronym.)
I’m not entirely certain how to solve the prep time merc issue other than to lock the potential war participants at effectively the beginning of the prep phase (current implementation, they could match after prep as someone suggested too though this might be an intensive process and you’d want to separate the two from a design principle as they did currently).
My alt is helping out a top 50ish alliance because they had gaps suddenly occur during the prep phase, and I managed to score just over 100 points which currently is in the top 10 for that fight now. Admittedly for me this is just a rare occurrence helping out some friends but it does illustrate the fundamental theory that this could be a severe problem when taken to the logical conclusion with even my matching idea.
Not sure what the right solution is, while I’m not mercing’s biggest fan I’d rather it live or die on it’s own merits and not because it winds up potentially breaking Alliance Wars.
Yes, but if it kills the match making for non-mercing alliances (which is almost all of them and what it was designed for) during alliance wars, then IMO an alliance would have to choose to not merc during AWs, or not have those players available for AW.
I think merc’ing takes backseat to other design considerations. If you can have it and not break stuff, go for it. From the devs point, it seems like merc’ing is tolerated but not designed around. Don’t see why that should be any different here.
Mercing should be fine, but the important thing is that the matching be done with the teams that are actually going to participate. If 4 allies are off mercing when the matching is done and then return, their alliance is going to be stronger than the matching system anticipated.
So, if you’re on the team at some magic time—be it 24 hours or 1 hour before the battle phase starts—and that’s when matching is done, then you can still leave to hit a Titan, come back and participate. If you weren’t in the alliance at matchmaking, then you can’t participate.
So to address this concern, do a tier reward system… if your total top 30 hero level is on average >3500 you will be matched in tier 1with similar strength alliance. If your average is >3000 but not 3500, tier 2 with great rewards but slightly weaker. Tier 3 >2500… etc. this makes matchmaking fair and gives incentive to level up heroes to top. Done
Our alliance war (ours is in the 90k range) was matched with a very competent opponent. We won by almost 400 points BUT it is because we figured out the strategy to maximize points very quickly.
When the matching works I think AW is brilliant; so far in Departed we’ve matched up against HHR and SW and all three alliances are unquestionably in the top 5 or so in game.
The problem is alliance score is a terrible metric: there’s another alliance, Mixed Nuts Extra Spicy who has an artificially low alliance score which isn’t going to be fixed with their current cup-dropping culture (and they would probably wreck the alliance if they tried more than their temporary cup pushes TBH) who is also in the top 5 and their matchups have been utter slaughters as I understand it.
Also they happen to merc quite a bit (though I suspect AW has crimped their style) and if the match ups are done when they’re at 22/30 members or whatever before they call everyone home, that’s going to break any system done on alliance score matching.
I’m not picking on Spicy in particular: FWIW I have immense respect for them as an alliance and their individual members as well; however, they’re an easy example at the high end of where the current matching system appears to break down, and if something more sophisticated isn’t done it’s going to be broken going forward as they aren’t the only alliance in this situation. Full disclosure, two of the 7D umbrella alliances have lower alliance scores than they might otherwise have as well if the matching is done on alliance score and it’s going to break hard on lower level alliances which are all over the map.
Ultimately, if the game isn’t fun at the lower levels, and I’d suggest AW mismatches aren’t fun, it’s going to be a serious problem for SG.
SG might have a better matching system already planned or even already implemented (if the current matching is just seed rounds for a more sophisticated ranking system later, they basically have that with cups already so it’s not inconceivable it will improve with time) but they aren’t saying and right now it really appears to be based on alliance score and to repeat myself in closing: that’s an awful metric with the current game mechanics.
IMO it should be what you proposed two weeks ago: aggregate power (or hp) total of the top 30 heroes of each player in the alliance. Match based on that.
BUT in order to account for merc’ing alliances or people who would sandbag right befoe prep time, either require the player to be in the alliance at the time of matching (right immediately when prep time starts?) in order to participate OR keep an aggregate of the top players across of a week and use that total. That’s easier on the players but much more computationally intense on the server/AWS side.
That will allow a much better metric (Total power or hit points) to be used than alliance score and account for alliances who merc OR alliances who would try to game the system. (Note those are 2 different things, if people want to merc, go for it. I just don’t think it should be allowed to create mismatches, either intentionally or un-intentionally, in AW matching.
Yeah I don’t see any other good way for the matchup calculation; I still think my idea is the best solution for both short and long term success; say for example Departed continues on it’s win streak, it really doesn’t make sense to match us with anyone outside the top 20 which would likely happen in the short-term, even one on a similar win streak when we’re talking elo system: it’s just going to be a pointless slaughter with the current game mechanics.
I’m not even sold on an elo system for the long term given the frequency of wars is so low and alliances can rapidly increase in power in very short order especially when you see an alliance split vis a vis Metallica back in the day: how does it make any sort of sense to match them with other 0-0 alliances when they reform into something new?
Ultimately I think they need to address both of those issues, and to make a functional matching system up and down the leaderboard I think they have to address the merc issue somehow which probably as you suggest means they need to do the matchup time differently.
I joined on my alt mid-prep phase to an alliance and scored in the top 10 overall and this occurred in a top 100 war; I’m nowhere close to stacked compared to many mercs (especially those in the aforementioned Spicy, hell many of them are more stacked than my main) and this sort of thing is problematic for matching when extended towards the obvious conclusion when it comes to matching.
Not sure it can be fixed really without just leaving mercs and the alliances that use / support them out in the cold from an AW perspective… or just killing mercing altogether, but I don’t really think either of those is an optimal solution from a community perspective either.
I agree on both points - at the top, it prob makes sense to cap it at top 20 or 30. Still that’s not repeating the same opponent until 10 or 15 weeks out. But that’s a problem just for a few at the top. Also agree the elo system only works well IMO with volume of data points where the “skill/power” of an opponent changes more gradually over time. Chess players aren’t jumping from expert to grand master in a few months.
). )
.