PDA

View Full Version : Matchmaking and rankings needs an overhaul



Morgenstern72
03-02-2013, 05:47 PM
To train new units effectivly I just had found out that you must go back to rank 1 and fill your units with rookies. Otherwise new units have as good as noc chance to make kills in a high rank team.

Problem: I have >50 matches done now and when i play with a rank 1 training squad I get new rank 1 players as opponents. They have no chance at all. Since the cahnce to get newbies that way is quite big you could climb up the ranks quite fast and train new units very effectively. But thats no fun for beginners!

So I would like to suggest that it has to count too how many matches you made so far. At the moment it seems only rank is important in matchmaking.

I have no idea how this could be done, but there should be something like a "Handicap" in golf. Something like
Higher Count of Matches / Lower count = some handicap you get

Example (rank 1)
60 matches (experiences player) / 3 matches (newbie) = Handicap 20

The Handicap could be divided / 2 (in this example 10) and this is armor AND strenght malus the better player gets.


I have no better idea at the moment and I will not exploit the system but I will train new units in rank 1 teams because otherwise it takes ages. it would be nice if new players would have a fair chance, at least in their first 20 fights.

What do you think?

franknarf
03-02-2013, 06:16 PM
Right now, team power (sum of unit ranks) has a lot of weight in determining whom you're matched with (as it should). But I agree that maybe something should be added to the matchmaker to make newcomers far more likely to be matched with each other than with old hands. I haven't really used zero-level units beyond my first couple dozen games, so I don't know how important this issue is.

I think just applying something special for n < 20 fights, as you say, makes sense and might help a lot (as opposed to making n matter for everyone). Stoic could do this by making a 0-1 for whether n < 20 and making it worth, say, 200 Elo in terms of their expanding "best match" window.

I'm not sure what you mean by "rankings need an overhaul" in addition to matchmaking. That doesn't seem to come up in your post.

Morgenstern72
03-02-2013, 06:25 PM
I mean the highscore lists with ranking. I think you can exploit the fact that you will get matched with a newbie when you build a rank 1 team and so you can win much more easily.

So maybe you should not get any points in highscore terms and much less renown when you match someone that has 1/10 of games than you (or even less). A Handicap would seem better since then every match really is a challenge.

franknarf
03-02-2013, 06:34 PM
Ok. I think a cleaner solution is just to eliminate the exploit by making those matches less likely. Besides, you're not going to get a very good ranking that way; the elo gain from victory naturally falls off with the difference between your elo and your opponents (and new players start at 1000).

Global solutions (like checking whether n1 > 10*n2) penalize those who have played a lot of games. I think it's better to treat all people above some n, like 20, the same.

Morgenstern72
03-02-2013, 06:42 PM
I thought more in terms of "Matches Won", "Win Streak", "Win Ratio" and so on. And most of all it should be fun for new players and not frustrate them with loosing again and again. Had the same guy right after my first match again, he was devastated to have to fight me again. I tried to help him get inot the game but with 30 seconds rule I can only type so much :)

stoicmom
03-02-2013, 07:10 PM
okay, fellow warriors, for what it is worth from a senior citizen, about all i can do is play base units. when i tried to upgrade even to 1, i was trounced nearly every time. yes, i have played many games, but won very few comparatively. even when i win and remain with whatever mix of base units, very few players are using complete 0 lineup. have won some against noob's, but more times than not, those who are practicing with base units at 0 come back and are victorious the next time we meet. so, all this to say, there is something for everyone in this game and i'm having fun. about the only achievement that i can reach probably now and forever is the number of games played. don't use 30 seconds cause i can barely do 60. i am sure that i am in the minority of why one would even play this game without some strategy skill or tactical plan, but i am learning little (very little) by little. thanks for the opportunity to play;)

KRD
03-02-2013, 07:12 PM
To my understanding, this is already implemented. An experienced player (with an Elo rating to show for it) training new rank 0 units is going to get matched against other experienced players doing the same, because both team power and Elo (as well as using the tournament timer) determine what the preferable opponent is for you.

The problem you describe only crops up when the matchmaking cannot find an appropriate match for you because at the time, there aren't any other experienced players queuing at that team power tier. And so because the game prefers to match you against equal power opponents rather than equal Elo ones, as it should, these uneven matches do happen. Still, isn't the alternative worse? At least this way, new players get to see low rank units played effectively, speeding up their own progression, at the cost of very little Elo in case they do go on to lose.

Edit: Whoops, didn't see your post there, stoicmom!

Morgenstern72
03-02-2013, 07:18 PM
Ok, if its only occasionly its ok. Since I only had newbies I had the impression something is not right with matchmaking.

I clearly do not want any handicap with players that have rank 4-6. Just played against a rank 6 and he complained that my rank 6 was overpowered. It was his 4th match, he bought the units and had no idea what he did or how to adjust stats. Such players should not get any mercy in a honorable fight :)

KRD
03-02-2013, 07:24 PM
I'll drink to that!

stoicmom
03-02-2013, 08:28 PM
ok, so i'm starring at my units in Proving Grounds and thinking how great they all look (upgraded and in color variations). However, i'm not using the upgraded units because i get so soundly defeated whenever i try to use them and frankly get overwhelmed. i continue to think, maybe i would better understand how to use these units if i were matched up with like upgraded units,i.e., if i have a level one backbiter, then i could be matched with a level one backbiter opponent. Could be very wrong in my thinking, but i believe i could handle that and learn from it as opposed to another level one unit being introduced into the frey by my opponent. Please help me understand if this is skewed thinking or a real possibility. Thanks!:o

raven and Gud understand my mentality, good luck to the rest of you!

Jorgensager
03-02-2013, 08:49 PM
First let me comment that playing many games in itself doesn't make you a good player. Most people will improve with a good amount of games, but statistically speaking, some people have to lose if we are to have over 1000 ELO. I.e. this modifier should


Only kick in over a certain ELO threshold (this could be floating ~ say on the top 30% or so ~ to assess general skill levels as we climb in ELO).
Be based on an ELO/<game count> relationship, rather than just the game count. A player achieving 1300 ELO in 50 games is likely better than a person achieving 1300 ELO in 100 games.




To my understanding, this is already implemented. An experienced player (with an Elo rating to show for it) training new rank 0 units is going to get matched against other experienced players doing the same, because both team power and Elo (as well as using the tournament timer) determine what the preferable opponent is for you.

The problem you describe only crops up when the matchmaking cannot find an appropriate match for you because at the time, there aren't any other experienced players queuing at that team power tier. And so because the game prefers to match you against equal power opponents rather than equal Elo ones, as it should, these uneven matches do happen. Still, isn't the alternative worse? At least this way, new players get to see low rank units played effectively, speeding up their own progression, at the cost of very little Elo in case they do go on to lose.


In my experience this is not the case. Matchmaker reliably matches me up with new players after only few seconds of waiting when I play with 4/5 basic units to get kills. It is probably a slight advantage for them to face good opponents who can give them tips on how to play the game, but likewise - too much of this is bad for the experienced players when they don't have to think about what they are doing.

I.e. the matchmaker should stay strict on the team power matchup, but also be stricter on ELO difference (and additionally account for the ELO/game count ratio for the players where that is active)... For all I know, I could be the only semi experienced player training basic units for kills, so that I wouldn't get any games with the above suggestions... so take it with a grain of salt. I have to trust Stoic on this one. :p


ok, so i'm starring at my units in Proving Grounds and thinking how great they all look (upgraded and in color variations). However, i'm not using the upgraded units because i get so soundly defeated whenever i try to use them and frankly get overwhelmed. i continue to think, maybe i would better understand how to use these units if i were matched up with like upgraded units,i.e., if i have a level one backbiter, then i could be matched with a level one backbiter opponent. Could be very wrong in my thinking, but i believe i could handle that and learn from it as opposed to another level one unit being introduced into the frey by my opponent. Please help me understand if this is skewed thinking or a real possibility. Thanks!:o

raven and Gud understand my mentality, good luck to the rest of you!

A problem with this idea is that teams would face equal teams, so part of the strategy aspect of building a good team disappears; Your team should be built in a way so that you're comfortable using it, and know the units you're using and how their abilities can best be used in different situations. The whole idea of trying new builds to counter what is typically used (to get an advantage) would no longer be viable in this situation.

Arnie
03-02-2013, 08:54 PM
Matchmaking looks first into Team Power (we need to start using correct terminology) which is the sum of all the ranks in your team. THEN we try to match you with the closest Elo opponent.
So in a perfect scenario we will find you a match with an equal Power team AND Elo rank.

KRD
03-02-2013, 09:49 PM
I haven't played enough yet to be able to say one way or another from experience, Arnie, but Jorgensager believes that the matchmaking should perhaps wait longer than it does currently after it has found an opponent perfectly equal in team power, but of a vastly higher or lower Elo rating.

Might warrant looking into after John's recovered from the hectic launch week?

Jorgensager
03-02-2013, 09:57 PM
I haven't played enough yet to be able to say one way or another from experience, Arnie, but Jorgensager believes that the matchmaking should perhaps wait longer than it does currently after it has found an opponent perfectly equal in team power, but of a vastly higher or lower Elo rating.

Might warrant looking into after John's recovered from the hectic launch week?

In the matchmaker's defence, there is [AFAIK] no way I can know my opponents' ELO [should be on the banners/matching screen in my opinion, to give a sense of whether you should brace for impact or not], but I highly doubt they are all up there, judging by the way some of them play (they're not all completely fresh, but some are, and there is no notable waiting difference on my side).

Morgenstern72
03-03-2013, 06:34 AM
I still think a hadicap system would be a very good solution. Extremeley rewarding to new players that can beat on old one through fair gameplay. Like in golf (what I do not play btw).

erom
03-03-2013, 12:01 PM
A handicap based on something like Elo instead of # games played makes more sense - otherwise someone who plays a lot but isn't the best player is just going to get murderized by the handicap, while someone who catches on quickly is going to get an advantage.

The question is, when you already have Elo and team power affecting the matchmaker, and the underdog renown bonus... well, I question if we would really better off layering ANOTHER new system over that, versus just trying to better tune what we already have.

Morgenstern72
03-03-2013, 04:32 PM
Erom, you are completely right, it has to be based on Elo.

And maybe it is affecting it already, but I only got newbies for a rank 1 match and they all lost very frustrated.

Finjinimo
03-03-2013, 11:27 PM
Does the underdog bonus only apply to Team Rating?

Perhaps there should be a ELO difference that it applies to as well, if there isn't already.

A handicap to playing might be too complicated a solution. Perhaps a penalty to earned renown if your ELO is significantly higher would help act as a deterrent to farming (in addition to the measures already in place -- ie low elo returns).

So at the moment you can average 7 or 8 renown or something per win, but with a renown penalty, if I have a elo rating more than 200 higher to my opponents then I receive the renown penalty of -2 or something. So I end up getting 5 or 6 instead. Whatever the number needs to be to make it less than playing at my own rank.

I can still play that way if I choose, but I earn less renown for doing so and I am being encouraged by the system to seek out matches of equal strength in both Team Rating and ELO.

Saying that, I actually think the system is working pretty well at the moment.

Jorgensager
03-04-2013, 04:10 AM
Perhaps a penalty to earned renown if your ELO is significantly higher would help act as a deterrent to farming (in addition to the measures already in place -- ie low elo returns).

So at the moment you can average 7 or 8 renown or something per win, but with a renown penalty, if I have a elo rating more than 200 higher to my opponents then I receive the renown penalty of -2 or something. So I end up getting 5 or 6 instead. Whatever the number needs to be to make it less than playing at my own rank.

I can still play that way if I choose (...)

The matchmaker chooses matches for us, so why would it pealise us if there are no other equal strength members online?

Bloodaddict
03-04-2013, 05:36 AM
Hi,

I also thought that it might be good to break out of the "equal team power" requirement earlier based on ELO differences (while I know that there were a lot of complaints in the other directions before).
When the tourney started last week I was basically not able to play the first days because of server issues. Therefore I decided to not join it but rather collect some kills for new units to be prepared better for the next tourney to come. For that I mostly used a mixture of 3 base and 3 promoted units. While it worked well for me in terms of collecting required kills, I think I won those matches too easily. And that was not because my opponents were bad players but just inexperienced players (it was clear from looking at the kills of their units that I played by far more than they did). So I thought that those matches would have been more fun for both of us when they had a team power of 4 compared to my 3.
Another thing that came into my mind related to this are the rankings. While the ELO is probably OK as it is now (I did not gain much because of these wins), it is a different story with e.g. the Win Streak. I am now at ~20 and I do not feel that I deserve it to be listed in the Top20 rating because I got that streak mostly in those "kill farming" battles. I do not feel really bad about it since I am still far away from the "all time win streak" Top20, but I still thought that it might make sense to only count 6vs6 wins in this streak statistic? However, I don't know if this still makes sense once we see level 2 and 3 units, so it is probably not a good idea...

Bottom line of my thoughts:
- With a much greater ELO rating it should be more likely to be matched with a team with higher team power (but probably just 1 higher)
- While ELO rating has a way to deal with wins of experienced vs inexperienced players this in not the case for e.g. Win Streak statistics

eduran
03-04-2013, 06:02 AM
- With a much greater ELO rating it should be more likely to be matched with a team with higher team power (but probably just 1 higher)

That's a good idea. The matchmaker could work like this:

1) Find an opponent with exactly equal team power and about equal Elo.
Assuming that does not produce a match -->
2) Allow a team power of +-1 and adjust Elo accordingly, e.g. you get -X effective Elo when playing against a team with more team power.
Still no suitable opponent? Allow team power +-2 and repeat step 2.

It would probably need some testing (or math :D) to find the right number for X and I guess there should also be a cut-off for maximum team power difference.

What are the advantages of this system? High Elo players who are trying to get kills for basic units or who are deliberately farming new players would get to play against other strong players or weaker players with higher team power.

Disadvantages? Matches with uneven team power, potentially frustrating for the player with the weaker team. However, the person with the weaker team is guaranteed to be an experienced player or a new player who is really good.

scase
03-04-2013, 03:27 PM
Hell to solve this at the base, I think the game needs to encourage people to refrain from promoting their units at the first chance they get. I've kept myself at 3 base and 3 upgraded, swapping out upgraded units until I am fully comfortable with their mechanics before bumping up my team power rating.

People would be much better served if most players got familiar with the game and its strategies instead of upgrading at the first opportunity they get. The game shows a very in-depth tutorial on unit upgrading which is great but, it also gives new players the idea that upgrading right away is a good idea. Which in the end, ends up putting inexperienced players with upgraded units against player who clearly are more experienced and keeps the real noobs from competing with other new players that have just started and have a low team power.

Works beneficially for the game in two ways, it keeps players against people of similar skills pitted against each other learning at a curve not so steep. And since they won't be getting curb stomped it keeps them from getting discouraged early from the game and bailing on it.

eduran
03-04-2013, 03:38 PM
I think it's natural to upgrade as soon as you can. Taking that away from newer players will also take away a major motivator. And once rank 2/3 are live the stronger players will field rank 12/18 teams anyways, which should be pretty much out of reach for newer players.

KRD
03-04-2013, 04:00 PM
We already know what that X is, eduran, it's part of the matchmaking formula. The exchange rate is 200 Elo for 6 team power difference, so 33.33 Elo per promoted unit.

eduran
03-04-2013, 04:11 PM
But does the MM actually use it? Has anyone ever been in a game with teams of unequal rank? If so, how come that using rank one teams you always seem to play new players regardless of your Elo?

Edit:
Just did a quick test. My Elo rating sits at 1304. I hire a couple of rank 0 units and create a team with five basic and one promoted units. I join the MM queue and get an opponent within less than 10 seconds. He has a Thrasher named Thrasher and five basic units, with a total kill count of 17 between all of them. While I can't see his rating the game is a one-sided slaugther and he makes a lot of obvious rookie mistakes. No way is his rating anywhere near mine and certainly not within 33 points. I proceed to win and gain four points of Elo where I should have gained nothing. Why does he have to face me on even terms? If this stays as it is there will be people farming fresh players and cause massive frustration among them.

Morgenstern72
03-04-2013, 04:45 PM
But does the MM actually use it? Has anyone ever been in a game with teams of unequal rank? If so, how come that using rank one teams you always seem to play new players regardless of your Elo?

Edit:
Just did a quick test. My Elo rating sits at 1304. I hire a couple of rank 0 units and create a team with five basic and one promoted units. I join the MM queue and get an opponent within less than 10 seconds. He has a Thrasher named Thrasher and five basic units, with a total kill count of 17 between all of them. While I can't see his rating the game is a one-sided slaugther and he makes a lot of obvious rookie mistakes. No way is his rating anywhere near mine and certainly not within 33 points. I proceed to win and gain four points of Elo where I should have gained nothing. Why does he have to face me on even terms? If this stays as it is there will be people farming fresh players and cause massive frustration among them.

That's exactly what happens and why I started this thread. I did not once get a player with high elo with a rank 1 team, only newbies that were completely frustrated. I normally did end theses fight with 4-5 units. That's no fun for new players. And it would have been easy to end it with 6 units, since 2-3 were always at (almost) full health.

And it will not end when rank 2 and 3 are coming, since you maybe still want to train new units and it's almost impossible to train a unit in a rank 5 team since it gets no kills. How should it work on higher ranks? You have to go down and then you should not face new players with no handicap. Just gives enormous win streaks :(

Jorgensager
03-04-2013, 06:32 PM
Apparently the MM goes after team power before anything else ~ i.e. will not consider ELO if two people are the only ones with equal team power (after my understanding from something Arnie said today)... but it still doesn't make sense that it matches on power only so quickly.

Make it wait at least half a minute before giving two players a skewed fight ( > 200 difference) so it's possible to avoid it. [by the way I thought it counts 200 ELO as 3 power, not 6? ;o]

I'm guessing the possibility to match unevenly on ELO has to be there to avoid lack of games in top tiers, and to avoid having an elitist group it takes long to break into (because you could only beat people with equal ELO for relatively slow gains ~ might not mean much right now, but could do so later)... but yeah, it shouldn't force these uneven matches so quickly.. ;o

Finjinimo
03-04-2013, 09:05 PM
. My Elo rating sits at 1304. I hire a couple of rank 0 units and create a team with five basic and one promoted units. I join the MM queue and get an opponent within less than 10 seconds. He has a Thrasher named Thrasher and five basic units, with a total kill count of 17 between all of them. While I can't see his rating the game is a one-sided slaugther and he makes a lot of obvious rookie mistakes. No way is his rating anywhere near mine and certainly not within 33 points. I proceed to win and gain four points of Elo where I should have gained nothing. Why does he have to face me on even terms? If this stays as it is there will be people farming fresh players and cause massive frustration among them.

Ya, that's the kind of farming I was talking about penalising. So if you are higher ELO by a certain threshold you should get less renown for that match. The same way you get less ELO.

Renown means your level of fame in the town yes? Why should a well-known and established player get the same amount of renown for thrashing a new player, as they would from beating someone more established.

Yes the matchmaker pairs us up with our opp so I don't actually have a say in who I battle... but I do. In order to farm wins, winstreaks, and renown... all I need do is load up an un-promoted team and I will be almost guaranteed to be playing against inexperienced players.

So, by decreasing the ELO and renown I earn for such matchups it would make me less inclined to game the system.

The ELO penalty is already in place (I think), so why not for renown too? ELO is worth nothing really, Renown is worth everything.

erom
03-04-2013, 09:13 PM
If you consider the history of matchmaking over the course of the game, it's getting a bit hilarious.

During soft launch: "It takes too long to get a game. Make it find matches easier!"
Immediately after launch: "I keep getting matched against teams too far from my power level! Make it take power level into account over Elo!"
Now: "I keep getting matched against teams not at my skill level at low power numbers! Make it take Elo into account over power level!"

If I was the devs, I would be sick of working on this particular problem, but matchmaker tuning is very important to get right in a game like this.

Zahar
03-05-2013, 02:56 AM
I think rank 1 units should NOT be allowed to deploy with 10 points to diminish team cost.

Jorgensager
03-05-2013, 05:29 AM
I think rank 1 units should NOT be allowed to deploy with 10 points to diminish team cost.

What are you talking about? They do not diminish team cost at 10/11. A rank 1 unit is worth 1 power regardless of stats.

raven2134
03-05-2013, 07:51 AM
Great discussion guys. Stoic is following this, and is thinking about how to improve matchmaking. :)

tnankie
03-05-2013, 04:55 PM
Matchmaking looks first into Team Power (we need to start using correct terminology) which is the sum of all the ranks in your team. THEN we try to match you with the closest Elo opponent.
So in a perfect scenario we will find you a match with an equal Power team AND Elo rank.
Ha!

oh look it is a flying pig.

Arnie, my point is that the algorithm may work well under ideal conditions, the conditions we face are not ideal. The pragmatist in me says that now matter how beautiful the algorithm/solution is if it doesn't fit the current problem then it isn't a good one.

The edge cases need a more robust treatment.

Elo rating for both players needs to be shown at the start.
Underdog needs to also come from Elo disparity.
Underdog needs to be stated up front (as in letting you know explicitly how much pain you are likely in for and why it is worth sticking around to get your pants pulled down and bottom smacked.)

brianrcampbell
03-06-2013, 03:25 AM
Edit: oh my god this is a wall of text. I'm rambling about the nuts and bolts of the Elo system. Don't feel obligated to read it if that's not interesting to you.

In practice, I've been really happy with the matchmaking system, though I've been tacking on rank 1 dudes periodically, so I haven't really experienced the "experienced player at team power 0" situation that spawned this discussion. I love that I can get a match in under 10 seconds. I would be annoyed if it took longer than that (see, e.g., XCOM... for this and about a billion other examples of really badly implemented multiplayer).

Initially I didn't see this as a problem in the abstract, but now that I think about it more, there might be something there. I don't know how this game's Elo system is tuned, but I'm going to assume for the sake of conversation that a 400 point differential indicates that, if they played 11 games, the higher ranked player would be expected to win 10 games and lose 1, and that the maximum number of points that can be gained or lost in a single game (k factor) is 32. So if you're a 1400 player, set your team power to 1 so that you play only brand new 1000 players, each win would net you about 3 points. But each loss would set you back 29 points. That's a fair bit of risk you're taking there.

But now I am wondering whether the "transient" nature of a system where there are lots of incoming new players whose "true" skill level shifts dramatically with time (e.g. a player after 10 games is significantly better than the same player before their first game) is adequately accounted for - or even how you would go about accounting for this. I feel like new players could be consistently overrated by the Elo system. I can't pin down exactly why this would be, though. But if it IS a problem, it could be at least abated by reducing the K-factor for an experienced player who wins against a new player. Basically, IF [loser has played less than 10 (or whatever) games], AND [winner Elo - loser Elo > 200 (or whatever)], THEN reduce winner's Elo gain by half (or whatever). This could help to address the problem (if new players are, in fact, overrated) of experienced players poaching newbies for Elo gain by shifting the risk-reward calculus and simply making it less time efficient.

I would think that they have the data available to determine if this really is a problem. Whatever factors they use, the Elo system assumes a particular distribution of skill which it implicitly equates to likelihood to win a head-to-head match, and uses this distribution to award or penalize points. They should be able to pull a big set of matches between new players (e.g. 5 or fewer games played) and very good players (e.g. >300 Elo points disparity), maybe removing disconnects if we don't believe that people have been disconnecting intentionally / strategically, and come up with an expected win-loss record, and compare that against the actual win-loss record. If, as I suspect (but can't figure out why), the expected win-loss ratio is lower than the actual win-loss ratio, this would be an indication that new players are overrated by the Elo system. Also, if they can pull a set of high-Elo players (>1400 maybe) who have lots (>50 maybe) of games played against new players, they could see how this set of players fares against competition with similar Elo scores (+/- 50 maybe). My guess (if new players are overrated) is that these players would lose more often than predicted by their Elo score.

Basically all of this system design should be done so that players don't have an incentive to "game the system" by seeking out or avoiding particular types of matchups. You don't want to have experienced players poaching newbies for disproportionate Elo gain, but at the same time you don't want to have experienced players unable to run a power level 1 build for fear that they'll run into newbies and have a disproportionate Elo risk (or maybe you do -- getting roflstomped by an experienced player is not a good introduction for a new player). Striking the right balance may be tough, especially if, basically, a newbie actually isn't an 1000.

I'm really interested in how they go about designing this. There are a lot of different factors in play, both in terms of what kinds of behavior they want to encourage and the actual nuts and bolts of determining the true skill of all players.

eduran
03-06-2013, 04:09 AM
So if you're a 1400 player, set your team power to 1 so that you play only brand new 1000 players, each win would net you about 3 points. But each loss would set you back 29 points. That's a fair bit of risk you're taking there.


I'm gonna go out on limb here and say new players don't care about Elo. So while this scenario is fair (assuming the system is tuned correctly) if you look only at the rating aspect, it still leaves the new player frustrated after a stomp. Most likely he won't even notice that he lost almost no Elo. And on the flip side, I doubt that people abusing the system do so for the ranking. They probably just hate losing and by playing new players constantly they can go on ridiculous win streaks.

Morgenstern72
03-06-2013, 12:22 PM
I doubt that people abusing the system do so for the ranking. They probably just hate losing and by playing new players constantly they can go on ridiculous win streaks.

I think this is actually happening and it would be great if developers could take a look in that.

raven2134
03-06-2013, 11:17 PM
Yep, Stoic knows and is looking at stuff to address this. No details yet.

Morgenstern72
03-12-2013, 01:30 PM
Ok, now it's even worse :(
Yesterday I had these matches
me = rank 1 (training a shieldbanger), ELO 1350

Fight 1: Rank 4, elo 1470 -> no chance at all
Fight 2: Rank 6, elo unknown -> no chance at all

Did give up then.

Today: Rank 6, elo unknown -> no chance at all

netnazgul
03-12-2013, 02:46 PM
Today: me pow6 ELO 1400, my opponent pow8 ELO 1550 :D

Bloodaddict
03-12-2013, 03:55 PM
Dont wanna say that the MM is bad just wanna share some experiments since unit upgrade were introduced: Seems that it is hard to find a 6vs6 match at the moment. Today I found none. First I played with my 6 against 10, than 7. Then is exited... I lost both and while I dont't wanna blame the MM for the second loss, I think that I had not a fair chance with the first one even when playing perfect...
As I said, I don't wanna blame the MM, I think at the moment there are just to many possibilities from level 0 to level 18 teams to find a lot of fair matches.

Morgenstern72
03-12-2013, 04:16 PM
I think there really has to be a handicap system OR you can predefine 3 (ore more) teams with diffeent ranks and matchmaking is trying to find a match for one of your teams.

At the beginning I really had fund playing this game but now it's not giving you (or your opponent ) a real chance to win :(

Vexbane
03-13-2013, 01:56 AM
I am waiting for the next patch before I play again due to the poor matchmaking. I hope something is done soon. That and the balancing of the 4x builds.

raven2134
03-13-2013, 07:26 AM
Alex is losing sleep over this, Arnie is deeply concerned and John is...being John and staring at the code :). They're on it and figuring out what to do, hang tight guys.

franknarf
03-13-2013, 07:59 AM
So, it sounds like...

Seriously though, we're adding a "ranked" toggle. It makes sense that it should only make perfect matches.
I was one of the folks that brought this up, but I think it will lead to another problem. People running ridiculous team-power levels (like, say, 10) and looking for ranked matches may not realize how hard it is to match them. I can think of two ways to address this, one of which is much easier:

Provide visual feedback, e.g., "There hasn't been a ranked match with anyone at that rank for x minutes."
Limit the team power levels at which ranked play can occur (as suggested (http://stoicstudio.com/forum/showthread.php?1251-A-new-ladder-system) in a couple forms, e.g., {6, 9, 12, 15, 18}, or {6, 12, 18}).

Besides being harder to get right, the first method invites Elo/renown/whatever farming, because you can make a ranked match with your shill much more easily when all the levels are available. You could always start with only the three full-rank levels {6, 12, 18} or the lowest three levels {6, 9, 12} and then expand it as the player base grows.

Then again, it could be that "Seriously though" at 3 am indicates exhausted internet sarcasm. :D

Vexbane
03-13-2013, 01:49 PM
After further thought imo there should be no underdog matches as far as rank power goes. All matches should be matched up against the same power as them. The difference is the longer the wait the higher skilled opponent you will face (elo and total game wise) which could then be the underdog bonus.

If you are having trouble getting a match switch around your rank. 6 is the standard and will usually always net you a decent match. Some kind of clue when you go to ranked lobby would be nice .It could say (low rank pop) so you would know there would be a longer wait.

There could be a list of the amount of rank bgs in "que" So when you click to get a rank match (but before you click find match) you would see a list of the number of rank bgs in that lobby (no names of players). It would update in real time. So if you see there are no rank 8 bgs in que you could see what was and play that. Yes that might "limit" what rank bg you can play based on the time you are on, but I feel that is the easiest way to balance it.

Bloodaddict
03-14-2013, 03:22 AM
If you are having trouble getting a match switch around your rank. 6 is the standard and will usually always net you a decent match. Some kind of clue when you go to ranked lobby would be nice .It could say (low rank pop) so you would know there would be a longer wait.


Would be nice to know, if that is true. I am playing with team power 6 all the time and in the last two days I did not find one match where I played 6vs6, I played mostly 6vs7, once 6vs10 and once 6vs5. I was thinking already about promoting one or two units to level 2 just for a better chance to find fair matches. But since I don't know at which level there is the best chance to find fair matches I left it as it is...
A statistic about the team powers used would be nice!