Perspectives War Robots WR Data and Analysis

War Robots Match-Making: What in the Holy Mother of Kang Dae is Going on?

Minski recorded data from over 100 matches to analyze the matchmaker. Here's what he found!

MECH*SPECTRUM is delighted to welcome this piece by Minski, and look forward to more great insight into War Robots!

rage-gif

“I came about 2 seconds from uninstalling this effing game last night – it’s getting ridiculous!”

…posted one of my long-suffering clan mates. Now, if a tactically astute, espresso-supping, Champion-league player is that unbearably frustrated by the game, then what feasible hope can the rest of us mere mortals harbour?

I started talking to pilots across different league levels, clans, and geographies, trying to nail down what loony rules the match-maker (MM) loosely followed. Urban myths included:

  • I heard that MM is based on the lowest league rank in the squad…
  • There is no MM in WR. Level 10-13-15 players are MM with level 25-28-30 players…
  • Pix wrote in code to inhibit constant success, to keep you down, and level the field a bit…
  • Activity cups also affect MM…
  • Change your weapon setups and bots and you will be matched with similar players…
  • If a player in Diamond league has weapons at level 10 and above it brings in Masters and/or another player with high level weapons in Expert

So, I decided to run my own test to see how perceptions compared with practice. My aim was to see how blue and red team league levels contrasted, how this was affected by squad or solo play, and also how ‘fair’ I thought the matches were. Granted, when measuring ‘fairness’, this blunt comparison doesn’t account for factors such as hanger strength or tankers (players consciously losing games to drop leagues and face easier competition), but hey, it’s still an indication.

The Test

100 battles were played on iOS, with none relating to the Leavers Queue. Using information available from the end-game summary screen, three things were noted: the league level of each pilot on both sides, if they were in a sub-squad, and whether the battle ‘seemed’ a ‘fair’ match-up.

Each league level between Bronze 1 and Champion was assigned a score, from 2 up to 20. The ‘league score’ total was added up for the blue team and also the red team. The difference between the two scores indicated the gap in individual pilot league levels between teams.

Health warning: whilst these tests are not statistically divine, the aim was to provide a ROUGH INDICATION as to ‘what in the holy mother of Kang Dae is going on’ with match maker. So I’d urge the reader to accept the findings in the spirit in which they were intended.

Blue versus red ‘league levels’

In 63% of games, the difference between the blue and red team ‘league score’ totals was +/- 6. Six league levels across 6 pilots, is on average, one league level between teams. So, in nearly two thirds of games, on average, that is like a full Diamond 2 blue team playing a full Diamond 1, or full Diamond 3 red team. Pretty good match-making in most games then.

In 11% of games, blue team were, on average, 1-3 leagues higher ranked than the reds. This is like your full Diamond 2 blue team, on average, playing red teams ranked Diamond 3 down to Gold 2. A bit of a gap but not too crazy-crazy.

In 26% of games, blue team were 1.5-9 leagues per pilot lower ranked than red team. This is like a full Gold 1 blue team, playing red teams, on average, anywhere from Diamond 2 up to Champion league. Granted, in reality, blue squads probably had, or were assigned, an individual who was far higher than the rest of the team, which pulled in higher league opposition. Nevertheless, in a quarter of games, on average, you probably have next to no hope of winning the battle.

The Squadding Effect

63 of these 100 battles involved a blue squad.

In 56% of blue squad battles, the red team had a higher ‘league score’. Slightly unbalanced, which still seems acceptable. But how large was the imbalance, and how did this alter by blue team squad size?

The larger the blue team squad, the wider the ‘league score’ average difference between teams (in red’s favour). Solo (1%) and two pilot blue squads (1.3%) were fairly evenly matched with red. However, red superiority increased the larger the blue squad became: 3 (11%), 4 (14%), 5 (17%). The exception to the trend was 6-pilot squads (7.5%), perhaps because the match-maker couldn’t insert a far higher league pilot who would attract higher league opposition. It is of course true that squads can operate as a more cohesive team by using common tactics, voice communication, and reducing the number of random players with different agendas.

This tells us is that you have more chance of playing a well matched team if you solo, or play in either a 2 or 6-pilot squad. The more competitive teams out there may want to pay more attention to the benefit that each additional squad-mate brings against the extra heat that larger squads attract.

‘Fairly’ matched battles

Each battle was subjectively marked as a ‘fair’ or ‘unfair’ battle, by how easy or challenging it felt to damage or be damaged by the red team.

66% of battles were marked as a ‘fair’ fight. This happened to be where the difference between teams was up to +/-10 points or, on average, 1.6 league levels per player. This is like, on average, a Diamond 2 blue team playing anything down to a very weak Diamond 3, or up to a very strong Diamond 1 team. This tells me that I’m actually more comfortable with the level of the opposition than I often feel. Of course, the other 33%, or third, of battles are a different matter entirely!

Minski

Myth busting

Myth: MM linked to bot / weapon levels / activity cups / pilot level who starts the game.

Pixonic: MM is only based on the highest pilot league level in the team.

Test: Pixonic’s response aligns with the results from the ‘squadding effect’ test which showed that unless the squad was all the same level, the highest level squad member would pull in higher level opposition. This could be you, one of your squad, or a random player assigned to your blue team. The bigger your squad, potentially, the more likely it is to be imbalanced, and the tougher the opposition.

In summary

Two thirds of battles are well matched, in terms of team league levels.

In a quarter of battles you have next to no hope of winning.

When squadding, red has a slightly greater chance of being a higher average league level. Best match-making occurs when solo or in 2 or 6-pilot squads, but not in 3-5 pilot squads.

Two thirds of battles seemed ‘fairly’ matched. This mirrors the squad-level test above.

Try to squad with players your exact league level for best match-making. This may be unlikely when just choosing from just your own clan, so reach out on Facebook, Discord and Line communities to expand your Facebook friend network, and in-game pool of pilots to squad with.

And finally, to my friend who destroyed his iPad by throwing it at a wall in the most extreme frustration – I hope this saves you money on your insurance bills.

About minski666

Minski is an avid Mech gamer based in London, England, who loves to delve under the hood of how games and tactics work. Myth busting is a favourite, and suggestions for topics to look into are always welcome! Discord: Minski#5577 Facebook: Min Ski

17 comments on “War Robots Match-Making: What in the Holy Mother of Kang Dae is Going on?

  1. Rad Mobile

    Nice work Minski! Appreciate the time and effort of keeping the stats and compiling it into a well written, short, and informative piece!

    Like

  2. Ashe Sterling

    iPad on the wall? Umm… I guess that a tablet mark is a peculiar decoration that might be appreciated by abstract artists… Anyways, I can confirm that the MM can get very frustrating sometimes, unless that you are in a Champion league with a full hangar of our dear sir Lancelot..

    Like

    • Very well put Ashe! Unfortunately, the match maker can be just as frustrating to the champ league crew as many of my clan mates will tell you. Imagine going for a solo run and being put up against a red team champ crew on comms? Or how about being paired up with lower level team mates or players who camp. All very tricky to navigate. It is interesting that you mention Lancelots in the champ league – I lift the lid on that in my next written piece!

      Like

  3. I several times wrote Pixonic about their unbalanced MM system. (My platform Gameroom) First and Last 2 seasons I ended in Legend League but I mostly lose than win because of bad MM. In most battles, my teammates level from 13 to 25 and opponents team has at least 3 Champion League players. Even I make more than 1 ml Damage in a battle and get at least 2 beacons, we are still loosing because teammates make from 0 -170 k damage it seems they don’t even know what beacons are for.

    Like

    • HI James – yes this is a common complaint. It is worth highlighting that Gameroom has a smaller pool of players than IOS or Android, so the matchmakers has fewer options to pair you up with. Not that this is a defence, but more of a suggested reason. Let’s hope that future solutions improve matters!

      Like

  4. Just make it fun. If you feel out powered just leave the battle and start another one. 😁

    Like

    • Doesn’t that land you in the LQ eventually though?

      Like

      • I mean fight with every bot you got and after all your bot destroyed. Hit the leave battle button and start a new one
        I never leave battle while i still got bots in hangar.

        Like

    • HI Ara – it sounds like you have a positive fighting spirit there! Keep it up and best of luck in your battles today.

      Like

  5. I realized over 2 years ago that the ONLY way to ensure a fight that wasn’t unfairly stacked against you is to max all bots and all guns as soon as possible.

    Like

    • Hi Richard, this certainly helps put you on an equal footing in terms of weapon/bot levels. However, the latest furore around bots that not everyone can obtain presents yet another barrier to fair fights. I imagine that the Dash overpower scenario will be rectified in the coming months however that just leaves room for the next batch of OP bots and weapons. From Pixonics perspective this makes good revenue, although from ours it obviously hollows out the soul of the game which is less than ideal.

      Like

  6. ҠƛMƖҠƛƵЄ

    Thanks for your time to record and document your findings very much appreciated

    Like

  7. Thanks Kamikaze – feedback is always very much appreciated. I’m pleased you read through until the end.

    Like

  8. I cannot agree with Minski on his conclusion that 2/3 of the fights are fair. I am expert level 3. I like to play squad with my buddies. usually we will be 3 or 4 people in squad. maybe 2-3 expert and one master. more than 80% of the times we will be paired with other players full squads where all enemies will be champions usually with level 12 all weapons and bots. the game just does not know what fair means. we have decided that in the future we will leave such battles.

    what is even more frustrating is that many times you will play a hugely undermatched game, you will lose by bots getting destroyed all of blue side while red side still has 6 bots (this is hardly a sign of a fair fight). and then clicking on a new battle the MM will consider appropriate to pit you again with the same squad that just kicked your ass, not giving you a chance at least to have a good fight for your next game.

    in case of squads the MM will just pit you with what is available in the server and it really seems the game has become quite swallow because many times you will be pitted against the same opponents being severely outgunned.

    for me the problem is not the dash bots or inquisitor, if pixonic wants to make them status symbols charge a lot of money and only allow few people to have them this is fine by me. they have obviously put a lot of work in them and they have to recover the R&D. for me the big problem is that the matchmaking is very very poor and will pit people or squads of people without dash against squads of people having 2-3 dash bots each.

    people having OP bots should fight each other, this would be a test of skill, not bit on other players who are ill-equipped and dont want or are unable to afford the dash bots.

    the matchmaking should be working on hangar power. each player should be assigned a hanger power level based on bots and weapons level they have. overall hangar power of both teams should be as close as possible.

    many times you see that the MM is not really that smart and even switching players from the chosen two teams would have made for a much fairer fight rather than putting high level players on one side and weaker players on the other side.

    so what is ruining the game and the experience is not that it is becoming P2P but that people are not matched accordingly to the level and are not getting fights they can win becoming frustrating.

    for me the leagues do not mean anything, there are silver players with 6000 wins obviously having full 12 hangars. should they be playing with other single players even at solo.

    if you want to make the MM really smart then it should take into consideration hangar power, win percentage and average damage – create a player score and then match players on this score.

    it is not possible when I am low expert with under 50% win percentage and 450K avg damage to be pitted against squads of people having 800K avg damage and over 90% win ration. when they are winning over 90% of their battles it means the battles they are playing are not evenly matched.

    I am sure PIXONIC could be doing much more on the MM but my impression is they just dont care, creating inequality in the MM means people are driven to spend to become more competitive and on the top of the food chain.

    really getting tired of this game, it also has lost its sense of accomplishment. so I am collecting gold with the daily tasks and occasionally gold from battles. but for what reason, to get another ancilot which can be destroyed in seconds by an orkan haichi? its not possible to buy with gold the good bots anymore so what is the point?

    I am not optimistic that things will get better despite recent claims by the company that its listening to the players comments

    Like

    • Hi dimkin, firstly can I thank you very much for taking the time to feed back. I know it took a while to write, and I think you have many valid points. I’d like to respond to some if I may.

      Of the 100 battles that I recorded, across all time zones, times of day, league levels, squad sizes/solo, and maps, 63% had a team league level difference of one league. So, for example, your average Expert3 team, on average, would be fighting an Expert 2, or Diamond 1. For this statistically material number of battles, I’m comfortable describing them as ‘fair fights’.

      You mentioned that you play in squads of between 3 and 4, which according to the results of the test will attract red teams 11-15 % stronger. If you’re playing in Expert squads, this will straight away attract master level players and probably champions, which matches with your story.

      I recognise the frustration you feel, and the urge to leave an unfair fight – I have felt this myself many times. I would caution you though, as this risks Pixonic placing you in the Leavers Queue, which is yet more uneven in play.

      Whilst your point about blue team being destroyed is probably very often true, the article didn’t cover win/lose outcomes, so I can’t comment on this point I’m afraid.

      I very much agree with you in that the game algorithm can only pit you against who is on the server at the same time. The higher league you are in, the fewer players, I’m told, this allows you to be matched with. One trick I learned, in case you don’t do this already, was that if you’re meeting the same opponents repetitively, perhaps wait a couple of minutes to let them go into battle, before you start a new game yourself.

      I thoroughly agree with you that the matchmaker is extremely poor. I have read many players suggesting that it is actually in Pixonic’s interest as this may drive additional revenue from players wanting to correct this imbalance with paid-for bots like Dash.

      Just prior to writing the article, I did see a communication with Nikita from Pixonic who suggested that they would not be going back to using hanger strength for matchmaking. Whilst this sounds like a perfectly sensible approach to use, I’m unclear for their reasoning. I also agree with you that leagues don’t mean much these days, as tankers and seal clubbers manipulate the system. For example, the top clan (by activity) in the game boasts an army of champion league level players, who choose to play in the silver leagues. One that I remember playing, has max damage of 1.9 million, longest win run of 103 games, and 12,000 victories (!)

      Let’s see what the developer does next. Chin up mate, I’m sure that you’re a solid player, perhaps consider playing in 0-2 or 6 player squads to give you a statistically better chance of winning? Just a thought.

      Thanks.
      Minski.

      Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: