88generalblue
GI. Patton
Posts: 137
Karma: 76
Platform: Android
Clan: Wik2
League: Silver
Favorite robot: Gareth w/ Taran, Aphids
|
Post by 88generalblue on Feb 8, 2017 1:06:53 GMT -5
*Disclaimer* What I am putting here is just personal conjecture as this is not a supported theory by the Wiki group. Also, I do not take credit for the pictures herein, and I was told by the person I got these from that they are not from Pixonic, they are fan-made only. */Disclaimer* Anywho, I have come to the conclusion that Avg Dmg (AD) is the greatest indicator of one's ELO in the new MM. The MM is rounding your AD to the nearest 50 or 100, rounding up if your Win Rate % (W%) is above 50% and rounding down if your W% is lower than 50%. For Example: Avg Dmg - Win % - Tier 325K - 55% - pushes up to 350 AD- Tier 375K - 45% - drops down to 350 AD- Tier The reason I've started thinking AD is the biggest indicator: Pixonic raised the health of most bots by 6% in the same update as the new MM, as having more health per level would separate each tier more proportionately, As it would be impossible to measure every individual player's 'skill' by nearly any other metric. We can't measure ability in dodging Aphids (vs just missing), or how well they use their Griffins jump ability, they are likely using some other standard that we can already see, As players are matched with similar-strength hangars & skill-sets, everyone in that bracket would begin having much higher AD. Better & stronger players, facing players with more HP, will be capable of raising their AD as a key indicator to how well someone performs, The people winning most (higher skill) will start moving up to the next AD- Tier as their W% rises over 50%, and those that lose go to back to the slightly lower AD- Tier, It also explains how the MM doesn't 'force' a certain War Robots %, it 'pushes' towards it. I currently have a 62% W%, if the ELO was based solely on W% it would be as rare to see hangars with a 30% W% as 70%, Let me know what you guys think of this idea, and check out the sweet picture below:
|
|
Deleted
Deleted Member
Posts: 0
Karma:
|
Post by Deleted on Feb 8, 2017 2:31:25 GMT -5
Interesting theory. Someone could test it maybe by switching from high damage rigs to beacon capturing setups. If they could keep their victory % about the same but sink their AD. MM might be on a rolling equation. The Elo emphasis switches day to day or once a week to prevent gaming the system.
|
|
|
Post by BlackGrouse on Feb 8, 2017 3:50:04 GMT -5
As it would be impossible to measure every individual player's 'skill' by nearly any other metric. We can't measure ability in dodging Aphids (vs just missing), or how well they use their Griffins jump ability, they are likely using some other standard that we can already see, I personally like this theory. As in my progress log here shown, my current average damage is 441 003 with 72% win rate pushes me to 450k average damage (Platinum rank) according to that Rank -table. In the time of old MM my hangar was somewhere in "trash-gold" or above tier and now since the new MM my average damage output is climed up slowly to current one. That's why I'm thinking your Rank -table (and theory) is in place. But I haven't thinked/theoried new MM at all, so this is only feeling according to my stats !
|
|
|
Post by whatttupG on Feb 8, 2017 14:29:15 GMT -5
I see a problem in your theory.
"As players are matched with similar-strength hangars & skill-sets, everyone in that bracket would begin having much higher AD."
This assumes the matches are even and scores would go up. I disagree, once a team faces a very similar team in skill, I'd think they start canceling each other out in skills and that would lead to decreased scores, not increased.
This I have seen, just bring aphids to dead city and single out the same guy across the bowl. Eventually you should find what I saw. Recently on that map, I took my spot and other aphid guy took his, and we traded volley's load after load trying to kill each other.
Quickly it became a Mexican standoff and clear we both knew how to throw and hide. It took over three minutes to catch him exposed and finish him. OK, maybe I had more patience I don't know, but he was a fun fight, it was a brutally even match, and it took forever to root him out. As a result my score was about half of normal so while this was a waste of time and silver in some sense, it was the type of rocket party I love to attend. My point is my score went way down, I know his did as well, it makes sense too because we both wasted so much time on each other.
|
|
Deleted
Deleted Member
Posts: 0
Karma:
|
Post by Deleted on Feb 8, 2017 16:30:19 GMT -5
It would explain why those that run dedicated beacon cappers have less volatility in their matches, and why I saw an 8/10 spiral patton the other day when I run a capper and 2 melee bots.
|
|
|
Post by truechill on Feb 9, 2017 4:08:35 GMT -5
Does this mean there is a zero damage tier out there filled entirely with Zenit/Noricum Natashas? ...
|
|
|
Post by Muhlakai on Feb 9, 2017 10:58:56 GMT -5
|
|
|
Post by loren on Feb 9, 2017 14:40:46 GMT -5
*Disclaimer* What I am putting here is just personal conjecture as this is not a supported theory by the Wiki group. Also, I do not take credit for the pictures herein, and I was told by the person I got these from that they are not from Pixonic, they are fan-made only. */Disclaimer* Anywho, I have come to the conclusion that Avg Dmg (AD) is the greatest indicator of one's ELO in the new MM. The MM is rounding your AD to the nearest 50 or 100, rounding up if your Win Rate % (W%) is above 50% and rounding down if your W% is lower than 50%. For Example: Avg Dmg - Win % - Tier 325K - 55% - pushes up to 350 AD- Tier 375K - 45% - drops down to 350 AD- Tier The reason I've started thinking AD is the biggest indicator: Pixonic raised the health of most bots by 6% in the same update as the new MM, as having more health per level would separate each tier more proportionately, As it would be impossible to measure every individual player's 'skill' by nearly any other metric. We can't measure ability in dodging Aphids (vs just missing), or how well they use their Griffins jump ability, they are likely using some other standard that we can already see, As players are matched with similar-strength hangars & skill-sets, everyone in that bracket would begin having much higher AD. Better & stronger players, facing players with more HP, will be capable of raising their AD as a key indicator to how well someone performs, The people winning most (higher skill) will start moving up to the next AD- Tier as their W% rises over 50%, and those that lose go to back to the slightly lower AD- Tier, It also explains how the MM doesn't 'force' a certain War Robots %, it 'pushes' towards it. I currently have a 62% W%, if the ELO was based solely on W% it would be as rare to see hangars with a 30% W% as 70%, Let me know what you guys think of this idea, and check out the sweet picture below: Great post 88generalblue, I would like to add that somewhere in the calculation they must be taking beacon capture in consideration as well. I have had to drastically change my game play and go from damage dealer to beacon runner. I grab as many beacons as I can with my two smalls then drop to my heavies and blitz the middle to hold. I do not get a lot of damage, mostly half eat'n heavily smoking hulls by the time they reach middle which I chew up with my Leo Pinata thunder. But even with a low damage output I still get bumped up with a low damage score but higher beacon count. Loren
|
|
88generalblue
GI. Patton
Posts: 137
Karma: 76
Platform: Android
Clan: Wik2
League: Silver
Favorite robot: Gareth w/ Taran, Aphids
|
Post by 88generalblue on Feb 9, 2017 15:04:50 GMT -5
Thanks for posting this, it was very helpful to see that AD is confirmed by Pixonic as one of the most important factors in MM. Now if only they could find a reasonable way to prevent players from purposely tanking, we will be all set!
|
|
88generalblue
GI. Patton
Posts: 137
Karma: 76
Platform: Android
Clan: Wik2
League: Silver
Favorite robot: Gareth w/ Taran, Aphids
|
Post by 88generalblue on Feb 9, 2017 20:11:38 GMT -5
I see a problem in your theory. "As players are matched with similar-strength hangars & skill-sets, everyone in that bracket would begin having much higher AD." This assumes the matches are even and scores would go up. I disagree, once a team faces a very similar team in skill, I'd think they start canceling each other out in skills and that would lead to decreased scores, not increased. This I have seen, just bring aphids to dead city and single out the same guy across the bowl. Eventually you should find what I saw. Recently on that map, I took my spot and other aphid guy took his, and we traded volley's load after load trying to kill each other. Quickly it became a Mexican standoff and clear we both knew how to throw and hide. It took over three minutes to catch him exposed and finish him. OK, maybe I had more patience I don't know, but he was a fun fight, it was a brutally even match, and it took forever to root him out. As a result my score was about half of normal so while this was a waste of time and silver in some sense, it was the type of rocket party I love to attend. My point is my score went way down, I know his did as well, it makes sense too because we both wasted so much time on each other. I see what you are saying, but thats exactly why it works. if you are stronger (meaning having higher level bots/weapons/skill) than others, you will get more damage, but you will naturally level- / cancel-out each other's effectiveness, and your AD will stabilize because you've reached a certain limit. now, there's always times where you are going to get more or get less damage based on many factors, but when you take the Average from 50 battles the outliers will even out. if you are finding battles where you negate each other's effectiveness, then it seems the MM is doing its job by creating a challenging experience for all players. Even those Zenit Tashas are gleefully bombing each other as you read this! Final note, I liked your use of "Mexican standoff" and "Rocket party" Cheers,
|
|
|
Post by amoebastudios on Feb 14, 2017 17:55:41 GMT -5
My experience matches what you have shown, that MM is based primarily on average damage. I've played long enough in the new MM to see this. Running 10/10 bots with 30k average damage and you'll see cossacks. Running 1/1 bots with 400k damage and you'll see some top-tier bots.
The MM breakdown feels like this: 70% Avg Damage 15% Avg Hangar strength 15% Win Rate
From what I've seen: <100k Avg damage: Noob/Low bronze. Mostly cossacks/destriders. Leave them alone 100k-175k: Bronze 175k-275k: Silver 275k-400k: Gold 400k+: Top Tier
I personally feel average-dam is indeed the best way to put bots of similar weight class and competitiveness together. Then move people slightly up and down depending on other factors like hangar and win rate. Note that I said "average" hangar score because it doesn't react to hangar strength right away.
|
|