Page 1 of 4 123 ... LastLast
Results 1 to 10 of 39

  Click here to go to the first staff post in this thread.   Thread: Defining a Classic war mismatch

  1. #1
    Forum Elder Thegreatpuma's Avatar
    Join Date
    Mar 2016
    Location
    Druidia
    Posts
    2,516

    Defining a Classic war mismatch

    I've been trying to think up objective ways to determine a mismatch in classic wars, so I have decided to start tracking Gold Weight, Hero Levels and TH Levels for our matchups.

    The current war we're working through in TMBG right now I would define as a mismatch, not a major one more of a minor one. I base this on:

    1. Total Gold Weight: us - 905, them - 907
    -----Mostly even here, very very slight advantage to them
    2. TH levels: us - 2/1/3/0/2/1/1 them - 2/2/2/0/2/1/1
    -----Again mostly even, but a small advantage to them with one more strong 12 and one less 11.
    3. Total hero levels: us - 389/392/181/40, them 407/421/205/33.
    -----They are 4.6% stronger at King, 7.4% stronger at queen, 13.3% stronger at GW, and 7% weaker at RC, making for a fairly solid advantage

    They also have a sandbagged th11 missing infernos and EA.

    Generally on matchups that seem to be mismatched at first I can see gold weight trade offs against offensive weight and heroes. I'm not seeing that on the above. I did notice that their second TH13 has mostly ignored Siege Machine upgrades as have their 12s (though all have all machines available). Also, their 11s and 12s have mostly ignored ice golem upgrades. This seem insignificant though.

    I guess the real question I have is what am I missing? Should I be tracking levels of ALL troop upgrades for each account too? Anything else I'm not thinking about? To add troop levels would make the work involved much greater unless someone knows an easy way to get all troops levels for an account.
    TMBG War Clan always recruiting - Wars with TH9-13
    TMBG Discord server:
    https://discord.gg/byNR2xZ
    TH11 - Doctor Worm 50/50/20 heroes/near max d
    TH12 - doug 65/65/40 heroes

  2.   This is the last staff post in this thread.   #2
    I imagine the matching uses all troop levels, so you do need to include them, although that isn't actually a very good measure, since most players have 1-3 go to armies, and can upgrade just the troops for those, while leaving others underpowered.

    And I suspect that doing that is what lead to the majority of current mismatches. I have no real evidence for that suspicion, it just seems hard to write an algorithm that will give an accurate measure of offensive strength when only a few troops are upgraded.

  3. #3
    Forum Elder Thegreatpuma's Avatar
    Join Date
    Mar 2016
    Location
    Druidia
    Posts
    2,516
    Yeah, I guess I will see if there's an easy way to get troop levels from Player Tag. I know it's in the API and on websites but I need to be able to get it into a spreadsheet to total each troop level to be able to do a real comparison.

    I mean even if there are only specific troop upgrades I should be able to identify where the offsetting weight is coming from.
    Last edited by Thegreatpuma; November 16th, 2020 at 09:17 PM.
    TMBG War Clan always recruiting - Wars with TH9-13
    TMBG Discord server:
    https://discord.gg/byNR2xZ
    TH11 - Doctor Worm 50/50/20 heroes/near max d
    TH12 - doug 65/65/40 heroes

  4. #4
    Forum Contender joshsgrandad's Avatar
    Join Date
    Jul 2017
    Location
    NE England, UK.
    Posts
    4,630
    I think the MMA did a good job..
    If it were a mismatch, then that missing 10 leading into 3 lower halls has maybe made the MMA try to follow your breakdown pattern somewhat..

    They have a slight royal advantage per base, but you are compensated with RC levels..
    They have a 6 level warden advantage per base, which would be my main concern if it were us..
    Gold weight, as you said, looks fine, so defensively they are quite similar..
    That slight hall advantage could probably be written off as maxer syndrome.. If you are gonna complete your hall levels, you are gonna feel some pain if facing those that dont worry about when they hit the jump.

    Matching is only going to be approximates with only 10k differential between halls at the top 2 levels, so big 12s should always be considered a risk with the war draw..
    We do the same, but it is despite my warnings, not because of my advice..
    And as touched on earlier, once the roster starts to spread, that invites the MMA to start looking for concessions in order to give a decent draw whilst utilising a similar breakdown.

    TBF, I dont bother with the maths, I simply look at the draw result and think Yeah, that looks okay, or Hmm, wonder what happened.
    Last edited by joshsgrandad; November 16th, 2020 at 09:29 PM.

  5. #5
    Forum Elder Thegreatpuma's Avatar
    Join Date
    Mar 2016
    Location
    Druidia
    Posts
    2,516
    Quote Originally Posted by joshsgrandad View Post
    I think the MMA did a good job..
    If it were a mismatch, then that missing 10 leading into 3 lower halls has maybe made the MMA try to follow your breakdown pattern somewhat..

    They have a slight royal advantage per base, but you are compensated with RC levels..
    They have a 6 level warden advantage per base, which would be my main concern if it were us..
    Gold weight, as you said, looks fine, so defensively they are quite similar..
    That slight hall advantage could probably be written off as maxer syndrome.. If you are gonna complete your hall levels, you are gonna feel some pain if facing those that dont worry about when they hit the jump.

    Matching is only going to be approximates with only 10k differential between halls at the top 2 levels, so big 12s should always be considered a risk with the war draw..
    We do the same, but it is despite my warnings, not because of my advice..
    And as touched on earlier, once the roster starts to spread, that invites the MMA to start looking for concessions in order to give a decent draw whilst utilising a similar breakdown.

    TBF, I dont bother with the maths, I simply look at the draw result and think Yeah, that looks okay, or Hmm, wonder what happened.
    Like I said, minor mismatch - we have had far worse and this is certainly within the range of acceptable.

    And I tend to agree. I fully expect the above when we match non-rushers - often times they have extra TH levels all the way down and there's an obvious trade off. In this case though, these guys are mostly maxers too though, as shown in their troop/hero levels, although don't appear to worry about maxing walls and Ice Golems/SM.

    If you want to see specifics: https://docs.google.com/spreadsheets...it?usp=sharing

    I have my main spreadsheet in excel and need to figure out how to add in troop levels.
    TMBG War Clan always recruiting - Wars with TH9-13
    TMBG Discord server:
    https://discord.gg/byNR2xZ
    TH11 - Doctor Worm 50/50/20 heroes/near max d
    TH12 - doug 65/65/40 heroes

  6. #6
    Our current war is a 5v5. We went with 5 th13s. The enemy has 4 th13s and a th3. I suspect they were trying to get an advantage. There also is a good chance they are currently complaining in chat about their "mismatch." Just a guess. It is hard to define it.

    Contact SC here. Click here for how trophies are calculated. Click here to see how war map placement of max halls is determined. An idea to improve legends here. I wish max players had a separate loot bank as described here. Caution, I often discuss for the sake of discussion and enjoy having my opinion challenged (or approved of) even when I care little about the actual issue. My balance wish: get rid of tornado trap, make it a decoration.

  7. #7
    Forum Contender joshsgrandad's Avatar
    Join Date
    Jul 2017
    Location
    NE England, UK.
    Posts
    4,630
    Quote Originally Posted by Thegreatpuma View Post
    Like I said, minor mismatch - we have had far worse and this is certainly within the range of acceptable.

    And I tend to agree. I fully expect the above when we match non-rushers - often times they have extra TH levels all the way down and there's an obvious trade off. In this case though, these guys are mostly maxers too though, as shown in their troop/hero levels, although don't appear to worry about maxing walls and Ice Golems/SM.

    If you want to see specifics: https://docs.google.com/spreadsheets...it?usp=sharing

    I have my main spreadsheet in excel and need to figure out how to add in troop levels.
    HaHa mate.. You been around Major too long.
    Cant help with spreadsheets.. Bed sheets is more my thing.

    GL with the data crunching.. We can all benefit from someone that can annalyse breakdowns like yourself.
    Cant access that data BTW.. Am always nervous when it asks me to sign in or give permissions.. My paranoia kicks in lol.

  8. #8
    Forum Elder Thegreatpuma's Avatar
    Join Date
    Mar 2016
    Location
    Druidia
    Posts
    2,516
    Quote Originally Posted by 2222 View Post
    Our current war is a 5v5. We went with 5 th13s. The enemy has 4 th13s and a th3. I suspect they were trying to get an advantage. There also is a good chance they are currently complaining in chat about their "mismatch." Just a guess. It is hard to define it.

    Yeah, that is a tough one. I don't view that as any more of a mismatch than ours is, TBH, but with that skewed 5th hall, my guess is y'all have a 20% advantage on every one of the metrics. Clearly it's doing something like weighing a top% of the lineup more stringently than a bottom percent, perhaps even on a sliding scale.


    Quote Originally Posted by joshsgrandad View Post
    HaHa mate.. You been around Major too long.
    Cant help with spreadsheets.. Bed sheets is more my thing.

    GL with the data crunching.. We can all benefit from someone that can annalyse breakdowns like yourself.
    Cant access that data BTW.. Am always nervous when it asks me to sign in or give permissions.. My paranoia kicks in lol.
    Odd, I access it no problem in incognito mode. Something must be weird in the settings maybe.

    I took the time to do the war in our baby clan as well, and I did both wars down to specific troop levels as well. In the baby clan we matched an 86k TH10 at the top and we only have up to th9, however, the tradeoff there is that we have more than 2x the AQ levels that they have king levels are almost the same. I'm still looking at the metrics in the main clan to see if there's a tradeoff somewhere.
    TMBG War Clan always recruiting - Wars with TH9-13
    TMBG Discord server:
    https://discord.gg/byNR2xZ
    TH11 - Doctor Worm 50/50/20 heroes/near max d
    TH12 - doug 65/65/40 heroes

  9. #9
    Senior Member
    Join Date
    Jun 2019
    Posts
    437
    Thumbs up from here!

    Take a look at the mean and standard deviation of your gold weights:

    You: 90.7 35.4
    Them: 90.5 31.8

    Though not an explanation for the match, it indicates that you're defensively matched based on an average (or total) and not on spread (not primarily anyway). If I were to make a suggestion, I'd suggest you look at how the gold weights, hero levels etc are distributed around the average.

    Also, have you considered looking into spell levels? That might be contributing.

    I don't think you're gonna get around troop levels if you want to explain a match like this. But, to simplify, you could start just including IG, bowler, witch and Yeti, ED, Miner. That's of course based on the assumption that the higher troops weigh more than say, an archer or minion.
    Last edited by Munitor; November 17th, 2020 at 02:34 PM.
    Just finished CWL? Post your data here:

    https://forum.supercell.com/showthre...ght-compendium

  10. #10
    Quote Originally Posted by Thegreatpuma View Post
    Yeah, that is a tough one. I don't view that as any more of a mismatch than ours is, TBH, but with that skewed 5th hall, my guess is y'all have a 20% advantage on every one of the metrics. Clearly it's doing something like weighing a top% of the lineup more stringently than a bottom percent, perhaps even on a sliding scale.
    In the past, clans could put th3s at the bottom and it would give them an advantage up top. For example, in this match I'm describing they would be happy to give up that bottom base if it gave them an advantage in the other 4 spots. The matchmaker changed, though, and it no longer works well. It is almost like the matchmaker just ignores extreme outliers like a th3 sitting at the bottom.

    Contact SC here. Click here for how trophies are calculated. Click here to see how war map placement of max halls is determined. An idea to improve legends here. I wish max players had a separate loot bank as described here. Caution, I often discuss for the sake of discussion and enjoy having my opinion challenged (or approved of) even when I care little about the actual issue. My balance wish: get rid of tornado trap, make it a decoration.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •