MATCHMAKING TWEAK -- FOOTING TH LEVELS (similar to League of Legends)

[Wryhorn]

We were in one of those wars you hate today. You know the story:

• Matchup
  
  • 45vs45
  
  
• Numbers by TH level
  
• THEM-- 10 TH10, 16 TH9, 19 TH8
  
• US -- 7 TH10, 13 TH9, 22 TH8, 3 TH7
• The mismatch
  
  • 3 'extra' TH10 villages (2 max hero bases - while our highest king/queen was 20/25 - LOL)
  • 3 'extra' TH9 villages
  • no TH7 villages

I started to compile stats from the war (how many of our TH9 and TH8 were hit by a level above for the 'scoring' stars they gave up) -- but then I realized the sample set is too small.

That said, common sense (based on 65+ war wins) tells me number of villages being notably different at each TH level gives a big advantage to the with the 'extra' clan. In short, most clashers would agree:

• Its hard to 3 star a max TH8 base if setup right as a TH8, however getting a 3 star as a TH9 on a TH8 is very doable (new units, troops/spell levels, camp/spell quantity).
• Its hard to 3 star a max TH9 base if setup right as a TH9, however getting a 3 star as a TH10 on a TH9 is very doable (new units, troops/spell levels, camp/spell quantity).

So when one clan has more villages in the TH10 and TH9 levels (although you could say the same happens at lower levels) they can use that excess to hit the lower level. So in our example war above our TH9 had to use their attacks to cover the 'extra TH9' (same for our TH10). Their TH10/TH9 could cover our TH10/9 and hit down.

In the end the difference in the war was about 20 stars (I'm writing this as were still attacking). 20 stars is about what you get from 6 extra bases in that:

1. Thats 12 'extra' hitting down attacks -- netting maybe 1 star the other clan won't get - So up 12 for them
2. Thats 6 extra bases we have to hit that isn't 'hitting down'. So maybe down 6 for us.

Interesting? I told our clan the same at the start.

ANYWAY, A POTENTIAL LINE OF THINKING TO MITIGATE THIS --

Do your match process as today. However, when match is found, add a new step -- a TH footing step -- similar to how League of Legends has a step where roles are eliminated after matching up two teams. In this Clash of Clans footing step, each clan would have time to drop villages from war (without seeing enemy) to foot to the lowest common denominator by TH by clan.

Example,

using our war above, the enemy clan would come to the footing screen and be told they had to drop 3 TH10 and 3 TH9 to foot to my clans # of villages. They could pick who they drop. Now in turn my clan would need to drop 6 villages from TH8 and TH7 so number of villages by TH match. As they do this they could see the power number associated to each village (or an abstraction of that) and the process could validate real time if the villages chosen to drop were good enough to keep the power levels of each clan in range (similar). We, in parallel, would see a similar screen telling us we needed to drop all TH7 and 3 TH8 with a power rating as a clan we needed to hit in doing so.

If clans didn't respond in a certain time -- Supercell servers could do this 'footing' based on simple rule set. Something like this:

1. Loop through villages based on TH Level (Descending from TH10 till no villages left)
   
   1. Is number of villages equal,
      
      1. if yes decrement to next TH level
      2. if no
         
         1. mark villages in excess clan that are candidates to be 'benched' from war (dont drop yet). Probably just mark from bottom up (sorted in descending power level by clan by TH level)
         2. decrement to next TH level
   
   
2. Recalculate power rating of each clan without villages marked for benching.
   
   1. If one clan is too strong
      
      1. Scan up the TH levels for the clan -- for each TH level -- check if has benched village(s). if so
      2. 
         
         1. shift the 'benched village(s)' up
            
            1. (so, in my example system would start at th7 -- skip that bench village list -- as benched villages are top of that TH list sorted desc by power (all are benched). then it would look at TH8 and start shifting up my bench list in th8 (slowly move the 3 benched there up the list of TH8 villages we had sorted by power desc).
            
            
         2. if not, increment TH level and check again
      
      
   2. If complete this pass and still power rankings are off, just go with original clan matchup (no benching). You do it now anyway -- but at least the above 'tries' to foot to an even match.

You could skip allowing the clan leaders to pick who they bench to keep it simple. Just bench for them. Add some intrigue -- OH NO JOHNNY IS BENCHED!!!

LOL.

Seriously though, I think the above would go a long way to getting more even matchups without much change to how the system works now. The biggest difference would be instead of it always being 50v50 or 45v45 etc ... it might be 47v47 or 21v21.

Thoughts?

- Wryhorn

[Gama321]

Except th doesn't substantiate to anything, only the upgrades that people do when they upgrade th, which are obviously accounted for. Making your suggestion moot

[X7386]

Except th doesn't substantiate to anything, only the upgrades that people do when they upgrade th, which are obviously accounted for. Making your suggestion moot

Exactly. Having a clan all full of people who just upgraded their THs will get a guaranteed loss in wars. Also, what do you do about TH rushers?

[Wryhorn]

Except th doesn't substantiate to anything, only the upgrades that people do when they upgrade th, which are obviously accounted for. Making your suggestion moot

Fair point. Benching on TH level alone is a simplification. So, lets break that down, what upgrades really matter in TH level (enabling a TH10 to have notably more success in hitting TH9, TH9 on TH8, and so on)? In short, my first attempt at such a list would be:

1. Camp size. Usually that is done within days of new TH. 10% more attack power.
2. Extra spell. 1 extra at TH10 and TH9. This is a notable difference. Usually built within week or so of new TH.
3. New hero. Queen at TH9. Big impact on TH8. Usually built within day of new TH.
4. New spell. Freeze at TH10, Jump at TH9, etc. Usually 'opened' within day of new TH. Impact of this varies.
5. New defensive buildings. iTwrs @ TH10 are game changers, xBows and 4th AD at TH9 have similar impact. Usually those building are in place in weeks of deployment. In short, xbows block a TH8 from easily hitting an early TH9; iTwrs block a TH9 hitting a TH10 easily.

The above is why our standing guidance to clashers in our clan is, on each upgrade, to first:

• - Upgrade camps
• - Upgrade spell factory
• - Upgrade CC (enables TH9 to get max golem from clan, plus 5 more troops)
• - Build Q (at TH9)
• - Build new/key defense unlocks (iTwr, AD, etc.)

Then the rest of time at that TH level (90%?) is upgrading troops/defensive buildings. However, very early on (1-2 weeks?), a village new to a TH level can have that most of the 'hitting down' advantage as the advantage is mostly from the items listed above (camp, spell qty, hero, key defensive buildings).

So ... how would my suggestion evolve if we wanted to get more detailed (use items above instead of just TH level)?

Well, I think you do the same process of benching. However, you could have a 'floor' where each TH level actually 'counts' as being viable at that level. Being at TH10 on day 1 doesn't mean you have the impact of a TH10. However, one could easily define, by TH level, what you want to see to make a village a viable village for that TH level. Till then they are counted in the prior level.

EXAMPLE USING UPGRADES AND TH LEVEL TO BENCH

TH10 footing -- Gross compare -- Enemy has 10 TH10, we have 7

Enemy clans bottom 3 in TH 10 (weakest first):

• #1 - 2 Lvl2 iTwrs, 220 camp (not upgraded), 5 spells
  
• #2 - 2 Level 1 iTwrs, 240 camp, 5 spells
  
• #3 - 2 Level 1 iTwrs, 240 camp, 5 spells

So all three of these I would say are 'viable' and would be counted when the sytem checked if clans had equal number of 'viable' TH X villages (in this case TH10). All three have the defensive advantage (iTwrs built) which makes it (1) not viable for TH9 to hit them effectively and (2) less likely that a TH10 hitting would 3 star which isn't the case if a TH10 hits a max TH9.

All of our TH10s were 'viable'

In short -- this means all three of the TH10's of the enemy listed above would be 'benched'

TH9 footing -- Gross Compare -- Enemy has 16 TH9, we have 13 TH9

Enemy clans bottom 3 in TH 9 (weakest first):

• #1 - 2 Lvl1 xbows, 200 camp (not upgraded), 3 spells (not upgraded), AD 6661 (weak 4th ad)
• #2 - 2 Lvl1 xbows, 205 camp (not upgraded), 3 spells (not upgraded), AD 6664
• #3 - 2 Lvl1 xbows, 200 camp (not upgraded), 4 spells (not upgraded), AD 6665

So here is where the 'tweak' happens. First 2 (weakest 2) are entry level TH9. Really don't have an attack advantage on TH8. Do have a defensive advantage (via xbows and extra ADs) which might equate to 1/2 a star but lets be conservative and say they both are 'non-viable'. As such they are not counted when footing TH9 to TH9 (they would be counted in TH8 group). The third however is viable as a TH9, and as such, would be counted in the TH9 footing (and benched). Interesting side note, that 3rd attacker 3 starred two max TH8 bases using GoWiWi (level 1 with) + Q. LOL. Great example of the advantage that 'extra' TH9 creates for them -- an extra village that can 3 star TWO max bases at the lower level.

All of our TH9's were viable. Looked at bottom 3. all had xbows, q, ads, and spell fac upgraded. 2 were not done w/ camps). We'll call the 2 w/o upgraded camps as semi-viable as they do have extra spell and Q.

In short -- this means just 1 (the 3rd from bottom) of the enemy listed above would be 'benched' .. the other two (weakest two) would be re-footed (included in that count) in next TH level down.

TH8 footing -- Gross Compare -- Enemy has 21 TH8 (2 from th9 non-viable), we have 22

Our bottom 1 in TH 8

• #1 - Has 3 AD (553), Camps not upgraded (200).

TH8 has less of a game change ATTACKING WISE when it upgrades from TH7. Basically its buff in the upgrade is defensive. No extra spell. You get an additional -- AD (big deal for drag defense) mortar (hard on mass troop attacks), tesla (consider it an extra Archer Twr), Wiz Twr (good for mass atks), Arch Twr (mass), etc.

So, I would think any new TH8 is viable as they all suck. LOL. TH8's are basically 3 star magnets for skilled max dragon players till ADs are maxed and usually people get max (lvl3) dragons way quicker than a TH8 can upgrade all its ADs.

In short -- this means just 1 (our bottom one) would be 'benched'

TH7 footing -- Gross Compare -- Enemy has 0 TH7, we have 3

We only carry a few (2-3) TH7 in our clan so my insight here is less but I suspect TH7 is like TH8 -- basically an early stage TH7 doesn't have a notable 'quick' (upgrade wise) attack advantage on TH6 so any TH7 is viable as soon as they upgrade.

In short -- Regardless of the above -- since the enemy has no TH7 this means all 3 would be 'benched' ... basically taking away the three 'free' 3 stars they get.

So, to your your guys comments:

• Originally Posted by Gama321
  Except th doesn't substantiate to anything, only the upgrades that people do when they upgrade th, which are obviously accounted for. Making your suggestion moot
  
  Exactly. Having a clan all full of people who just upgraded their THs will get a guaranteed loss in wars. Also, what do you do about TH rushers?

You can see the process works for TH rushers. You can also see (Gama) that the upgrades people do are not "obviously accounted for" -- at least not well -- in my example. The enemy had three viable TH10's 'more' than we did and 1 extra TH9. Thats 8 'extra' attacks to go hit down for 3 stars. Big advantage (say they enable 3 stars on bases that would of been two starred). That's 8 a net star advantage.

Still, I'm sure this isn't the perfect answer -- rather a good start on one way one might approach it.

-Wry

[Gama321]

Okay to be honest i didnt read it at first but now I have. What you are asking for is for people to have to sit out of a match simply because there is a mismatch of undefined proportions (you didnt give the levels of every single troop, spell, hero and every defence of all players on both sides so nobody can comment on the extent odlf the mismatch).
There is no way that will happen or is in anyway fair to simply have to bench people who want to war.

I also have to reiterate that camps and everything else are accounted for in matchmaking, whether someone is a new th or a mature th is irrelevant. Everything that helps in war is accounted for to an unknown extent that is not revealed to us. mismatches will happen as there is no perfect match but you just have to deal with it. Making people sit is not the answer, and is completely unfair to those who want to war. Keeping in mind that th is meaningless the only way it would work is if you had to kick a certain amount of people until the matchmaking formula is equal for both sides. This would require sc to allow people to see the formula and skew matchups in their favor, while still retaining the problem of forcing people to sit.

[TheRealchrism]

+1...something has to be done about the match making system..I have been on both sides of the OP coin. In my last clan...we didn't even bother to attack they were so OP. I am a max TH9 def/troops/spells (lvl 15 hero's/legos) and top dog in my clan. The lowest TH10...and there were 12 of them( we have 11 TH9, no TH10).. had lvl 20 hero's and close to max defense..... when we saw the match up...we agreed why bother with the war...just chalk up a loss and save the resources. Please review and revamp the match making system.

this was a 35 v 35 they were TH10 and TH9 only

They did attack...got 92 stars score was 92-0 lol I do not think our TH7's and TH8's could have done much in this one.

[Wryhorn]

Okay to be honest i didn't read it at first but now I have.

LOL. At least your honest. No worries, your rejection based on title made me think about it more.

What you are asking for is for people to have to sit out of a match simply because there is a mismatch of undefined proportions (you didnt give the levels of every single troop, spell, hero and every defence of all players on both sides so nobody can comment on the extent odlf the mismatch).

LOL (again). I would beg to differ. I think the things I listed are close to what needs to be looked at to be considered 'viable' at that TH level (impactful the war result more than would of been prior TH upgrade). So we can agree to disagree there. Point is there is some factors of a village that, once achieved, at TH10 and TH9 make it impactful. Supercell can statistically tell that more than me. I am just listing the factors I think that matter based on a bazillion wars.

There is no way that will happen or is in anyway fair to simply have to bench people who want to war.

Agree to disagree (again). Anything can happen if the motivation is there. Supercell wants people to play the game and keep doing it for years. What is more compelling? A vibe of matchups are fair -- but some villages 'may' be benched in a given war sometimes --- or the current vibe that matchups are not fair? At what point do people get tired of match ups being unbalanced too much of the time and just walk away from the game? How fun is it to have 10 wars of which 8 are lopsided easy wins, 1 is tight, and 1 is heavily slanted to one clan? Its ok -- but it could be better. Hence my suggestion for improving the match process. Parity is the key to any great league or system of competition -- look at the NFL. People said they would never do the salary cap model but they did -- and now -- in any given 'season' any team can win. This is the reason they are growing some say (versus other professional sports that don't take as hard of a stance on parity being job #1).

I also have to reiterate that camps and everything else are accounted for in matchmaking, whether someone is a new th or a mature th is irrelevant.

I gave you an example where its not irrelevant. Maybe I'll just post a review of each war were there is an inequality on this thread. 80% of the time it is the case. 80% of the time this process or something similar could be making it a better war.

mismatches will happen as there is no perfect match but you just have to deal with it.

LOL. Is that how you talk to customers? Deal with it? I would if I felt the mismatches were rare and/or possible to overcome. However, if mismatches happen consistently and are not easily overcome -- then 'dealing with it' may be clashers quitting a frustrating/unfulfilling game. Is that what Supercell wants? I suspect not. I suspect their response to my thread would be to ponder if the algorithms for matching factor in what I am talking about.

I'll do this. I'll post on this thread each war. Notating how the matchup slants and if this benching might of helped (who). Lets see where it goes. I'll also notate a good matchup (of course).

This would require sc to allow people to see the formula and skew matchups in their favor, while still retaining the problem of forcing people to sit.

Not really, i mentioned (you must not be reading again) that supercell probably should/could do this in the background (as they do now for the matchup process). Its actually better, from a usability perspective, that way. just present the clan leaders with the bench list (and/or let them reject war based on benchings but without seeing enemy). Either way.

Good banter Gama.

-Wry

[TheRealchrism]

Keeping in mind that th is meaningless the only way it would work is if you had to kick a certain amount of people until the matchmaking formula is equal for both sides. This would require sc to allow people to see the formula and skew matchups in their favor, while still retaining the problem of forcing people to sit.

I am editing this as I go.....input appreciated

There are enough smart people out there that the formula for the match making is probably already out there.....I am sure there is an Asian clan already exploiting it...lol. Until someone, wikileaks, posts the formula, the rest of us are merely guessing on the weighting of each piece of the puzzle...heck...we may be over thinking the whole thing and may just be as simple as 1, 2, 3.......

This would be a simple weight scale
defense
lvl 1 AT = 1 point lvl 12 AT = 12 pts
cannons, tesla, xbow

AD - 4 points for every level

inferno - 5 points for every level

walls - .04 points for each wall level multiplied by TH level

or for defense ( HP/DPS ) x level for each def then add them up....this was the simplest and weights it better


tier 1 troops - 1 point for every tier 1 troop available
tier 2 troops - 3 points for every tier 2 troop available
tier 3 troops - 6 points for every tier 3 troop available
dark troops - 5 points for every dark troop available
spells - 5 points for each spell and 1 point for each level available

Percentage of max level for each troop add them up then average them out for number of troops available same for spells.

for troop availability weighting multiply the average by the TH% each TH gets 10% per level TH7 - 70%...TH 10 100%

(hp/dps(including summond units) / (housing space percentage of available camp) or ( hp divided by dps ) times housing space percentage

camps 1 point for every 5 troop capacity
CC 1 point for every level

heros, 2 points for every level.

rough estimation....I have not worked it out......lol,.... im bored at work and this popped in my head....someone check this out if you are as bored as I am.

[X7386]

Fair point. Benching on TH level alone is a simplification. So, lets break that down, what upgrades really matter in TH level (enabling a TH10 to have notably more success in hitting TH9, TH9 on TH8, and so on)? In short, my first attempt at such a list would be:

1. Camp size. Usually that is done within days of new TH. 10% more attack power.
2. Extra spell. 1 extra at TH10 and TH9. This is a notable difference. Usually built within week or so of new TH.
3. New hero. Queen at TH9. Big impact on TH8. Usually built within day of new TH.
4. New spell. Freeze at TH10, Jump at TH9, etc. Usually 'opened' within day of new TH. Impact of this varies.
5. New defensive buildings. iTwrs @ TH10 are game changers, xBows and 4th AD at TH9 have similar impact. Usually those building are in place in weeks of deployment. In short, xbows block a TH8 from easily hitting an early TH9; iTwrs block a TH9 hitting a TH10 easily.

The above is why our standing guidance to clashers in our clan is, on each upgrade, to first:

• - Upgrade camps
• - Upgrade spell factory
• - Upgrade CC (enables TH9 to get max golem from clan, plus 5 more troops)
• - Build Q (at TH9)
• - Build new/key defense unlocks (iTwr, AD, etc.)

Then the rest of time at that TH level (90%?) is upgrading troops/defensive buildings. However, very early on (1-2 weeks?), a village new to a TH level can have that most of the 'hitting down' advantage as the advantage is mostly from the items listed above (camp, spell qty, hero, key defensive buildings).

So ... how would my suggestion evolve if we wanted to get more detailed (use items above instead of just TH level)?

Well, I think you do the same process of benching. However, you could have a 'floor' where each TH level actually 'counts' as being viable at that level. Being at TH10 on day 1 doesn't mean you have the impact of a TH10. However, one could easily define, by TH level, what you want to see to make a village a viable village for that TH level. Till then they are counted in the prior level.

EXAMPLE USING UPGRADES AND TH LEVEL TO BENCH

TH10 footing -- Gross compare -- Enemy has 10 TH10, we have 7

Enemy clans bottom 3 in TH 10 (weakest first):

• #1 - 2 Lvl2 iTwrs, 220 camp (not upgraded), 5 spells
  
• #2 - 2 Level 1 iTwrs, 240 camp, 5 spells
  
• #3 - 2 Level 1 iTwrs, 240 camp, 5 spells

So all three of these I would say are 'viable' and would be counted when the sytem checked if clans had equal number of 'viable' TH X villages (in this case TH10). All three have the defensive advantage (iTwrs built) which makes it (1) not viable for TH9 to hit them effectively and (2) less likely that a TH10 hitting would 3 star which isn't the case if a TH10 hits a max TH9.

All of our TH10s were 'viable'

In short -- this means all three of the TH10's of the enemy listed above would be 'benched'

TH9 footing -- Gross Compare -- Enemy has 16 TH9, we have 13 TH9

Enemy clans bottom 3 in TH 9 (weakest first):

• #1 - 2 Lvl1 xbows, 200 camp (not upgraded), 3 spells (not upgraded), AD 6661 (weak 4th ad)
• #2 - 2 Lvl1 xbows, 205 camp (not upgraded), 3 spells (not upgraded), AD 6664
• #3 - 2 Lvl1 xbows, 200 camp (not upgraded), 4 spells (not upgraded), AD 6665

So here is where the 'tweak' happens. First 2 (weakest 2) are entry level TH9. Really don't have an attack advantage on TH8. Do have a defensive advantage (via xbows and extra ADs) which might equate to 1/2 a star but lets be conservative and say they both are 'non-viable'. As such they are not counted when footing TH9 to TH9 (they would be counted in TH8 group). The third however is viable as a TH9, and as such, would be counted in the TH9 footing (and benched). Interesting side note, that 3rd attacker 3 starred two max TH8 bases using GoWiWi (level 1 with) + Q. LOL. Great example of the advantage that 'extra' TH9 creates for them -- an extra village that can 3 star TWO max bases at the lower level.

All of our TH9's were viable. Looked at bottom 3. all had xbows, q, ads, and spell fac upgraded. 2 were not done w/ camps). We'll call the 2 w/o upgraded camps as semi-viable as they do have extra spell and Q.

In short -- this means just 1 (the 3rd from bottom) of the enemy listed above would be 'benched' .. the other two (weakest two) would be re-footed (included in that count) in next TH level down.

TH8 footing -- Gross Compare -- Enemy has 21 TH8 (2 from th9 non-viable), we have 22

Our bottom 1 in TH 8

• #1 - Has 3 AD (553), Camps not upgraded (200).

TH8 has less of a game change ATTACKING WISE when it upgrades from TH7. Basically its buff in the upgrade is defensive. No extra spell. You get an additional -- AD (big deal for drag defense) mortar (hard on mass troop attacks), tesla (consider it an extra Archer Twr), Wiz Twr (good for mass atks), Arch Twr (mass), etc.

So, I would think any new TH8 is viable as they all suck. LOL. TH8's are basically 3 star magnets for skilled max dragon players till ADs are maxed and usually people get max (lvl3) dragons way quicker than a TH8 can upgrade all its ADs.

In short -- this means just 1 (our bottom one) would be 'benched'

TH7 footing -- Gross Compare -- Enemy has 0 TH7, we have 3

We only carry a few (2-3) TH7 in our clan so my insight here is less but I suspect TH7 is like TH8 -- basically an early stage TH7 doesn't have a notable 'quick' (upgrade wise) attack advantage on TH6 so any TH7 is viable as soon as they upgrade.

In short -- Regardless of the above -- since the enemy has no TH7 this means all 3 would be 'benched' ... basically taking away the three 'free' 3 stars they get.

So, to your your guys comments:

• Originally Posted by Gama321
  Except th doesn't substantiate to anything, only the upgrades that people do when they upgrade th, which are obviously accounted for. Making your suggestion moot
  
  Exactly. Having a clan all full of people who just upgraded their THs will get a guaranteed loss in wars. Also, what do you do about TH rushers?

You can see the process works for TH rushers. You can also see (Gama) that the upgrades people do are not "obviously accounted for" -- at least not well -- in my example. The enemy had three viable TH10's 'more' than we did and 1 extra TH9. Thats 8 'extra' attacks to go hit down for 3 stars. Big advantage (say they enable 3 stars on bases that would of been two starred). That's 8 a net star advantage.

Still, I'm sure this isn't the perfect answer -- rather a good start on one way one might approach it.

There are enough smart people out there that the formula for the match making is probably already out there.....I am sure there is an Asian clan already exploiting it...lol. Until someone, wikileaks, posts the formula, the rest of us are merely guessing on the weighting of each piece of the puzzle...heck...we may be over thinking the whole thing and may just be as simple as 1, 2, 3.......

This would be a simple weight scale

Well, why don't we come up with a perfect solution? Supercell has the players' strength "scores" out there (a weight scale), to match wars. They can use a range involving those - Example:

10 v 10 war. Suppose the scores for 1 clan are
90
87
40
40
40
40
40
40
40
40

and the other clan is
50
50
50
50
50
50
50
50
50
50

Theoretically, this is a close to perfect "match" - the combined scores are different by 3 points. But 1 and 2 in the first clan will bring it to victory.

If you eliminate by scores instead of TH levels, elimination would look like this - eliminate players 1-2 and 2 other players in the first clan, and 5 players from the second clan. This will be 6 v 5, but it will be a better match.