Collecting Clan War matchups...for science...and stuff!

[WH7DX]

I posted an example of my match break down earlier which includes an extreme amount of detail. Still waiting to hear from Darian what else he thinks he needs before I put in the effort of typing our details for a 30 vs 30. Besides, I think it was an even match. Warden levels were almost equal, we probably won in the hero department. I think overall they won defense and offense (sheer number of wardens and troop levels).

Had they actually been able to do something against our top bases other then....haste loons to pull cc, use heroes to kill cc, dump loons/hounds/warden/support with spells/get 1 star then rinse and repeat, it might have actually been a close war.

Ok I guess that was the one from 2 days ago. Got it.

[SuperStorm103]

Just got a 25 vs 25 match in 27 minutes. One of the best I have seen to date. Both sides 6 th11's we have 7 th10's they have 6. 11 9's for us 13 for them, and our last is a th8. Slight edge in heroes and warden levels to them. They have an extra eagle, but we are up 4 infernos. Will do a detailed match if I hear back from Darian.

Going to be a brawl.

[tanaka04]

dear supeecell, how can you be unfair on us? we faced many th 11 with max heros, troops, n defense as our opponents, we got tougher opponents every war after new updating, it made me to get bored playing war.
please fix your systems on war matchups, and please be fair on us who didnt spend much money for buying gems.
this is one of our opponent whom we face this war: #9CYVGCJ8
thanks..

clash on...!

[urizano]

I understand this is a tedious request, but it does give us a greater understanding of what you are seeing.

In fact, too tedious.

Consequently, very few contributers will do it, to that level of detail.

I suggest, in this thread, remarkable match-ups are posted as
#<clan1> - #<clan2> - date
together with a comment on why this is perceived as unfair - in plain language.

In this form, Supercell developers have an easy time to analyse it if they wish to. I actually do think, this is the case, and they are working hard for a solution which works for the vast amount of wars.

Finding really fair matches is an increadibly difficult task as clan teams can be different in a large number of dimensions. Changing any parameter has a huge effect in the overall system.

In this thread, as I understand it, developers looking for the worst outliers in the perception of the player.

Making it tedious is probably meant as some education for the players. To acknowledge how increadibly difficult the task to make a fair matchmaking algorithm in fact is.

[urizano]

I may start with this:
#L0PCGRPU - #2JRJCUJ0 - 28-09-17 20:35 GMT
This match is remarkable as #L0PCGRPU was extremely balanced in offence and defence in each account, a clear cut example of a "non-engineered clan". #2JRJCUJ0 however, had some extreme lopsided accounts in it: defenseless Rh11 with some max troops. Apart from that, predominantly .5 and .25 accounts.
The overall match, however, was reasonably fair. #L0PCGRPU has apperently given up at some point, but with the available troops it could have been really close.

[Tomville]

In fact, too tedious.

Consequently, very few contributers will do it, to that level of detail.

I suggest, in this thread, remarkable match-ups are posted as
#<clan1> - #<clan2> - date
together with a comment on why this is perceived as unfair - in plain language.

In this form, Supercell developers have an easy time to analyse it if they wish to. I actually do think, this is the case, and they are working hard for a solution which works for the vast amount of wars.

Finding really fair matches is an increadibly difficult task as clan teams can be different in a large number of dimensions. Changing any parameter has a huge effect in the overall system.

In this thread, as I understand it, developers looking for the worst outliers in the perception of the player.

Making it tedious is probably meant as some education for the players. To acknowledge how increadibly difficult the task to make a fair matchmaking algorithm in fact is.

I agree it is tedious. I have been one of the few doing it, but then it’s much easier for me than many due to small war sizes with many identical bases at fully max.

The problem with any statement as to why it’s perceived as unfair is that the perception of unfairness may well be off and where it is there is no non-onerous way to investigate to establish that fact. So the only practical way is asking for some evidence beyond the perception.

Lots of things have been called “unfair” here from the obviously completely unfair (they had double the number of max th11s) to the moderately unfair mostly dependent on roster position (my th10 has matched a max th11 at the top of the map) to the lineball (my multiple th9.5 matched multiple defence max th10s) to the completely ridiculous (my single 9.5 matched a max 10; they had one extra th x in the middle/at the bottom).

[blombardo]

ugh. come on folks; build a spreadsheet and create a google form to speed things up... dont make me do it.. I have better things to do with the next 90 minutes of my life lol...

Was hoping Darian would have someone create it and he could edit post 1 to include a link to it... I fear if a random scrub like me puts one up it could get buried on page 14 and nobody will properly use it..

[Vikingchief]

ugh. come on folks; build a spreadsheet and create a google form to speed things up... dont make me do it.. I have better things to do with the next 90 minutes of my life lol...

Was hoping Darian would have someone create it and he could edit post 1 to include a link to it... I fear if a random scrub like me puts one up it could get buried on page 14 and nobody will properly use it..

its a team effort, and a great idea, post it i reckon and maybe Darian will get on board, but please allow us to record the 2 max army comps at least, so we can input level 3/3 BOWI, lev 7/6 lavallool etc. And freeze spelll level If possible.

now that would add greatly to forum discussion, and If supercell would place defence into the API or APK data, between you, dorsan and others who knows the possible value that could be added.

[urizano]

So the only practical way is asking for some evidence beyond the perception.

Sure, for the communication among us (Supercell users) that would be helpful.

But the Supercell developer have data at hand. This is what I was talking about.

[LaceyStacey]

Our last three matches were pretty fair. The war match we just pulled is really bad / unwinnable...

• The opponent has 1 more eagle than we do.
• The opponent has 4 more wardens than we do.
• The opponent top 5 defenses are significantly higher than ours.
• The opponent heroes are higher than ours.

No	TH	BK	AQ	GW	Loot Entry	Def Weight	Def Diff	Def Weight	Loot Entry	BK	AQ	GW	TH
1	11	35	37	15	20600	103	-6	109	21800	45	45	20	11
2	11	31	38	16	20400	102	-7	109	21800	45	45	20	11
3	11	45	45	20	20400	102	-7	109	21800	45	45	20	11
4	11	44	45	20	18800	94	-11	105	21000	45	42	19	11
5	11	45	45	20	18200	91	-5	96	19200	44	44	20	11

Next match is here and another lopsided match. This one is the opposite of the last war (yet the same players on our side). We have no chance to lose this one and it is not fair for the other clan... This match took 2 hours and 6 minutes. We just lost the last war above 114 to 117 (obviously).

• The opponent top 5 defenses are significantly lower than ours.
• The opponent has 3 fewer eagles than we do.
• The opponent has 3 more wardens than we do.
• The opponent heroes are all significantly lower than ours.

No	TH	BK	AQ	GW	Loot Entry	Def Weight	Def Diff	Def Weight	Loot Entry	BK	AQ	GW	TH
1	11	35	37	15	20600	103	11	92	18400	22	24	10	11
2	11	31	38	16	20400	102	11	91	18200	20	30	9	11
3	11	45	45	20	20400	102	11	91	18200	24	33	13	11
4	11	44	45	20	18800	94	4	90	18000	25	27	10	11
5	11	42	45	20	18200	91	1	90	18000	25	29	7	11

I looked at the info requested Darian, and I do not have the 10 hours to collect this data. It is my belief that with the info I provided above I can predict when we have an easy/fair/impossible matchup.

• I do not personally think that the other bases matter much outside of the top 5 (maybe top 10).
• I do not think that the troops matter (besides heroes). I assume that if someone has a th11, then they have the troops maxed that they are best with.
• I do not think that individual defenses matter like you have in your spreadsheet example. The defense gold storage weight pretty accurately tells me how hard a base will be.
• It does not matter too much that they have engineered bases. as long as the hero levels and number of wardens are taken into consideration.