Next on war docate should be manicured and rostar engineering

[2222]

lol, ok so the three options are

1 townhall weighting (we agree is terrible)

2 minimum base weight based on townhall level (your option)

3 weight tax on offence (nerfs option)

Its possible to combine 2 and 3 then, a townhall level gives the base a minimum weight of say a Max townhall 2 townhalls below BUT as the offence get stronger than defence you also have a weight tax applied to this different and add it to the minimum weight assuming the base defence weight is lower than the base minimum weight.

Thats an option, but the same result can be achieve with 3 alone, so allthough 2 is not a pure townhall weighting system (Its a dynamic one where the townhall level determines a minimum weight as apposed to a permanent weight), 3 for me is the best of all worlds.

We agree on everything other than I've lost faith in SC's ability to make a complicated matchmaking system work. I love this game and think SC was brilliant in creating it. I defend SC regularly. However, I can only keep the hope alive for so long. Eventually, I just say I give up and ask them to make a simple basic blunt change that everyone can understand. It isn't perfect, but it is better than we have now. If they would instead implement an even better change, I would be very happy. Frankly, I don't think either one will happen.

[dorsan]

I wouldn't do the "ignore bottom 20%" method, but it would help v. roster engineering. There needs to be a way to "filter out" filler accounts to prevent roster engineering.

Personally, if I could re-write the war mm to deal with roster engineering, I'd do:
1) Check for off/def weight per each base & tabulate
1a) Determine def weight distribution at clan level (i.e. # of bases at th11 def weight ranges, # of bases at th10 def weight range, etc.)
2) Take top 60-80% of off weights & count twice (to account for 2 attacks). Not sure on exact %, but something in that range. Most [casual] wars are decidedly by maybe 2/3 of available attacks.
3) Based on 1a, somehow determine "anticipated" attacks required to 3*. Low def weight bases set to "1", mid def weight to "2", and increase as weight increases.
4) Match clans with these "lane" checks - effectively, "match" required attacks to maximize stars v anticipated defenses required

Poorly explained (and not fully thought out) and not feasible without an increase to pool size, but promotes accuracy, maybe.

How is that better and less complex than simply checking offense vs offense and defense vs defense in 2 separately sorted lists?

In fact how is any suggestion ever better than that? And how come supercell can't understand this?

[Vikingchief]

How is that better and less complex than simply checking offense vs offense and defense vs defense in 2 separately sorted lists?

In fact how is any suggestion ever better than that? And how come supercell can't understand this?

They do, they actually i think discussed your suggestion and commented on it in announcement, remember they said said it would lead to war searches taking to long. I think it was in the announcement, could be wrong.

you joining the CWSF ?

[Vikingchief]

We agree on everything other than I've lost faith in SC's ability to make a complicated matchmaking system work. I love this game and think SC was brilliant in creating it. I defend SC regularly. However, I can only keep the hope alive for so long. Eventually, I just say I give up and ask them to make a simple basic blunt change that everyone can understand. It isn't perfect, but it is better than we have now. If they would instead implement an even better change, I would be very happy. Frankly, I don't think either one will happen.

3 is not hard ask nerf, its a simple formula, unless nerf has also given up.

[dorsan]

They do, they actually i think discussed your suggestion and commented on it in announcement, remember they said said it would lead to war searches taking to long. I think it was in the announcement, could be wrong.

you joining the CWSF ?

Nah, can't take any accounts out of my main clan plus running the clan already takes all the time I have for playing.

There is also a solution for the search not taking too long, but unless they are ready to compare offense to offense and defense to defense for each position it will never work well.

[Vikingchief]

Nah, can't take any accounts out of my main clan plus running the clan already takes all the time I have for playing.

There is also a solution for the search not taking too long, but unless they are ready to compare offense to offense and defense to defense for each position it will never work well.

I think they are happy to to do that, Iether they have not seen your variable prep day solution, or dont think it will work.

[dorsan]

I think they are happy to to do that, Iether they have not seen your variable prep day solution, or dont think it will work.

I believe these forums are the last place on the planet they would look for a solution. I like proposing solutions because it's fun, but if you think that they actually take anything we write seriously, you are wrong.

[Vikingchief]

I believe these forums are the last place on the planet they would look for a solution. I like proposing solutions because it's fun, but if you think that they actually take anything we write seriously, you are wrong.

I hope you are wrong, because Its looking like the forums are their only hope.

[stealthmtn]

How is that better and less complex than simply checking offense vs offense and defense vs defense in 2 separately sorted lists?

In fact how is any suggestion ever better than that? And how come supercell can't understand this?

Definitely not simpler. More robust, maybe. Overly complicated, probably. But I do that, alot.

But completely agree off v off & def v def should be checked separately (my method does that).

I'm not sure the best method to factor out roster engineering. % does not seem robust enough for all situations. Needs to be more flexible than that.

[STRIKER44]

What if they used the matching they had (which seemed to work for us about 2/3 of the time) and had a final check to the match to ensure the top offensive ths don't have a lopsided advantage, like when engineered are paired vs others? Like a final check to see that the opposing clan don't have 3 th top war offence attack troops vs 1 in the other?