Adaptive clan war matchmaker

Hello,

This isn't a feature request for a visible feature, but a server side feature request to promote fair play.

Problems:
- the matchmaker is manually adjusted with frequency low enough for players to exploit the system
- buildings weights do not accurately reflect their contribution to the war success of a clan
- offensive strength isn't weighted strongly enough to create fair wars

Solution:
- start with the current system of unit/building weights
- sort players separately by their offensive and defensive strength
- allow matches only if players on all position fall within the offensive/defensive threshold for that position (this might be modified to weight stronger players higher than weaker players)
- use dithering to redistribute player strength (if defense of player 1 in team 1 is higher by X value than the defense of player 1 in team 2, then add that X value to the player 2 in team 1 so that the matchmaker doesn't err to the same side for every player)
- after wars are completed adjust all building/unit weights by a small percentage

Example of a weight adjustment after a battle:
clan 1 has an advantage at 2 extra players having level 4 hogs and 2 canons upgraded to level 8
clan 2 has an advantage at 1 extra player having level 5 goblins and 3 air sweepers upgraded to level 5

War ends and clan 1 wins.

increase the weight of level 4 hogs upgrade by 2x the default balancing percentage
increase the weight of level 8 canons by 2x the default balancing percentage
decrease the weight of level 5 goblins by the default balancing percentage
decrease the weight level 5 air sweepers by 3x the default balancing percentage


If you repeat this process after every single war, the game will get to a state where all the upgrades weight exactly as much as they contribute to a clan strength. And if new strategies emerge or units are buffed or nerfed, the system will quickly adjust itself without any human intervention needed.

Thanks for your time,
Dorsan

This sounds nice but legitimately hard to code and store. I think the algorithm is fine as is as there are still drawbacks to prolonging 8.5.

Quote:

---

Originally Posted by Radiating

This sounds nice but legitimately hard to code and store. I think the algorithm is fine as is as there are still drawbacks to prolonging 8.5.

---

There is no additional storage needed because the weights for all units and buildings are already stored and so are the war day upgrades of every player in both teams.

The code itself is trivial, I could do it in 2-3 hours for both the matchmaking part and the auto adjusting part.

And I'm not just talking about 8.5s, I'm also talking about defenseless bases and to a smaller extent bases being upgraded with war weight in mind. When deciding for an upgrade, it's weight shouldn't be a concern of the player, it's usefulness vs cost should be.

PS: Any coding hours spent into this would pay back in the long run because there would be no need for constant manual adjustments.

Quote:

---

Originally Posted by dorsan

There is no additional storage needed because the weights for all units and buildings are already stored and so are the war day upgrades of every player in both teams.

---

But you would need different values now for each clan. That takes up some space.

Quote:

---

Originally Posted by Radiating

But you would need different values now for each clan. That takes up some space.

---

The idea is that if a L3 inferno is worth 1000 points for matchmaking for all clans.

And when a clan wins a war when it has one EXTRA L3 inferno than the other clan, then we upweight the L3 inferno (by probably a TINY amount) so that whenever ANYONE goes to war from then on and it does matchmaking it weights a L3 inferno as 1000.5 points rather than 1000)

Eventually buildings (walls, heroes and troops too) will have an effective war weight that means something based on empirical data rather than the subjective opinions of the developers.

I saw a list once that gave weightings for heroes on a purely linear scale. It made no note of the huge difference every five levels of hero makes. If that's truly how heroes are weighted, then they need all the help they can get.

This would be fantastic, and is a simple and clever idea.

Well done Dorsan.

Quote:

---

Originally Posted by dorsan

Hello,

This isn't a feature request for a visible feature, but a server side feature request to promote fair play.

Problems:
- the matchmaker is manually adjusted with frequency low enough for players to exploit the system
- buildings weights do not accurately reflect their contribution to the war success of a clan
- offensive strength isn't weighted strongly enough to create fair wars

Solution:
- start with the current system of unit/building weights
- sort players separately by their offensive and defensive strength
- allow matches only if players on all position fall within the offensive/defensive threshold for that position (this might be modified to weight stronger players higher than weaker players)
- use dithering to redistribute player strength (if defense of player 1 in team 1 is higher by X value than the defense of player 1 in team 2, then add that X value to the player 2 in team 1 so that the matchmaker doesn't err to the same side for every player)
- after wars are completed adjust all building/unit weights by a small percentage

Example of a weight adjustment after a battle:
clan 1 has an advantage at 2 extra players having level 4 hogs and 2 canons upgraded to level 8
clan 2 has an advantage at 1 extra player having level 5 goblins and 3 air sweepers upgraded to level 5

War ends and clan 1 wins.

increase the weight of level 4 hogs upgrade by 2x the default balancing percentage
increase the weight of level 8 canons by 2x the default balancing percentage
decrease the weight of level 5 goblins by the default balancing percentage
decrease the weight level 5 air sweepers by 3x the default balancing percentage


If you repeat this process after every single war, the game will get to a state where all the upgrades weight exactly as much as they contribute to a clan strength. And if new strategies emerge or units are buffed or nerfed, the system will quickly adjust itself without any human intervention needed.

Thanks for your time,
Dorsan

---

Ok first of all, how the heck needs level 5 goblins in war! Also adjust with TH compositions because some don't work as well and depends if they are really well upgraded or rushed. Its so so so annoying looking at these freaking mismatches because the system isn't in their favor. Just deal with it.

Quote:

---

Originally Posted by PureEnergy66

Ok first of all, how the heck needs level 5 goblins in war! Also adjust with TH compositions because some don't work as well and depends if they are really well upgraded or rushed. Its so so so annoying looking at these freaking mismatches because the system isn't in their favor. Just deal with it.

---

If you don't understand the idea, why do you feel the need to comment on it?

Just to avoid any further misunderstandings, the weights are not stored per clan, they are stored globally and are affected by every single clan war.

And the example I've provided was just that, an example. ;)

If the original post is unclear, please do ask, I'll clarify gladly.

Thank you for a great post. I really enjoyed reading it and you obviously put a LOT of thought into it. I don't understand about coding and that kind of stuff but you made it easy to understand. That's impressive.

The flaw I see with this system is that you assume everybody is trying their best every war. My clan always puts in a solid effort but I've helped other people's clans here and there and I've seen some absolutely pitiful war attempts.
I've seen someone use 3 giants against an entire base, saw someone use all wizards to no effect, all goblins, just sniping resources then quitting, and the list goes on...

Also, how will war farmers skew this system?

As I type, I'm thinking that perhaps your idea would only create adjustments if the losing clan got a certain score or percentage destroyed or something.

What do you think?

Quote:

---

Originally Posted by RedDestroyers

The flaw I see with this system is that you assume everybody is trying their best every war. My clan always puts in a solid effort but I've helped other people's clans here and there and I've seen some absolutely pitiful war attempts.
I've seen someone use 3 giants against an entire base, saw someone use all wizards to no effect, all goblins, just sniping resources then quitting, and the list goes on...

Also, how will war farmers skew this system?

As I type, I'm thinking that perhaps your idea would only create adjustments if the losing clan got a certain score or percentage destroyed or something.

What do you think?

---

You are underestimating the number of wars going on in the game. The low effort vs high effort and low skill vs high skill wars are going to generate mathematical noise in the system, that is true. However, the more you sample the less important the noise becomes.

Think of it like this. You are trying to take a photo in the night. Now if you expose the photo for 1/100 seconds, all you get is darkness, and if you amplify it, you will get an extremely noisy picture. However, the longer you expose the picture, the noise will be lower an lower, and if you expose it for 10 minutes, there will be no noise at all. This isn't happening because the noise is reduced, the noise is same for every 1/100 seconds for those 10 minutes. But the noise isn't consistent, so it will eventually cancel itself out.

The same thing will happen if you run the matchmaker for a long time with many clans participating. An automatic system surely wouldn't work if you'd have 100 or less clans. But with the number of clans we have in this game, the noise created by low effort or low skilled players will cancel out very fast.

To show it in a small example:

war 1:
- clan 1 has a level 3 inferno better than clan 2
- clan 2 doesn't attack and loses the war
result: level 3 inferno gets an artificial weight increase

war 2:
- clan 1 has a level 3 inferno better than clan 2
- clan 1 doesn't attack and loses the war
result: level 3 inferno gets an artificial weight decrease

If you look at both wars as single events, they both introduce an error to the system. But if you look at them at the same time, the error disappears. With millions of wars taking place, every time a no-effort clan introduces an error, sooner or later another no-effort clan will bring it back to balance.