The final solution to CW matchmaking

Google uses it. Netflix and amazon use it to recommend shows and products. Every research area from drug design to linguistics is using it. Why hasn't the gaming community embrace it? I'm talking about deep learning. Over the years, SC has acquired millions if not billions of data points from each war: what troop levels and building defense levels each player had in each war. These could be used to train a deep neural network, which would predict exactly one thing: given a player's set of offensive troop capabilities, how many stars would they obtain on a given base.

Matchmaking is really easy then. Just predict for each player's profile (troop/hero/spell levels/camp size) how many stars they'd get on any other player (this will not be 0,1,2,3, it'll be something like 0.47 or 1.68 to reflect uncertainty). Now use the optimal hungarian algorithm for scheduling tasks to decide how to do the optimal matchups for the first and second attacks (bonus! this would allow them to finally make the recommended base actually work!). This is quite easy and fast computationally (much faster than animation algorithms that are used all over in CoC), wouldn't add a bit to war search time. The MM algorithm would then only match up wars where each side would have an equal number of expected stars, within some tolerance.
This neatly and perfectly handles the case of those engineered bases; the deep learning algorithm would quickly see that the offense could easily trump a maxed TH of similar level. From a software engineering point of view, this is the only right way to address the current problems that exist in match-making.

The end result is no further reward for engineered bases. They'd just end up matched against other engineered bases. No more fluffy-bunny throws the war. Only close wars that are determined by skill and player participation.

For the record, I'm a deep learning practitioner and would be happy to serve as a consultant if SC wants to finally fix the problem by going this route.

I find no problem at all in the current war matchmaking and this would be abused.

Sounds interesting. Why is this Hungarian algo better than the current MM system?

You compare the power of Google or Amazon to ... Softbank (Supercell's holding) ! lol
Our small dev team can't even dream to deep learning.
Electronic Arts maybe. But not Supercell.

This suggestion is advanced. Sometimes i do feel sorry for opponent clans and wish to be matched with other clans of equal calibre.

Quote:

---

Originally Posted by Radiating

Sounds interesting. Why is this Hungarian algo better than the current MM system?

---

The hungarian algorithm would just be used to decide which player from team A should attack which player from team B, in a way to maximize total war stars for team A (given the expected # of stars that each player from team A would get on each player from team B's bases). This is a complex algorithm but available in myriad free software packages. It takes about 300 lines of code if you write it neatly from scratch, and for most programmers wouldn't take more than a couple days to program even without using a freely available implementation. This is already something that an intelligent war clan could do if they had a wide separation between attack vs defense strengths; too wide to do the standard mirror strategy. But most don't because this is a game and everyone likes to believe all their TH9s could 3* any other TH9 and so forth (rarely the case, but estimating exact numbers for each attacker is hard for us mere mortals).

The deep learning / artificial neural network would just estimate how many stars each player from team A would get on each of team B's bases, which is thus the input to the hungarian algorithm.

Quote:

---

You compare the power of Google or Amazon to ... Softbank (Supercell's holding) ! lol
Our small dev team can't even dream to deep learning.

---

My reference to netflix/amazon/google was simply ground the field in technologies that people are already familiar with. While google/netflix are among the forefront of developers for deep learning technologies, the software is freely available and ready to use by anyone. For example, the packages Caffe ( http://caffe.berkeleyvision.org/ ) and theano http://deeplearning.net/software/theano/ already provide implementations; all that is needed is the data from the wars that have occurred.

The end result would be that farming upgrades, like goblins for example, would have 0 effect unless that was the only troop I had, for example. While golems and valks and wizzies would have a lot of weight, and indeed I would expect them to be synergistic off one another (e.g. the effective weight for having max valks and max golems would be be much greater than the weight for having either max separately). The main point is that it would allow us all to upgrade our base however we please without worrying about how it will change our weight in war. Rushed bases wouldn't be an issue; your opponent might end up with a rushed base, or more likely the algorithm would realize his offensive strength is great enough and the rushed base is low enough on the war map it'll get 3*'ed anyway.

It would probably slow down clash even more. It would also be useless.

Quote:

---

Originally Posted by zoelem

It would probably slow down clash even more. It would also be useless.

---

It's just a better way of doing matchmaking. You have no idea what effect it would have on matchmaking speed or effectiveness. The Hungarian algorithm is fast, not slow...

Quote:

---

Originally Posted by zoelem

I find no problem at all in the current war matchmaking and this would be abused.

---

How on earth would this be abused? You clearly don't even understand how it would work.

One thing I think is sc haven't stored data of all those wars.

Quote:

---

Originally Posted by PhilISuppose

It's just a better way of doing matchmaking. You have no idea what effect it would have on matchmaking speed or effectiveness. The Hungarian algorithm is fast, not slow...


How on earth would this be abused? You clearly don't even understand how it would work.

---

I think zoelem is just a guy who keeps spamming on posts without understanding them just to get better rank like form veteran.

I was thinking of this the other day when we faced off against a highly engineered clan. We lost the war in a tie. But only because our top bases did not preform. So my thought was that the only real advantage a engineered clan has is a couple of extra attacks at the most, up top to get the stars or percentage points it needs to win, but if your guys just take care of business then there is no worry when facing an engineered clan.