What is known about the match-making algorithm? Help us flush this out.

Hi Everyone,

First, apologies for rehashing an old topic. I tried to stick this to an existing thread but couldn't quite find a good one that struck the correct tone.

Second, I believe the match making algorithm is less than ideal (at best). However, I don't want this thread to be a complaint session. Rather, I view this as a fact-finding mission to help clans understand how the algorithm works so perhaps clan leaders can make some decisions to bend it to a more favorable result (I'm looking for a clean, even war - NOT an advantage).

Some background, we started our clan in late 2014 and have warred about 40 times now. During our first 25 wars I really didn't notice any major mismatches. However, it seems just about all of our wars are mismatches these days (a few in our favor, but most working against us).

For example, our two most recent wars. Both 25 man wars. Here are a couple side-by-side comparisons:
War 1
Us - 3 TH10s, 1 TH9, 7 TH8s, 11 TH7s, 1 TH6, & 1 TH5
Them - 3 TH10s, 3 TH9s, 10 TH8s, 9 TH7s.

War 2
Us - 2 TH10s, 1 TH9, 9 TH8s, 11 TH7s, 2 TH6s
Them - 5 TH10s, 3 TH9s, 13 TH8s, 2 TH7s, 2 TH6s

Clearly, something is amiss here. I can go into greater detail on this second war if needed. There are some rushed bases on the opponents side which makes this a little more even than it appears (I do not believe the other side is running any .5 bases as their troops are average) but it is still quite lopsided.

I would say out of our last 10 wars there has been maybe 1 war which was a good, even match. What I can't place my finger on is what changed from our first 25 where I would say 20-23 of them were good even matches.

To further confuse me, I have another account that I use with a completely different clan. Ironically that clan is very similar to ours. They are on a 40 some war win streak and as far as I can see they haven't been on receiving end of a mismatch for at least 6 months. A further kick in the gut is they have yet to score as many stars as my original clan has in all of our past 15 wars. And we have two losses in there, they haven't even had an opponent within 6 stars.

Both clans start our war search at roughly the same time and have similar makeups (I keep expecting us to draw each other but it hasn't happened). I'm trying to do some research to see if there are some small differences in the clans which are affecting the match making process.

What information do we have about the match making algorithm? I know there are some clans who organize themselves so they can square off against each other. And other clans who seek to 'ambush' by using their knowledge of the match making algorithm to get a specific opponent. What do these clan know that the rest of us (or maybe it's just me) don't?

Ever notice how each base has a different-ish war loot value if you attack it?

After the war is over, try to visit the other clan and see what each of your bases is "worth." Compare that to what their bases are "worth."

Good luck!

I have no idea what goes into the system, but we pretty much just decide to deal with the punches. We encourage our members to max offensive structures, farming/warring troops while doing ADs and splash defenses on defense, then moving up in our clan, as we know that it will get us matched against lower THs that do the "max everything route," but as far as to what extent, we have no clue.

There is 1 TH10 account (my main), 12 TH9s, ~10 TH8s, then a hodge podge of TH7s down to TH5s in our 50 member wars, and 2 wars ago we got matched up against a clan with 16 TH9s and about 15-17 TH8s, but no TH10s. Won it 149-115, our closest ever to a perfect war with 50 members. And on the surface it seemed a bit outmatched, although we had 2 TH10 level offense attacks in hand which was huge, so maybe the mismatch was in our favor.

Cut to last war, where we faced 5 TH10s (only one of them below any TH9s on the board), 9 TH9s, and 17 TH8s. 4 of the TH10s had TH10 level troops/offense, and all 5 had infernos. Seemed like a pretty significant mismatch against us (our hero levels aren't great either to offset it, ranging from 20/20 down to 5/5 for most of our TH9s and my main). We still won 140-107, as their organization and execution was atrocious, but we knew we had to settle for 2 star attacks on the TH10s, with the occasional unlucky 1-star.

These matchups require us to adapt based on our opponents. in terms of the types of cc troops we use as well as how we assign our TH9/10 attack targets, but it would be nice to know how the algorithm works/is weighted with more transparency.

I don't know why one of your clans would have a different experience than the other if the makeups are very similar, but I would suggest adding a couple offensively strong TH9's and getting rid of the 5's, 6's, and some of the 7's. I believe you should have a minimum of 3 or 4 TH9's that can three star other 9's. This will leave your 10's to deal with the enemy 10's and you won't have to get your 8's to attack enemy 9's. We like our makeup to be about 1/8 TH7's, 1/4 TH8's, 1/2 TH9's, and 1/8 TH10's. Too many 10's is bad and too many low guys seems to be bad as well.

Quote:

---

Originally Posted by Uthrac

Ever notice how each base has a different-ish war loot value if you attack it?

After the war is over, try to visit the other clan and see what each of your bases is "worth." Compare that to what their bases are "worth."

Good luck!

---

Correct me if I am wrong, but those values are dependent on the level of the attacker as well, correct? Do I, as a TH8, get the same war win bonus for attacking an opposing TH5 as one of my teammates who is a TH5?

Thanks for the suggestion. Hopefully our current string of Chinese and Korean opponents will end so I can speak the same language as the opposition to try it.

I have not read it directly from an SC post or any specific source (just random posts here on the forum) but after the last patch supposedly TH level was worked into the equation as a secondary system. That is to say, after walls/defenses/troops/heros are added up and each clan is scored the new system supposedly tries to find a clan with a similar number of TH matches to pit you against. I am not sure I believe it yet but I have read posts to this effect more than once.

That said, when comparing strength the TH number is meaningless. It COULD mean that they are maxed, but it often means they are rushed or using a .5 strategy. There are a boatload of TH9's out there now with nothing upgraded except a few walls and lots of troops. The fact that you only posted the TH numbers from each war implies that you did not analyze the matchups closely. I know I have posted here about clans that had far more TH10's and TH9's than we did and we still rolled them. We have also been worked over by very organized clans that appeared to be an even match.

Moral of the story is, TH has nothing to do with how strong/week/good/bad/developed/rushed a given village is. All it indicates is what the COULD be.

Quote:

---

Originally Posted by Reddigulous

Correct me if I am wrong, but those values are dependent on the level of the attacker as well, correct? Do I, as a TH8, get the same war win bonus for attacking an opposing TH5 as one of my teammates who is a TH5?

Thanks for the suggestion. Hopefully our current string of Chinese and Korean opponents will end so I can speak the same language as the opposition to try it.

---

Good question. Worth investigating - though if you are the one looking at the value of the enemy bases vs. your own bases, the same "factor" would be applied.

Google Translate is a sloppy, but effective, way of communicating with foreign clans.

Quote:

---

Originally Posted by Goobers

Cut to last war, where we faced 5 TH10s (only one of them below any TH9s on the board), 9 TH9s, and 17 TH8s. 4 of the TH10s had TH10 level troops/offense, and all 5 had infernos. Seemed like a pretty significant mismatch against us (our hero levels aren't great either to offset it, ranging from 20/20 down to 5/5 for most of our TH9s and my main). We still won 140-107, as their organization and execution was atrocious, but we knew we had to settle for 2 star attacks on the TH10s, with the occasional unlucky 1-star.

---

You bring up a good point which I've been wondering about. Does the matching algorithm take into account some kind of attacking strength? As you stated, the opposition in this last war really had the advantage but wasn't organized and failed to capitalize. That happened in 'war 1' listed in the original post. Their TH10s were far higher than our TH10s. If they had been merely average attackers we never would have had a chance but ended up squeaking out a 68-64 win.

Does anyone know if the matching algorithm somehow accounts for 'average total stars' from past wars by chance?

Quote:

---

Originally Posted by Reddigulous

Does anyone know if the matching algorithm somehow accounts for 'average total stars' from past wars by chance?

---

I would be surprised if the algorithm is much more than:

1. Calculate the "value" of each base (total defense and offense).
2. Sum the values.
3. Match with a clan with similar sum.

Quote:

---

Originally Posted by Uthrac

I would be surprised if the algorithm is much more than:

1. Calculate the "value" of each base (total defense and offense).
2. Sum the values.
3. Match with a clan with similar sum.

---

SC has been very open with exactly this information. They have provided nothing with more details about the value of each troop/building nor what the "similar" range is.