War Matchmaking Iteration

[Chadwicke]

Thank you good work today

 

[Kanechoigo]

You should shut down the WW before fix all the problems

 

[Chadwicke]

BHG_Muet this is the 3rd war for us (Viking rum club) that we have searched for 30 mins and got a team that is exactly 35% bigger than us the decimal is in the wrong place if it's opening to 35% variance in 30 mins as you said it's supposed to be 3.5% per 30 mins

 

[oddin]

sorry, I don't understand something. What do you mean by < a team that is exactly 35% bigger than us > ? How can you possibly measure that? Can you explain more pls?

 

[Theoneandonly]

Chadwicke,

you are doing the same mistake as last Time. As you don’t know anything on the search of the opposite team you are not able to judge. The other team might have searched six hours or more and got matched then against Viking Rum. Plus the war weight bug on digital buildings ...

 

[jagadeeshgarapati]

Just consider overall level than any individual. This is going round robin. One day one building will not be calculated. Next some troop, next some new levels. Again we see same issue.

 

[Cannibals]

Presuming matchmaking is working anything like what has been communicated, the first part of what you said makes no sense. Each team’s search parameters widen over time. In order for a match to occur, both teams’ search parameters must be wide enough to accept the gap in strength between the teams. The bigger team’s search parameters will widen faster since it is based on a percentage of the total, so if anything the bigger team will have searched less time. The mismatch is more likely due to SA and DA weights being off.

 

[SomeRandomPlayer]

I highly doubt this is true, as I highly highly highly doubt you know exactly how war weight is calculated and performed these detailed calculations on every base (both opponent and yours). Come on, man.

 

[Theoneandonly]

You are not right on this point. The system doesn’t compare all parameters, only the weight and the time.

 

[BHG_Muet]

War weight fix for Digital Age should be live now. There were 13 different entries that were missing weights (I don't have the full list in front of me at the moment or I'd share them). Let me know how weighting feels in your next Wars!

 

[Cannibals]

There is ample evidence that SA war weights may have some issues as well. Are those fixed too?

 

[kumapuu]

Just try a matching system that completely depends on "glory", like other games with rank system.
Current matching system using "weight" would not work well.

Hey all,

We've been looking at the pain points discussed here over the last several months for high-end Alliances and match quality. After reviewing a lot of wars and auditing queue health, we've decided to move forward with an iteration to the queue that considerably extends the wait time before reducing the potential match quality. This change went live today and we're keeping an eye on its performance.

What impact will this have?:

• Increased time to find an opponent. This update has greatly extended the wait time before the matchmaking system starts making large trade-offs in match quality. As a result, you may notice longer wait times to find an opposing Alliance as the system slowly extends its search.
• Higher quality matches. We do understand the impact a poor match can have in DomiNations (given how long wars last, among other pain points) and are hoping this change will result in better matches for more Alliances. We're not expecting much of a change for mid-tier Alliances (match quality there was already pretty good after the system overhaul). But, the extra wait time should allow the smaller pool of top-end Alliances to better find each other instead of punching below their weight class.

This of course doesn't mean that all war matches will be perfect moving forward. But, you should hopefully see an improvement in your opponents in exchange for waiting a little longer in queue. As always, keep your feedback coming and let me know what you think of your upcoming wars.

Thanks!
Muet

 

[Cannibals]

My understanding is that in order for two teams that have a 35% difference in weight to match, both teams’ must have waited in the queue for enough time for their weight parameter to grow to 35%. One team that is in the queue for a long time would not be matched with another team that just entered the queue based only on the first teams parameters.

 

[SomeRandomPlayer]

This is a bad idea as glory should not be included at all. War lineups change, alliance members change, new alliances are born and merged, and so on.

 

[Ramesses the Great]

Here are our 4 wars (15v / 20K glory range) since this MM update:

1- 2SA (275,255) 5CWA(230s-220s) 3AA(200-180)+GA to GP - 20K vs 2SA(267,254) 6CWA(240-)+GA to GP - 19K
Search Time: 1h45m

Equal match. We won with relative ease.

2- 2SA(275,255) 4CWA 5AA+GA to GP - 20K vs 1DA(292 with Zhukov 78) 3SA(268 with Z49,263,250) 1CWA6AA+GA to GP - 20K
We rotated in lower weight players in the top
Search Time: 1h

Mismatch (2 SA v 1DA 3SA). Opponent had advantage with SA numbers. We lost.
Enemy #1 was reported for cheating and we assumed he got a temp ban (?). Still his base defended as it wasn't reset during war.

3- 2SA(275,255) 4CWA 3AA+GA to EA - 20K vs 1SA(257 with Z70) 3CWA(G60s) 4AA+GA to CA - 20K
We were slightly less weight.
Search Time: 1h

Okay match. We had SA advantage and they had #1 with lvl 70 Zhukov + 3D8. Could've gone either way due to their top maxed generals. We won.

4- 2SA(275,255) 4CWA 4AA+GA to IA - 20K vs 1DA(286/Zhukov62) 5SA(281/Z42, 274/Z42, 271, 269/Z62-H42, 257) 4CWA(245,241) 2AA1GA2CA - 20K
Search Time: 1h30m

Pretty bad mismatch (2SA 4CWA v 1DA 5SA 4CWA). Currently in that war, but unfortunately, the result is expected with big glory loss.

--

Worth noting we are similar weight with minor changes (2-3 rotations per war) in all these wars, yet the opponents vary wildly.
Also that we're generally a low weight alliance.

--

Our matches quality hasn't improved. Only changes we've noticed:
1- It takes 2-3x longer than before to match.
2- We get same glory range (19-20K glory) opponents. Before it ranged from 16.5-20.5K.

--

Statistics to consider, since April (new MM):
1- We win 80-90% of wars in which we have equal number of SA bases (equal).
2- We win ~50% of wars in which the enemy has 2+ more SA (disadvantage). - very common nowadays
3- We've never lost a war in which we had more SA than the enemy (advantage).

 

[Net55]

Another bad mismatch
after 10 hours of searching, and a reset of search 2 hours before our match, we got another really bad match

our team Their team
DA 2. 3
SA 2. 5
CWA. 0. 3

as you can see, even if our top players 5 star both their attacks, that leaves out AA players to have to attack their SA’s

fyi their top levels are also much higher than ours

 

[FrostMr]

Why not introduce a league for wars. If I want to hit low weights i have to drop leagues and try stay down there. When I opponent search I get my own league or those near by either side. I don’t know the exact details behind it for war but there are plenty of brains here who could work it out.

These war weights seem to be missed often, why not base simply of EXP if doing it that way and not looking at league.

 

[Bootney Lee Fonsworth]

Back in the day I wondered why alliance leagues weren't a thing. Hell, the system was already there for individuals it merely had to be adapted. Now? I don't think there are nearly enough active alliances to bother.

 

[SomeRandomPlayer]

Our war matching has improved lately; in fact, we've had two excellent matches in a row. I only mention this anecdotal evidence as I complained in the recent past.

Granted we've become paranoid and won't let war search run for more than an hour; and ideally we could let it go forever yet still keep the good match variance low (even if it takes 100 hours of searching). Having to worry about restarting war searches is a real pain and stressful.

 

[Rhabarberbarbar]

Fully agree!