War Matchmaking Iteration

[BHG_Muet]

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

 

[boomboomboom]

Glad to see this working - not. Biggest mismatch in the history of matching. Thanks for making it worse then before

 

[SomeRandomPlayer]

Excellent news!

If given the choice between waiting 24 hours for a good match or waiting 24 minutes for a bad one; most would prefer waiting 24 hours for a good match.

That said, many alliances would restart a war search after an hour thinking that was when the search widened beyond a desirable point. All tips are welcome. For example, does the criteria do a big jump after X hours now or is the criteria expansion gradual over time?

 

[oddin]

Great news! This is a welcome change. As a leader of a mid tier alliance (~around 18500-19000 glory points), with a variety of age distribution, I can say that our matchups are mostly fair. What we find to be a little out of place is glory loss or gained at the end of a war.

We fought the other day, against a 16000 glory alliance that was way stronger than us and we lost 450 points. I was told that glory gained/lost is relative to each team's position on the leaderboard. Shouldn't the system take into account the relative strength of each alliance in this calculation? Wouldn't that further improve glory points awarded?

In addition, since you are responsible for the core programming and general direction of this game, I would like to receive more communication from you. Key points are:
a) bug fixes
b) cheating
c) future updates

Last of all, I strongly believe that the quality of each update, could vastly improve by implementing beta testers from a pool of players from all ages. Is that something you are planning to do?

 

[SomeRandomPlayer]

This anecdotal observation helps nobody; you don't even know if the new system was involved in your match. Please wait a couple wars before making such a blanketed statement. *sigh* https://www.xkcd.com/386/

 

[BHG_Muet]

It is certainly more gradual and the goal isn't to wait quite that long; it should prioritize a match well before 24h. But, as an example, in the 10v10 queue, the old set of data allowed for a 20% difference in the two Alliance's Offensive/Defensive ratings after just 30min. The new set of data is keeping a tighter clamp, only expanding to 3.5% difference by 30min.

 

[SomeRandomPlayer]

Thank you, this is helpful; the more data we have to play with the better. Also 24h was an example to show an extreme possibility; I'm a sucker for using extremes in examples to drive home points. Although honestly, waiting 24h for a good close match would be fine for me but I understand our current ADD world culture might not handle it well.

 

[Wicked Lord]

Good news.
After half a year (not sure) since the "new" system gots implemented I can say, it´s way better than before (Mid tier alliance).

Our 10 last results:

74 : 74
74 : 43
57 : 75
75 : 56
72 : 72
43 : 39
49 : 49
67 : 67
67 : 48
75 : 63

well done

 

[BHG_Muet]

I hate to give the standard "coming soon" response. But, for your a) and b), stay tuned. We certainly hear the complaints around bugs and security and have had a fairly long-running investigating on the tech side to discuss solutions for some of the bigger pain points in those areas. We've got some good stuff in the works internally but given DomiNation's history of not addressing these concerns, I'd rather show you than tell you. You'll need to wait a little longer but the first steps are coming soon and I'm excited to share what we've been working on when it is ready.

For c), I'm interested to hear what you would be looking for here? More information per release or more future planning? The 2020 roadmap is hard to share as items shift around quite frequently and we can't guarantee release dates (for example, I have a target release date for Information Age in 2020 but it is too early to commit to that date with you all). If you're looking for more information per release, I agree! Arguably, 8.0 went out a little too quiet. We can do better and you'll hopefully see improvements here as we go into 2020. The first (small) change you can expect is an improvement to our patch notes in the upcoming 8.1 release. Listing "bug fixes" as a line item on the patch notes doesn't tell you anything and we're looking to give actual details in our next release. I know it is a small change but it's at least one I can talk about.

Please share your thoughts on what you want to hear more about. I'd like to compare it to the process changes we've been discussing.

 

[Chadwicke]

Now can assult rally at least have time added to it so it is useful again BHG_Muet the better option is putting it back like it was before at 8 seconds with rate of fire

 

[Chadwicke]

Could you please explain what a top teir alliance make up is to us both at the top end of that spectrum to the lower class of top tier , as well exactly what is a mid teir , leaving people out because they are too big or too small is not fun either

 

[Kanechoigo]

The only thing you need to do is that correct war-weight calculations that’s all. My alliance stopped war since early May, because maxed lvl310 SA was sitting behind lvl250 CWA. Exact situation happened 2years ago, you guys never learn?

 

[Chadwicke]

Yes with war weight broken yet again this will never work BHG_Muet plus if this seach we are on is the new Normal at 12 hours and counting you need to double event building duration as this is cutting by 1/2 the wars a building will last . I can't figure out why war had to change in the first place sandbags was not a big deal compared to every war since the changes . And if im waiting now 12 plus hours to get slaughtered because war weight is off thwn there's is no point to war and if no point to war no point to cards no point to rush with crowns. You really want to fix the game fix what caused this mess the hacked accounts out there ruining the squew

 

[oddin]

is war weight not calculated correctly again?

 

[BHG_Muet]

War weights were updated with the release of Digital Age to include Digital Age (and Space Age was fixed a few patches ago). Are you all seeing Digital Age or Space Age accounts incorrectly weighted?

When I mention "top-tier" Alliances, I'm generally talking about the small percentage of Alliances at the high end of the matchmaking pool. With regards to the exact makeup, it would be those Alliances with high Glory and high Offensive/Defensive ratings.

 

[BHG_Muet]

12 hours shouldn't be possible as the system would have forced your Alliance into a match well before that. It's possible that your Alliance stopped trying to match but got stuck in the queue. I'll investigate today.

If you're ever in queue that long, cancel and re-enter the queue as something went wrong.

 

[Xabar]

BHG_Muet ,
For mid-tier Alliances, the matchmakings are not too good, as you say. The main problem is the players Ages distributions between alliances. For example in a 20vs20 WW, Alliance 1 (5 SA, 5 CWA, 5 IA, 5 Clasical), match with Alliance 2 (1 SA, 2 CWA, 2 AA, 4GA, 4 Industrial, 2 Enlighthment, 4 GunPowder, 1 Medieval). The Alliance 1 is always going to win, because the 20 attack of their top10 players, can easily defeat the 20 players of the alliance 2. The players of the Alliance1 attack oponents of an age below.
The Offensive/Defensive ratings, seams not fix the sandbags problem

 

[BHG_Muet]

Yes, correct. The system doesn't remove the ability to sandbag but it does attempt to prioritize Alliances with similar compositions (something it didn't do prior to the update earlier in the year). So, even if your example Alliances had even Glory and even ratings, it would not pair these two Alliances at first and it would try to find other opponents with similar compositions.

That said, it's always possible that there just aren't other Alliances in the queue with a composition that is similar at any given time. So, eventually it would prioritize finding an opponent over preventing sandbagging and pair the two example Alliances you mentioned. But, that shouldn't be a fast pairing and should only occur if the system can't find a more similar composition.

 

[oddin]

BHG_Muet you are responding to our comments and some are NOT SHOWN. But if I go toyour profile, I can see parts of your responses. Check it out cause we would like to read your comments!

 

[SomeRandomPlayer]

Nice find; several missing Muet comments here. Also, click "See More" there to see his full responses. Folks would appreciate to see them here