Squad War Matchmaking Data

All Squad Wars related topics. Questions, comments, issues, shout outs, call outs.
User avatar
Sigma582
Admiral
Posts: 600

Re: Squad War Matchmaking Data

Post#31 » Wed Aug 30, 2017 2:31 pm

Midge wrote:Brigade, PLC, Havoc, etc. would literally almost always have to war with each other and nobody else.

Let them fight each other day in day out, it looks less of a problem than 30% BR overmatch (which is almost guarantees victory of the stronger squad unless tako is in play).
As a solution for that DI could add a "repetition factor" which would decrease "matchability" of squads you fought recently so that the algo would, for example, prefer new squad with 7% BS difference over one with 4% difference you recently fought.

Midge wrote:get deeper and grow your squads like they did?

It takes 600 droid-days to upgrade 20 turrets and 4 shields from lvl7 to lvlX. With 3 droids constantly working on defences (while 1 is on dekas and 1 on all other stuff) it will take 7 months to max out defences. In reality it will be around a year on average. Are you suggesting every squad to suffer being underdogs for about a year while its members upgrade their bases?
Image
New home of former Windows IMPERIALS® squad. PM me to join us on Android server.
-
Layout Manager
http://www.swcommander.com/viewtopic.php?f=31&t=19986
-
Defense Tracker
http://www.swcommander.com/viewtopic.php?f=31&t=20403


User avatar
Midge
Admiral
Posts: 702

Re: Squad War Matchmaking Data

Post#32 » Wed Aug 30, 2017 6:25 pm

Sigma582 wrote:
Midge wrote:Brigade, PLC, Havoc, etc. would literally almost always have to war with each other and nobody else.

Let them fight each other day in day out, it looks less of a problem than 30% BR overmatch (which is almost guarantees victory of the stronger squad unless tako is in play).
As a solution for that DI could add a "repetition factor" which would decrease "matchability" of squads you fought recently so that the algo would, for example, prefer new squad with 7% BS difference over one with 4% difference you recently fought.

Midge wrote:get deeper and grow your squads like they did?

It takes 600 droid-days to upgrade 20 turrets and 4 shields from lvl7 to lvlX. With 3 droids constantly working on defences (while 1 is on dekas and 1 on all other stuff) it will take 7 months to max out defences. In reality it will be around a year on average. Are you suggesting every squad to suffer being underdogs for about a year while its members upgrade their bases?


Are you suggesting they shouldn't face big squads unless they are exactly even?

Bitching about matchmaking is ridiculous when you are 54-21. Period. We got stuck against Enclave, Shadow Company, Imperial-Snipers TWICE when we were rolling with 2 10s. Didn't bitch. Got better. Even earned a tie against one of em.

The problem is people think base scores and medals and HQ are all that matter. They don't. If you are winning 2/3rds of your damn wars it is pretty lame to whine about matchmaking.

Try growing your squad past 20? Try joining a more active squad? Join the LINE community and learn how a ton of squads like yours are winning. Though, again, winning two thirds. What are we gonna be satisfied with? Winning 80%? 90? Only fair if you win em all?
Leader of The Mob Syndicate

Home of TheMidnightMob, TheMorningMob and TheMobWantsYou

Level X Rebel : 100k+ Medals : Joined the Rebellion 12/28/2015


User avatar
Pirate
Major General
Posts: 436

Re: Squad War Matchmaking Data

Post#33 » Wed Aug 30, 2017 8:36 pm

Midge wrote:This biggest problem you have there is that you ASSURE rebels and imps both of a 1 1/2 hour wait time. There are a many squads that don't have but a handful of squads within their stratosphere. It HAS to be a NUMBER OF SQUADS algorithim because a % base score one doesn't work. What if Brigade enters and nobody gets with 30% of their base score for the three hours you propose there? Or within 60% within six hours? That is VERY VERY possible with a squad like theirs.

It also does what the main issue for a LARGE gap in the algorithim is ... Brigade, PLC, Havoc, etc. would literally almost always have to war with each other and nobody else.

It has to allow for teams to face much lower teams because there are just too many of them and it is worldwide. You cannot tell a team they are only gonna match teams within 5% of them, then 10%, then 15% etc. when they can see that there are only like 10 teams in the game within that range they would only face the same squad over and over and over.

The matchmaking may be a little too large of a pool, but it makes sense logically. They have to pair teams and Imps are already waiting forever with the current setup. If you are a squad that is rolling with 4 10s and a few 6s ... then get those 6s moving quickly, or you simply are going to be outmatched sometimes.


Then open up matchmaking at +/-15%, that's fine. I'm not advocating that every match has to be super tight (the challenge of attacking up and facing better squads makes players and squads better) but once you start getting north of 25% or 30% difference in basescore, it isn't a war, it's a slaughter. Throwing Obi's squad to the wolves at 36% difference after 10 seconds is dumb. If it had been 2 or 3 hours and there is no better match, fine, have at it, sorry for your luck. But 10 seconds?

Let's say that MM is capped at 25% difference. DB enters with avg BS of 2964. They could face any squad with avg BS of 2223 or above, that's a pretty reasonable pool of squads. That team with 2223 avg BS could face anyone from 1667 avg to 2964 avg. The problem is definitely at the high end but I would think the big dogs would still find matches. Maybe I'm wrong and it would totally F matchmaking times, the current system seems pretty flawed though.

My intention when I started this thread was to try to better understand MM based on actual numbers. I don't know if there is much to understand though. Seems as much RNG as anything else.
ⒹⒶⓇⓀⓈⒾⒹⒺ
Imperial Squad

Windows Survivor


User avatar
Hoops8
Commodore
Posts: 131

Re: Squad War Matchmaking Data

Post#34 » Thu Aug 31, 2017 5:44 am

Midge wrote:
Hoops8 wrote:
Pirate wrote:
Well, I think you win the mismatch prize, that is terrible. I would get it if that was a match after 8 hours of searching but 10 seconds in plain stupid. I guarantee there were plenty of imp squads looking for a match shortly after you guys that had similar total basescore.

I don't know if there is any point to tracking data that has no correlation. I have tracked over 20 of our matches with the input from here as well and all that I have learned is that matchmaking is a joke. I get that squads of all heavy 10s shouldn't have to wait many hours for a good match but they also shouldn't throw the first squad with less than 30% difference in basescore at them.

First 30 minutes of search: +/- 5% basescore
next 30 minutes of search: +/- 10% basescore
next 30 minutes of search: +/- 15% basescore
...etc...

How easy is that?

At least this way you have a chance to bail out if you've been waiting a long time and don't want to draw a big dog. It's really not that hard to solve and better matches are better for business. No brainer, come on Disney



Quality Idea! I doubt you would find anyone who would object to this! We had a stretch of 20 wars recently where we were the underdog in 19/20 wars! All of which we were insta-matched. We still fared well but more even matches would be swell, and more fun for those non-lvl 10s in our squad.


Show me that 20 war stretch and I will rip it to shreds. What makes you the underdog? They had higher base scores? They had higher levels? They were much better war squads? You're sitting at 54-21 ... exactly how unfair is the matchmaking? Will it only be fair if you're 70-5?


Like I said we did well, but we are typically fighting an uphill battle, though recently matchmaking has been better.

I based the term "underdog" on base score of the war parties. I kept track. I am not sure of the dates but I can get that, though I kept track of the wars on my work computer, if you really need this info.

We are 65 - 30 not 54-21.

I am not complaining just stating the fact that we had a smaller base score of war members when compared to our imp counterparts in 19/20 wars during one stretch. Not demanding easy wars, or cupcakes. Read my post again all I want is fair matches, and would be willing to wait for a fair one rather than get matched against a cupcake of a monster??? Is this something you are against???
OMISQUAD and OMISQUADBLACK Always looking for those that enjoy good company and good base crushing.


DarkRebel
Major General
Posts: 423

Re: Squad War Matchmaking Data

Post#35 » Thu Aug 31, 2017 6:09 am

Just a quick 2 cent here:
My IMP squad generally had good matchup. We had some uphill battles, but also some cupcakes, but in general I found the matching was fair, for me.
People who were willing to provide data here were more likely than not, a victim of mismatch. True ?
So you are more likely get data from the folks that experienced mismatch, and not for the entire SWC player base. So be careful about drawing conclusion from incomplete data set.
Squad war is not won or lost just because of base rating. It takes skills, team work, and being active to win, and even if you are outmatched, better coordination and outposts boost can compensate the base rating gap.
There is a trade off between tight match and wait time. The key is to find the right balance, but balance is up to everyone interpretation of what they think is the "right" balance. Some prefers tight match, some prefers quick match.


User avatar
Midge
Admiral
Posts: 702

Re: Squad War Matchmaking Data

Post#36 » Thu Aug 31, 2017 7:07 am

Hoops8 wrote:
Midge wrote:
Hoops8 wrote:

Quality Idea! I doubt you would find anyone who would object to this! We had a stretch of 20 wars recently where we were the underdog in 19/20 wars! All of which we were insta-matched. We still fared well but more even matches would be swell, and more fun for those non-lvl 10s in our squad.


Show me that 20 war stretch and I will rip it to shreds. What makes you the underdog? They had higher base scores? They had higher levels? They were much better war squads? You're sitting at 54-21 ... exactly how unfair is the matchmaking? Will it only be fair if you're 70-5?


Like I said we did well, but we are typically fighting an uphill battle, though recently matchmaking has been better.

I based the term "underdog" on base score of the war parties. I kept track. I am not sure of the dates but I can get that, though I kept track of the wars on my work computer, if you really need this info.

We are 65 - 30 not 54-21.

I am not complaining just stating the fact that we had a smaller base score of war members when compared to our imp counterparts in 19/20 wars during one stretch. Not demanding easy wars, or cupcakes. Read my post again all I want is fair matches, and would be willing to wait for a fair one rather than get matched against a cupcake of a monster??? Is this something you are against???


One of the two squads you claim is 54-21 and the other one is 65-30. EITHER WAY ... both are doing pretty damn well to be complaining. And ... as I said ... calling yourself "underdogs" based solely on base score is kind of ridiculous.

YES ... sometimes you are going to face teams that are probably out of your league ... the Browns have to go to New England and face the Patriots sometimes ... just how it goes. Hell, they are allowing the Jets to continue playing football.

Fair matches happen sometimes, easy matches happen sometimes and difficult matches happen sometimes. That SHOULD be how it is because it means squads are getting matched in a timely fashion. There is NO logical way for them to make it "fair" EVERY SINGLE TIME and keep a regular pool flowing when there are in excess of 15,000 active squads and players moving from squad to squad to squad on a regular basis.

I find it hard to take any complaint about matchmaking seriously from a guy in two squads that are doing better than 70% in wars. Seriously? You expect every single match to be fair? How many squads have you jumped into and hung out in. I assure you not enough. I have visited more than 500 (and that is not an exaggeration) of rebel squads and I assure you that there is NO WAY TO NUMERICALLY measure how good a squad is in WAR based on medals, base score or HQ level. You could MAYBE measure it with war results, which I do think Disney should incorporate somehow. BUT without using that, there is no way to measure it.

YOU ARE NOT AN UNDERDOG JUST BECAUSE THE OTHER TEAM HAS HIGHER BASE SCORES. It is a ridiculous assertion. What if half of that squad speaks one language and the rest of them all speak three other ones? There are a ton of those out there. So, that very casual squad should be punished with only facing teams the same base score as them even though they will get totally wrecked every time?

Anyways ... all that aside ... Pirate, I got no beef with your love of numbers and wanting to dig in and see what is out there. I don't think there really is a way to get enough of a sample size to be relevant here, but it is cool to see some of those tables.
Leader of The Mob Syndicate

Home of TheMidnightMob, TheMorningMob and TheMobWantsYou

Level X Rebel : 100k+ Medals : Joined the Rebellion 12/28/2015


DarkRebel
Major General
Posts: 423

Re: Squad War Matchmaking Data

Post#37 » Thu Aug 31, 2017 7:28 am

Pirate wrote:
ObiWanKenobi2016 wrote:...


Well, I think you win the mismatch prize, that is terrible. I would get it if that was a match after 8 hours of searching but 10 seconds in plain stupid. I guarantee there were plenty of imp squads looking for a match shortly after you guys that had similar total basescore.

I don't know if there is any point to tracking data that has no correlation. I have tracked over 20 of our matches with the input from here as well and all that I have learned is that matchmaking is a joke. I get that squads of all heavy 10s shouldn't have to wait many hours for a good match but they also shouldn't throw the first squad with less than 30% difference in basescore at them.

First 30 minutes of search: +/- 5% basescore
next 30 minutes of search: +/- 10% basescore
next 30 minutes of search: +/- 15% basescore
...etc...

How easy is that?

At least this way you have a chance to bail out if you've been waiting a long time and don't want to draw a big dog. It's really not that hard to solve and better matches are better for business. No brainer, come on Disney

Not a bad idea. But matching is to match 2 squads so you need to consider both squads' wait time.

For example,
Dark-Bridgade has been in the queue for nearly 2 hours and had no high level rebel squad to match up with.
A new rebel squad just entered into match making with lower base rating than DB.
Do you match DB with this low level rebel squad, or you make DB wait longer ?

According to your formula above, DB should be matched as they had waited long enough. But then it's not fair for the rebel squad.

As mismatch will be difficult to avoid, how about compensating the lower level squad (not necessarily inferior) with the extra boost for the outposts. i.e. A squad whose average base rating is 20% worse than the opponent will enjoy an extra +20% boost on any of the outpost in play.


User avatar
Midge
Admiral
Posts: 702

Re: Squad War Matchmaking Data

Post#38 » Thu Aug 31, 2017 8:46 am

DarkRebel wrote:
Pirate wrote:
ObiWanKenobi2016 wrote:...


Well, I think you win the mismatch prize, that is terrible. I would get it if that was a match after 8 hours of searching but 10 seconds in plain stupid. I guarantee there were plenty of imp squads looking for a match shortly after you guys that had similar total basescore.

I don't know if there is any point to tracking data that has no correlation. I have tracked over 20 of our matches with the input from here as well and all that I have learned is that matchmaking is a joke. I get that squads of all heavy 10s shouldn't have to wait many hours for a good match but they also shouldn't throw the first squad with less than 30% difference in basescore at them.

First 30 minutes of search: +/- 5% basescore
next 30 minutes of search: +/- 10% basescore
next 30 minutes of search: +/- 15% basescore
...etc...

How easy is that?

At least this way you have a chance to bail out if you've been waiting a long time and don't want to draw a big dog. It's really not that hard to solve and better matches are better for business. No brainer, come on Disney

Not a bad idea. But matching is to match 2 squads so you need to consider both squads' wait time.

For example,
Dark-Bridgade has been in the queue for nearly 2 hours and had no high level rebel squad to match up with.
A new rebel squad just entered into match making with lower base rating than DB.
Do you match DB with this low level rebel squad, or you make DB wait longer ?

According to your formula above, DB should be matched as they had waited long enough. But then it's not fair for the rebel squad.

As mismatch will be difficult to avoid, how about compensating the lower level squad (not necessarily inferior) with the extra boost for the outposts. i.e. A squad whose average base rating is 20% worse than the opponent will enjoy an extra +20% boost on any of the outpost in play.


Yeah ... that is what we need ... we need Tako to be 20% stronger!
Leader of The Mob Syndicate

Home of TheMidnightMob, TheMorningMob and TheMobWantsYou

Level X Rebel : 100k+ Medals : Joined the Rebellion 12/28/2015


DarkRebel
Major General
Posts: 423

Re: Squad War Matchmaking Data

Post#39 » Thu Aug 31, 2017 8:57 am

That's a different problem, LOL.
Yeah, Tako is a bit OP. I would cut that down by at least half of where it is today.


User avatar
Sigma582
Admiral
Posts: 600

Re: Squad War Matchmaking Data

Post#40 » Thu Aug 31, 2017 9:15 am

Does this look like an underdog?
Image

We are not stupid, we have lots of war experience, but when you face 10 lvlX bases, some maxed, some with lvl9 defences while having mostly lvl8-9 team you just technically can't get enough points out of them without tako. No matter how hard you try if your troops are just too weak to get 3 stars and defences are too weak to withstand lvlX attacks.

This I call mismatch and this makes no sense for either side.
Image
New home of former Windows IMPERIALS® squad. PM me to join us on Android server.
-
Layout Manager
http://www.swcommander.com/viewtopic.php?f=31&t=19986
-
Defense Tracker
http://www.swcommander.com/viewtopic.php?f=31&t=20403


Return to “Squad Wars Discussions”

Who is online

Users browsing this forum: CommonCrawl [Bot] and 1 guest