Jump to content
You need to play a total of 10 battles to post in this section.
BoBaLoT

Another post about Matchmaking! :-)

19 comments in this topic

Recommended Posts

10
[LGNDS]
Member
16 posts
7,848 battles

Okie Dokie folks,

I have been capturing data from my games since 02 Oct this year to get a better understanding (at least in my own mind) of whether the matchmaking was indeed giving me the raw end of the deal or if it was just all in my head. I mean we all do tend to remember the bad more than the good which can impact on our ability to accurately describe what's going on and I didn't want it to be opinion based so I did the only thing I could think of doing and started recording them. So anyway here is my numbers since 02 Oct.

As you can see I have not included my T10 battles as this has little to no value (that I could really see anyway) as you are always going to be top or same tier. I did find it interesting that the numbers @Hapa_Fodder  was talking about on an official WG stream a couple of months ago were absolutely accurate, albeit a tad misleading. He was saying that after ~40 games (I think) or so the matchmaking will only bottom Tier you ~40% of the time. Turns out this is very accurate but what you may not have realised is that that is actually the highest bracket of the 4 variables. That means that if you're bottom tier 40% of the time then only 60% is left to cater for the 3 other tier types (Top, Middle and Same tier), giving each one around 20%.

Essentially, if you're feeling like you're bottom tier 'all the time' then it is probably because you are.

I will of course continue to capture more information. I am hopeful to get more data on Tiers 4, 5, 6 and 7 in particular (my numbers are not really high enough yet to make a solid call for these tiers) but honestly they are just horrible lol. Tier 7 in particular is atrocious with ~50% of my games since recording them being bottom tier. It is quite challenging to get motivated to play against Tier 9's in my Atlanta right. But nevertheless I will continue to gather information and see where it leads. If anything I am hopeful that this will bring light to the comments everyone is making about matchmaking? Personally, I would love to see a change from the two tier up/down system to a one tier up/down system, even if it means waiting for a few minutes. I am confident/hopeful we have the numbers of players to make this viable option.

I would love to hear from others who have captured similar data. It might also be important to note that I started capturing this data when I was already over ~5500 games I think. I wonder if that makes a difference?

Oh I almost forgot, I was also capturing the number of games that I was unfortunate enough to have a double CV. Other than Tier 9 it would seem that ~15-20% of the time you will get double the pleasure. I thought WG said this wasn't going to happen anymore above Tier 4 games?

Enjoy 🙂

Matchmaking stats.JPG

  • Thanks 4

Share this post


Link to post
Share on other sites
3,003
[TLS]
Member
4,368 posts
20,667 battles

This is indeed interesting and does confirm our suspicions where there is a decent sample size.

It would be also good if you can stratify if you were just the 1 or 2 bottom tier ships or bottom tier ships make the majority of the team because it does make a difference at t8/7 as t9/10 get range/reload mods.

Also, you could share how you gather the data so that volunteers can add to your data pool. 

  • Cool 1

Share this post


Link to post
Share on other sites
10
[LGNDS]
Member
16 posts
7,848 battles
4 minutes ago, dejiko_nyo said:

It would be also good if you can stratify if you were just the 1 or 2 bottom tier ships or bottom tier ships make the majority of the team because it does make a difference at t8/7 as t9/10 get range/reload mods.

Also, you could share how you gather the data so that volunteers can add to your data pool. 

Awesome idea, I had not considered that. I will start a new column to capture whether I am majority bottom/top tier as well.

As for capturing the data, I do it manually as I join a game.

Share this post


Link to post
Share on other sites
3,168
[CLAY]
Beta Tester
4,820 posts
21,559 battles

Am I over simplifying it to say T9 is the "best" tier in this regard then?

Share this post


Link to post
Share on other sites
256
[BONED]
Member
600 posts
7,777 battles

Some of the issue is the progression system. 

To grind up lines you need to play a lot more games at each respective tier.

From WOWSW numbers on asia server number of battles by tier are

Overall

snip_20201217162006.png.da90ca615a38e46e770cc4f16cd3cd8f.png

Tech Tree

snip_20201217161934.png.4aced3958eb917c0baf51ffcefa34a54.png

Premium 

snip_20201217162022.png.14f721debd1ea476f5f305e93e384d92.png

It shows how as you go down the tiers how the number of available ships in lower tiers is reduced. MM is at least made easier by the lack of a template of tier numbers (as in WoT) where being bottom tiers is much more common, as are same tier battles. 

Share this post


Link to post
Share on other sites
10
[LGNDS]
Member
16 posts
7,848 battles

So my mate was doing the same thing since 04 Oct this year and I got him to send me the data. His spreadsheet was slightly different and captured a bit more information than I did but I collated it and put it into my spreadsheet to see the similarities / difference. Here is his numbers. He has about the same amount of total battles as I do.The number of games captured for this data is a bit higher than mine.

Matchmaking - Norrie.JPG

Edited by BoBaLoT
  • Cool 1

Share this post


Link to post
Share on other sites
10
[LGNDS]
Member
16 posts
7,848 battles
3 minutes ago, Max_Battle said:

Am I over simplifying it to say T9 is the "best" tier in this regard then?

That's pretty much what I am thinking also. To be honest Tier 9 is often the most enjoyable (because it has a higher chance of being same tier game and that means NO CV 🙂 )

  • Cool 1

Share this post


Link to post
Share on other sites
2,781
Member
5,164 posts
9,111 battles
21 minutes ago, Max_Battle said:

Am I over simplifying it to say T9 is the "best" tier in this regard then?

Probably. I posted my question about T9 battles before in a thread - of why T9 battles are very often being mono-tiered.

One of the reasons are attributed to the changes to MM for T8 ships. You know, MM would put a T8 ship in T10 battles ~40% of the times and T6 ~40% of the times. Since there are more people playing in high tiers, there are more chances to get into T9 mono-tiered battles if you play a T9 ship.

Also, there is no odd-tiered CVs (thanks, CV rework), so mono-tiered T9 battles will never see a CV.

ALSO, T9 premiums have the best credit farming capability of all the tiers. T10 special ships (or T10 ships in general) probably only gain credit a bit better than T7 premiums with premium camos.

Also ALSO,... ... I ran out of things to add.

Edited by Paladinum
  • Cool 2

Share this post


Link to post
Share on other sites
44
[LGNDS]
Community Contributor
32 posts
3,194 battles

So T9 generates the best credit cap, has the best percent for Mid tier and on average still gets the 20% top or at tier. In short yerp its the best haha

  • Cool 2

Share this post


Link to post
Share on other sites
3,168
[CLAY]
Beta Tester
4,820 posts
21,559 battles
20 minutes ago, Paladinum said:

ALSO, T9 premiums have the best credit farming capability of all the tiers.

 

3 minutes ago, DobbyM8 said:

So T9 generates the best credit cap, 

Just the premiums.

Share this post


Link to post
Share on other sites
3,168
[CLAY]
Beta Tester
4,820 posts
21,559 battles
27 minutes ago, Paladinum said:

Also ALSO,... ... I ran out of things to add.

 

  • Haha 1

Share this post


Link to post
Share on other sites
1,850
[151ST]
Member
2,996 posts
11,069 battles
2 hours ago, BoBaLoT said:

Okie Dokie folks,

...

First up, nice study!

I like data, you can only argue with the interpretation... And given that I feel like no one else is being devil's advocate so looks like it's on me...

Your data isn't quite complete.

Because of the way MM works and how it gets wonky with certain queue numbers you are missing time and class information.

For example...

If you queue at a super low server population time, as a T10 you will be top tier (rather than same tier) way more frequently as the MM teies to fill games.

You are also far more likely to be pulled up tiers as a t8 because of that...

This will also lead to more 2v2 CV games...

The class thing is a big deal too as if you are playing a DD or CA you will wait less than a BB and thus it will affect your statistics.

So while I think it's very interesting, I don't think we have the whole story...

 

24 minutes ago, DobbyM8 said:

So T9 generates the best credit cap, has the best percent for Mid tier and on average still gets the 20% top or at tier. In short yerp its the best haha

Yeah, I think we all knew this was the case, it's nice to see some data to back it up however.

 

Edit: I wonder how MM would look if even tiers had same tier coding similar to T10...

TBH we need +/-1 MM

Edited by S4pp3R
  • Cool 2

Share this post


Link to post
Share on other sites
249
[NZAUS]
Beta Tester
361 posts
8,563 battles
36 minutes ago, S4pp3R said:

TBH we need +/-1 MM

Yes ! 

Entire game almost fixed in one go . 

CentaurBeer.jpg

  • Funny 1

Share this post


Link to post
Share on other sites
10
[LGNDS]
Member
16 posts
7,848 battles
39 minutes ago, S4pp3R said:

First up, nice study!

I like data, you can only argue with the interpretation... And given that I feel like no one else is being devil's advocate so looks like it's on me...

Your data isn't quite complete.

Because of the way MM works and how it gets wonky with certain queue numbers you are missing time and class information.

For example...

If you queue at a super low server population time, as a T10 you will be top tier (rather than same tier) way more frequently as the MM teies to fill games.

You are also far more likely to be pulled up tiers as a t8 because of that...

This will also lead to more 2v2 CV games...

The class thing is a big deal too as if you are playing a DD or CA you will wait less than a BB and thus it will affect your statistics.

So while I think it's very interesting, I don't think we have the whole story...

 

Yeah, I think we all knew this was the case, it's nice to see some data to back it up however.

 

Edit: I wonder how MM would look if even tiers had same tier coding similar to T10...

TBH we need +/-1 MM

 

I totally agree mate. And whilst it would have been awesome for me to capture every tiny detail, I didn't want to end up being consumed by capturing so much information that it took away from actually playing the game 🙂  I think it is safe to say that regardless of time of day and class it really should only be one tier up/down. I propose that it would stop at least some of the yolo'ing. I can't tell you how many times I have seen a T7 just go up the guts to end it's misery and get onto the next game because it got into a T9 game lol

Share this post


Link to post
Share on other sites
3,003
[TLS]
Member
4,368 posts
20,667 battles
23 minutes ago, BoBaLoT said:

I can't tell you how many times I have seen a T7 just go up the guts to end it's misery and get onto the next game because it got into a T9 game lol

It wouldn't be a problem if the ship had the tools to match a T9 but the vast majority of ships don't have those tools. But giving a T7 those tools will result in T5s unable to combat T7s effectively. Such a wide MM spread makes balancing a pain.

  • Cool 3

Share this post


Link to post
Share on other sites
765
[SMOKE]
Member
2,215 posts
17,119 battles

+/- 2 , is this already pretty much acknowledged as one of the factor that MM so screw up, made it +/-1 , then we would not had T7 matched against T5 or T8 against T6 , but would WG do that , I doubt it.

Share this post


Link to post
Share on other sites
712
[BLESS]
Member
1,308 posts
11,695 battles

It is interesting to observe how WG engineered changes to the game over the last year or so to address the issue.

For those who don't know, back in the day everyone played TX, so much so that if you took a T8 into randoms, you'd be sitting facing 11 TX ships. T7, on the other hand, was gifted to be top tier almost all the time, so again, if you were stupid enough to play T5 you were shafted.

The main change is the introduction of T9 coal ships, which game players premium (read - credit making) ships to enjoy at T9. T9 is the new T10, T7 is the new T8. MM was also tweaked significantly. Just about all tiers are viable in Randoms now, though T7 is still pretty rough and T9 is still gifted.

  • Cool 1

Share this post


Link to post
Share on other sites
2,781
Member
5,164 posts
9,111 battles
1 hour ago, dejiko_nyo said:

It wouldn't be a problem if the ship had the tools to match a T9 but the vast majority of ships don't have those tools. But giving a T7 those tools will result in T5s unable to combat T7s effectively. Such a wide MM spread makes balancing a pain.

CONCEALMENT MOD. 1 AND UPGRADE SLOT 6!!!

Share this post


Link to post
Share on other sites
1,850
[151ST]
Member
2,996 posts
11,069 battles
2 hours ago, Tagnbag said:

Yes ! 

Entire game almost fixed in one go . 

CentaurBeer.jpg

I think that's the most modest centaur I've seen from you...

Share this post


Link to post
Share on other sites

×