Jump to content
Sign in to follow this  
You need to play a total of 20 battles to post in this section.
Slimeball91

Match Making data for 0.10.5.

5 comments in this topic

Recommended Posts

4,010
[PVE]
Members
7,465 posts

If you're reading this then you are probably already aware that I've been tracking team compositions for 0.10.5.  This patch has removed the 2/7 "bug", and it has also brought some other changes.  So what are those changes?  As you'll likely already aware, team compositions can now be any mix of humans and bots, but the bigger change has been the change in the limit on the number of BBs, and DDs in a match.  That number seemed to be capped at 4 in the past, it has gone up to 5 now.  The talk around here says there was never any limit on the number of cruisers.  That could be, but my experience is they generally stayed within the same 4 limit that BBs and DDs have (I'm aware there were exceptions).  My data does support the idea that cruisers don't seem to have the same rules to limit them as the other ships (or the MM rules are the same and cruiser population just overrides the rules more often).

Here's the details on the data collected.  100 co-op games played, all played solo, NA server.  All games were with T8-T10 ships, and at normal hours, mostly between 7pm-12am EST (no late night/early morning games).  I'm not sure it has any bearing, but the games played were roughly a 50/50 mix of tech tree and premium BBs, cruisers and DDs (no games in CVs).  The data collected was for team composition: number of human players, and the number of ship type for each match.  I did not collect data on XP/credit earning, top/bottom tier MM, or WR.  This is just about collecting data so you have a better idea of what you might expect to see in your matches as far as bots and ship types.

The first piece of data, the percentage of games with full human teams: 68 games = 68% full human teams.  The remaining 32 games all had some mix of human players and bots.

Here are the number of games/percentage for each human/bot combination.

9 Human Players/0 Bots 68 68%
8 Human Players/1 Bot 4 4%
7 Human Players/2 Bots 5 5%
6 Human Players/3 Bots 4 4%
5 Human Players/4 Bots 7 7%
4 Human Players/5 Bots 5 5%
3 Human Players/6 Bots 4 4%
2 Human Players/7 Bots 1 1%
1 Human Player/8 Bots 2 2%

As you can see the mix of humans/bots is pretty varied, with no clear bias like we had with the 2/7 bug.  With a larger sample we might see a more clear trend with the 5/4 matches.  Now that I've mentioned sample size, I don't want to pretend 100 games is enough to definitively say how the new MM works in this patch.  I do think its a large enough sample to have some idea of what you might expect if you play at these tiers at normal hours, and it is certainly enough to have a conversation.  You're welcome to leave your thoughts and share your experiences. 

I'm going to post the data on ship types in another post so this one doesn't get too cluttered.

Edited by Slimeball91
  • Cool 1
  • Thanks 2

Share this post


Link to post
Share on other sites
27
[WOLFX]
Members
96 posts
19,311 battles
26 minutes ago, Slimeball91 said:

If you're reading this then you are probably already aware that I've been tracking team compositions for 0.10.5.  This patch has removed the 2/7 "bug", and it has also brought some other changes.  So what are those changes?  As you'll be aware, team compositions can now be a any mix of humans and bots, but the bigger change has been the change in the limit on the number of BBs, and DDs in a match.  That number seemed to be capped at 4 in the past, it has now gone up to 5 now.  The talk around here says there was never any limit on the number of cruisers.  That could be, but my experience is they generally stayed within the same 4 limit that BBs and DDs have (I'm aware there were exceptions).  My data does support the idea that cruisers don't seem to have the same rules to limit them as the other ships (or the MM rules are the same and cruiser population just overrides the rules more often).

Here's the details on the data collected.  100 co-op games played, all played solo, NA server.  All games were with T8-T10 ships, and at normal hours, mostly between 7pm-12am EST (no late night/early morning games).  I'm not sure it has any bearing, but the games played were roughly a 50/50 mix of tech tree and premium BBs, cruisers and DDs (no games in CVs).  The data collected was for team composition: number of human players, and the number of ship type for each match.  I did not collect data on XP/credit earning, top/bottom tier MM, or WR.  This is just about collecting data so you might have a better idea of what you might expect to see in your matches as far as bots and ship types.

The first piece of data, the percentage of games with full human teams: 68 games = 68% full human teams.  The remaining 32 games all had some mix of human players and bots.

Here are the number of games/percentage for each human/bot combination.

9 Human Players/0 Bots 68 68%
8 Human Players/1 Bot 4 4%
7 Human Players/2 Bots 5 5%
6 Human Players/3 Bots 4 4%
5 Human Players/4 Bots 7 7%
4 Human Players/5 Bots 5 5%
3 Human Players/6 Bots 4 4%
2 Human Players/7 Bots 1 1%
1 Human Player/8 Bots 2 2%

As you can see the mix of humans/bots is pretty varied, with no clear bias like we had with the 2/7 bug.  With a larger sample we might see a more clear trend with the 5/4 matches.  Now that I've mentioned sample size, I don't want to pretend 100 games is enough to definitively say how the new MM works in this patch.  I do think its a large enough sample to have some idea of what you might expect if you play at these tiers at normal hours, and it is certainly enough to have a conversation.  You're welcome to leave your thoughts and share your experiences. 

I'm going to post the data on ship types in another post so this one doesn't get too cluttered.

Thanks for the research.I want my 2/7s back!But not with all my ships.If they made a 2 human coop where we could knowahead of time and pick our ships that would be great!Well just dreaming.

  • Cool 2

Share this post


Link to post
Share on other sites
4,010
[PVE]
Members
7,465 posts

In collecting my data I also kept track of each ship type played so we can have some idea of what the population is for each type, and how often we get a match heavy with one ship type.  Up first, the overall population.  In 100 games that means there will be 900 total ships.  I did not track the tiers, the ships could be any tier within the T8-T10 MM spread.  The data is arranged in rows (left to right), ship type/number out of 900/percentage of the total.  For example, there were 33 CVs, that makes up 3.67% of the 900 ships in the 100 games played.

CVs 33 3.67%
BBs 295 32.78%
CAs 311 34.56%
DDs 261 29.00%

This was a little surprising to me.  With cruisers arguably being the hardest ship type to play in co-op I find it surprising that it is the most played.  I thought BBs would have been the most played.  This data isn't the best since I didn't track if ships were played by humans or bots.  While unlikely, it is possible the bots could have a bias for some ship type that could skew the data.  In other words, if bots favored playing cruisers then that could help explain why they are the most played.

The other big change in this patch was the increase to 5 BBs and DDs per match.  Here are the numbers for 5 of each ship type in a match.  Data in rows (left to right) ship type/number of occurrences/percentage out of 100 games played.

5 BBs 7 7%
5 CAs 10 10%
5 DDs 7 7%

You can see cruisers are the most likely to have 5 per match, with BBs and DDs being equal.  Things get a little more interesting when we combine games with 5 (or more) and games with 4.  The point of this table is to see how common it is to be in a match heavy in any ship type.  This table below shows the number of games with 4 or more of each ship type.

4 or more BBs 32 32%
4 or more CAs 39 39%
4 or more DDs 20 20%

I wouldn't have thought there would be so many cruiser heavy matches, and so few DD heavy.  What is also really interesting is that 91% of all matches were heavy in one ship type or another.  One more point of interest, there were no games with more than 5 BBs or DDs, but there were 4 games with 6 cruisers, and 1 game with 7.  My takeaway from this, MM could probably use some work on better distributing ship types.

Edited by Slimeball91

Share this post


Link to post
Share on other sites
87
[WOLF9]
Beta Testers
510 posts
21,776 battles
3 hours ago, Slimeball91 said:

 For example, there were 33 CVs, that makes up 3.67% of the 900 ships in the 100 games played.

Small samples like you said,  I still would expect a little higher for CVs than the 33% of matches you have here.  Based on previous stuff I recorded with very small(smaller than yours) and larger sample numbers.   They got as low as 42% and high of 55%.

 

Makes me think the rocket change might've turned off the fair weather or casual CV players.  And for those who regularly play them,  maybe it's the last straw and they aren't peeved off by any perceived performance losses but on the principle of it,  being punched in the face again.  Loss of options, and CVs are unrewarding, dull, handicapped in Co-op enough as is.

 

Would be interesting and maybe fun to help you out.    But GPU is on its death throes.   Half the time it hard locks on the loading screen.  It's a blessing if I can string some games together. 

 

Anyways,  good stuff there!

Share this post


Link to post
Share on other sites
4,010
[PVE]
Members
7,465 posts
1 hour ago, LabradorRetriever said:

Small samples like you said,  I still would expect a little higher for CVs than the 33% of matches you have here.  Based on previous stuff I recorded with very small(smaller than yours) and larger sample numbers.   They got as low as 42% and high of 55%.

I'm pretty sure I know why CVs games are so low.  In the past 8/1 games where the lone bot was a CV were semi-common in the past.  That seems to be gone now. 

Share this post


Link to post
Share on other sites
Sign in to follow this  

  • Recently Browsing   0 members

    No registered users viewing this page.

×