WWII Logs Homepage
Forum Home Forum Home > OSS Application > General
  New Posts New Posts RSS Feed - The Sweet Spot?
  FAQ FAQ  Forum Search   Events   Register Register  Login Login

Welcome to the WWII Logs forum.

The Sweet Spot?

 Post Reply Post Reply
Author
Message
*Maj_Fox View Drop Down
VETERAN
VETERAN


Joined: 2017-Jan-15
Location: England
Status: Offline
Points: 121
Post Options Post Options   Thanks (0) Thanks(0)   Quote *Maj_Fox Quote  Post ReplyReply Direct Link To This Post Topic: The Sweet Spot?
    Posted: 2020-Apr-15 at 08:52
Is it possible using the data available to understand the game BR spread and the resulting sweet spots?

What do I mean by that?

If I take a BR 4.0 line up into Ground AB what is the percentage chance that match BR will an up tier or down tier?

Using my Chi-Nu-II data from 31 matches I see the following Match BR Spread.

BR 3.7 : 13%
BR 4.0 : 23%
BR 4.3 : 25%
BR 4.7 : 39%

Not a Sweet Spot as I am in the bootom half of my BR spread 64% of the time. But will this be the same for all my BR 4.0 line ups across the nations?

This sort of data would be very useful over on WT forums.

Any thoughts?

Cheers
Back to Top
Sidiros View Drop Down
Admin Group
Admin Group
Avatar

Joined: 2016-Feb-24
Location: Florida
Status: Offline
Points: 317
Post Options Post Options   Thanks (0) Thanks(0)   Quote Sidiros Quote  Post ReplyReply Direct Link To This Post Posted: 2020-Apr-16 at 19:41
I have played a variety of different ways in Ground RB.    I played with a one vehicle lineup,   a full lineup of vehicles, Air heavy, no Air, as a scout, a variety of nations, as support, and across multiple countries and BRs.   I played Air Arcade extensively in the bombing roll to make cash and I believe my B-25's have over a 80% win rate.

With all my experiences and tinkering around with my matches my belief is that Gaijin has a hidden player rating used for matching that is constrained by the +1 BR spread.    Do very well and you'll be expected to carry the team more, matched with poorer playing players and tiered higher a majority of the time.   Play poorly and you'll start getting matched with better players more often and get tiered better.  

Granted this is at the mercy of player populations in the BRs; however, that's what I experienced when playing.    

Ultimately I found controlling my score (between 1-2k), playing a supporting role, , and playing a lineup that could be successful even when uptiered was ideal.   I had a way more enjoyable experience.

So the sweet spots were a good lineup of vehicles at a given BR that could do well on an uptier (mobile, good guns.)

Not sure if that was helpful...    Just my thoughts on it.
Back to Top
Bruce_R1 View Drop Down
Admin Group
Admin Group


Joined: 2020-Jan-13
Status: Offline
Points: 23
Post Options Post Options   Thanks (0) Thanks(0)   Quote Bruce_R1 Quote  Post ReplyReply Direct Link To This Post Posted: 2020-Apr-17 at 13:30
By definition, differentials in BR tiering would apply to all nations, as we understand the matchmaker today.

I answered this question in the 1.93 Ground RB context using data from this site here:


In that context there were definite gaps where you could reliably count on being uptiered (5.0 being one of the big ones). But yes, I think wwiilogs is probably the only independent data source with the collection methods that could support answering this question right now. For AB, I'd suggest recruiting a few more reporters besides yourself, as there could be geo- or time-based issues that means you're not getting an average sample of matches, before reporting on it. For RB, after a little more data cleanup I was looking at doing something along these lines with the 1.95 data.
Back to Top
LtChambers View Drop Down
Member
Member


Joined: 2020-Apr-16
Status: Offline
Points: 2
Post Options Post Options   Thanks (0) Thanks(0)   Quote LtChambers Quote  Post ReplyReply Direct Link To This Post Posted: 2020-Apr-17 at 21:00
How to download the data for analysis like that? Or is it only your personal missions?
Back to Top
Bruce_R1 View Drop Down
Admin Group
Admin Group


Joined: 2020-Jan-13
Status: Offline
Points: 23
Post Options Post Options   Thanks (0) Thanks(0)   Quote Bruce_R1 Quote  Post ReplyReply Direct Link To This Post Posted: 2020-Apr-18 at 14:29
In my case for that article (that was back in 1.91/1.93) I copied all and downloaded the tables I needed from the site pages and scrubbed them up using Excel. Took a while.

I'd say you need about 1000+ game returns from a range of users, for a given version and play type, before I'd trust it very much, but for ground RB across those two versions I ran the significance calculation and it could be considered a representative sample.

It stands to reason, as I explain in the post, that there would tend to be BR spikes, leading to under/over tier situations around them, every four steps in the ladder, going down from 10.3, as competive tiers crowd out those to either side.  The undertier points will always be in the troughs plus or minus two of those spikes in that model. The game will optimize to that, assuming what we have been told so far about the matchmaking is correct. What's interesting and a little unpredictable without stats like wwiilogs can produce, is where and how that 10.3, 9, 7.7... peak pattern gets interfered with and altered at lower BRs. 
Back to Top
Bruce_R1 View Drop Down
Admin Group
Admin Group


Joined: 2020-Jan-13
Status: Offline
Points: 23
Post Options Post Options   Thanks (0) Thanks(0)   Quote Bruce_R1 Quote  Post ReplyReply Direct Link To This Post Posted: 2020-Apr-18 at 16:04
Here's what that graph looks like with 1.95 data added in, if you were curious.



Edited by Bruce_R1 - 2020-Apr-18 at 16:04
Back to Top
LtChambers View Drop Down
Member
Member


Joined: 2020-Apr-16
Status: Offline
Points: 2
Post Options Post Options   Thanks (0) Thanks(0)   Quote LtChambers Quote  Post ReplyReply Direct Link To This Post Posted: 2020-Apr-18 at 19:46
So there's no way to download the raw database?

Also is it possible to upload "incomplete" missions? If I die 3 times in the first 3rd of the round (I tend to play too aggressively) I'm not sticking around for the rest of the time so it's a bummer if those rounds can't be counted because that'll severely bias my results.
Back to Top
 Post Reply Post Reply
  Share Topic   

Forum Jump Forum Permissions View Drop Down

Forum Software by Web Wiz Forums® version 11.04
Copyright ©2001-2015 Web Wiz Ltd.

This page was generated in 0.047 seconds.