Jump to content
You need to play a total of 5 battles to post in this section.
Lert

Lert's tier 5 experiment

260 comments in this topic

Recommended Posts

27,652
[HINON]
Supertester
21,729 posts
15,511 battles

Starting tomorrow, I will play 100 tier 5 matches and record how many times I was top tier, how many times I was middle tier in a tier 6 match and how many times I was bottom tier in a tier 7 match.

 

"What ships will you be driving?"

 

Texas, Gremyashchy, kamikaze, Murmansk, various ARP Kongos - a selection of premium ships I have at the tier. Whatever tier 5 I have in my port that I am in the mood to drive.

 

"What are you hoping to prove?"

 

Well, first of all, 100 matches is way too low a sample size to 'prove' anything. Whenever a tier 5 or 6 ships is discussed though there will invariably be a comment about sucky MM and 'it'll be bottom tier 100% of the time'. I've seen numbers thrown around between 40% and 100% with those claims. Personally, I'm just curious what the actual ratio is, and set myself this experiment.

 

"But if it's not going to prove anything, what's the use?"

 

To settle my own curiosity. To provide a data set people can look at. 100 matches won't prove anything, but it'll show a trend, with a certain margin of error.

 

"Why tier 5 and not tier 6?"

 

I have more ships at tier 5 that I enjoy driving than I do at tier 6, and collecting a dataset of 100 matches is already enough work, don't you think? I might do a similar T6 experiment after this, I don't know yet.

 

"How long will this take?"

 

I don't know. A while. Between a week and a month. I have no idea.

 

"What will be the format for your data points? Will you post replays? Team screenshots?"

 

No. I'm just going to update each day with how many battles I played in tier 5 ships, and what the tiering was.

 

"But without proof, this experiment isn't scientific!"

 

Im not a scientist. This isn't a scientific research project. This is just me collecting datapoints.

 

Any other questions? Feel free to ask them. Anyone want to join in on this experiment? I'd welcome the data.

  • Cool 27

Share this post


Link to post
Share on other sites
1,930
[-K-]
-Members-, Members, Supertester, WoWS Community Contributors
5,553 posts
18,010 battles

Sounds like a fun idea.  Alternately, makes me wonder if I can go back to my last 100 tier 5 replays and see...

Share this post


Link to post
Share on other sites
1,422
Beta Testers
4,496 posts
9,954 battles

If your doing this, i would suggest noting the current population when you hit battle.

 

Different times of days/population can show a difference.

Also if in a division can matter also. 

Edited by Spyde
  • Cool 4

Share this post


Link to post
Share on other sites
6,805
[SYN]
Alpha Tester, Beta Testers
30,523 posts
5,372 battles

Have fun Lert. Curious to see the results. If you want a wing man, I'll join ya.

Share this post


Link to post
Share on other sites
3,371
Members
5,202 posts
9,487 battles

Would be even more statistically intriguing if others on here joined your endeavor. Get ten people who track 100 games and things start to become meaningful. 

 

Edited by Kombat_W0MBAT
  • Cool 1

Share this post


Link to post
Share on other sites
27,652
[HINON]
Supertester
21,729 posts
15,511 battles

Different times of days/population can show a difference.

Also if in a division can matter also. 

 

Those are good points. These will all be solo battles, but I will note server population when entering the queueueueueue.

 

Would be even more statistically intriguing if others on here joined your endeavour. Get ten people who track 100 games and things start to become meaningful. 

 

I asked Mouse if she would, and you're welcome to join in as well.

Share this post


Link to post
Share on other sites
Members
1,819 posts

 

Those are good points. These will all be solo battles, but I will note server population when entering the queueueueueue.

 

 

I asked Mouse if she would, and you're welcome to join in as well.

 

Will you be tracking your wins? 

Share this post


Link to post
Share on other sites
27,652
[HINON]
Supertester
21,729 posts
15,511 battles

Will you be tracking your wins? 

 

That falls outside the scope of the experiment, but I see no harm in tracking that too.

Share this post


Link to post
Share on other sites
3,501
[NDA]
Supertester
5,272 posts
15,897 battles

Lert The Sinksalot fleet salutes your bravery in this endeavor.

 m40PPXz.gifH1rhjxh.gif

 

 

  • Cool 1

Share this post


Link to post
Share on other sites
2,755
[V_KNG]
Beta Testers
11,063 posts

I second the notion of you tracking the server population when you click the battle button. I suspect too you should return to port after every battle so that you can note the population. 

 

You might also want to play half non premium, half premium. Oh... and no gold rounds. :red_button::hiding:

Share this post


Link to post
Share on other sites
558
[SOV]
Beta Testers, In AlfaTesters
1,682 posts
6,777 battles

What about tracking divisions you see? Or more specifically, divisions with a mixed tier (5&6, 6&7, etc). 

Share this post


Link to post
Share on other sites
117
[TAFY3]
Beta Testers
450 posts
9,172 battles

I haven't noticed any problems with tier five, or six MM.  Even if there are tier seven ships in the game bottom tier can still be effective.  If it bothers people so much then play tier four and wait for favorable MM.

Share this post


Link to post
Share on other sites
334
[5IN]
Supertester
2,050 posts
9,675 battles

This is very interesting. I intend to follow your progress as Im curious to see how it turns out.

 

Personally, the MM "issues" dont really bother me as much as it seems to bother many others. Not that I particularly enjoy going up against T7s in a T5 frequently (it seems), but you never know how a match will turn out, and sometimes a plucky "little" T5 can turn the tide and help bring a win. ;)

 

B

Share this post


Link to post
Share on other sites
27,652
[HINON]
Supertester
21,729 posts
15,511 battles

You might also want to play half non premium, half premium.

 

I'm not going to re-buy ships and train captains for this. I'll be using prems, and use this as an opportunity to farm a lot of credits.

 

What about tracking divisions you see? Or more specifically, divisions with a mixed tier (5&6, 6&7, etc). 

 

Yeah, not gonna do that. Don't want to collect too much data per match, this is supposed to be a quick and dirty, simple run where I record the tier spread of each match. You're welcome to though.

Share this post


Link to post
Share on other sites
22
[5D]
Members
135 posts
8,416 battles

if you can while your doing this can you do a tonnage/HP record im doing my own experiment :)

 

Share this post


Link to post
Share on other sites
788
[OO7]
Supertester
2,179 posts
9,480 battles

Why are you going to willingly torture yourself by playing tier V? The only ship there that's reasonable to play on a regular basis is the Gremmy, well because its the Gremmy. Ships like Texas and Murmansk are two good ships plagued by terrible MM....

Share this post


Link to post
Share on other sites
-Members-
6,958 posts
9,224 battles

Since I'm currently grinding at tier 5 right now I conduct my own experiment, as well. Don't know when I'll actually get to it, though. Some weekend days I play 25-50 matches, while other days I barely play any.

Share this post


Link to post
Share on other sites
108
[BAKED]
Members
918 posts

If you play the mid tiers enough you can already do this with replays if you have them enabled. It was posted to reddit a while ago. It reads through your replays and outputs what percentage you were top top-1 top-2 at each tier.

 

https://jsfiddle.net/771nnk0e/41/

 

Edited by Dodgy_Cookies
  • Cool 2

Share this post


Link to post
Share on other sites
Members
1,091 posts
4,681 battles

I think you could be wrong on one point: 100 matches MIGHT be a significant sample. 

 

For example, if you come away with a result that most people would consider an outlier, like say 80% or even 90% tier 7 battles, then you could definitely draw some conclusions from that. Remember, matchmaking is not really random. 

 

I agree with Spyde that server population is interesting. One would tend to expect a higher rate of bottom tier results when the server population is lower. 

Edited by DingBat

Share this post


Link to post
Share on other sites

  • Recently Browsing   0 members

    No registered users viewing this page.

×