Login

russian armor

The Ladder: 1v1 Matchmaking

5 Jul 2016, 07:06 AM
#2
avatar of pigsoup
Patrion 14

Posts: 4301 | Subs: 2

did i already say... MIND BLOWN?
5 Jul 2016, 07:39 AM
#3
avatar of momo4sho
Senior Caster Badge
Donator 22

Posts: 466 | Subs: 1

Just gonna post what i said in staff for relevance.

Great work mate as always, but I don't really think stats can alone dictate the faction match ups accurately.

Your stats are much more accurate than the previous one I took a look at, I dont remember the exact source but it just had a blanket win rate across the top 200 ranks [ which means almost nothing imo ]. As your data suggests, OKW is slightly better against SOV than OST, which I find to fall quite flat on its face in reality.

At the end of the day, the stats don't represent all the numerous cheesy strategies that factions can choose to use.

For example, if all 1v1 Soviet players adopted a super abusive strat vs OKW that had a high success rate, the data would be heavily skewed against OKW, but since there is no such precedent we see more tame numbers.

In my opinion, at the end of the day, the best numbers to look at would be high level tournament games that also factored in the map and spawns of each player [ as your previous work touched upon]. The Maps dictate the results of a game just as much, if not more than the actual faction matchup itself Automatch stats can't even come close to giving us the full picture of the state of the game.

Cheers
5 Jul 2016, 07:56 AM
#4
avatar of MoLarr

Posts: 17 | Subs: 1

Need to put the date that this data was captured, at the very top, and what patch.

Great analysis, but as stated above doesn't account for other factors like what map, state of cheese present in the game. Could be useful as a litmus test for the state of balance, but the game will need to be in a state of stability I would think.







5 Jul 2016, 08:19 AM
#5
avatar of Siphon X.
Senior Editor Badge

Posts: 1138 | Subs: 2

jump backJump back to quoted post5 Jul 2016, 07:56 AMMoLarr
Need to put the date that this data was captured, at the very top, and what patch.


Added.

Just gonna post what i said in staff for relevance.

Great work mate as always, but I don't really think stats can alone dictate the faction match ups accurately.

In my opinion, at the end of the day, the best numbers to look at would be high level tournament games that also factored in the map and spawns of each player [ as your previous work touched upon]. The Maps dictate the results of a game just as much, if not more than the actual faction matchup itself Automatch stats can't even come close to giving us the full picture of the state of the game.



Yeah, I agree mostly.

The main point of the post is actually the matchmaking, not the last section with the win percentages at the top. I moved some stuff around and reformulated some paragraphs to deemphasize that part, compared with earlier versions.

However, one of the main criticisms of coh2charts is that it doesn't only include games vs. other top players but against all players. Since I actually have the option to factor that in, I figured I'd have to include that due to public demand.

So, I tried to show that coming up with numbers there is actually quite tricky by showing stuff like error estimates and different ways to compute the percentages...

Regarding maps:
The game history also still contains the information about which map was played and I will definitely do a post about that as well (also including the other modes). However, it seems like the starting position is not in the history (at least I couldn't find it), which will probably seriously dilute the results...

Regarding ladder vs. tournaments: I guess you mean that in a tournament setting people will use the cheesiest of cheesy strategies to win, while they might be more willing to derp around on the ladder, try out different things or are willing to quit if confronted with an unfun strategy? Yeah, likely.

Problem with tournies is that you rarely will have enough matches between top players to make any meaningful statistic. The ESL series was great in that regard. But even there: Ok, it's a bit tricky to say due to the EU/NA divide, but VonIvans SOV/Partisan/Maxim-spam strategy seemed to be overall the most effective, yet he was more or less the only player using it, at least consistently...
5 Jul 2016, 08:38 AM
#6
avatar of NEVEC

Posts: 708 | Subs: 1

A lot of numbers that mean nothing for me.
5 Jul 2016, 08:38 AM
#7
avatar of momo4sho
Senior Caster Badge
Donator 22

Posts: 466 | Subs: 1

Yeah I agree about the tiny sample size for high level tournament games, forgot to include that. Just guess it comes down to what resources we have available to dissect. Just wanted to throw that original text out there to deliver some more context for the stats.

Cheers.
5 Jul 2016, 11:04 AM
#8
avatar of __deleted__

Posts: 4314 | Subs: 7

5 Jul 2016, 11:26 AM
#9
avatar of vasa1719

Posts: 2635 | Subs: 4

Permanently Banned
Greate work man.
+1 to Momo too.
5 Jul 2016, 12:35 PM
#10
avatar of RitaBrush
Lead Artist Badge

Posts: 810 | Subs: 2

Wow, great statistics again! And pretty nice artwork too! You are really one man army :D
5 Jul 2016, 14:44 PM
#11
avatar of Swift

Posts: 2723 | Subs: 1

Good stuff, I wish I really grasped it all :]
6 Jul 2016, 02:17 AM
#12
avatar of kavil

Posts: 12

Thank you mate. good insight on the background stuff.
6 Jul 2016, 09:32 AM
#13
avatar of Tittendachs

Posts: 115

First of all great work mate!
I appreciate your effort very much!


The ladder range in which a player has to be found in order to create an even game is very narrow, typically less than one level.




Most games happen outside of that range, as the range is often more narrow than that in which 33% of the games happen. In other words, the outcome of more than two-thirds of games are strongly affected by the significant skill difference of the two players.



In my experience this is very true! It's the most interesting part of your analysis for me. The search times have been increased significantly lately (maybee 2 months ago?) in order to get more evenly matched games. I had the feeling that it did help slightly. Considering only 1/3 of the games are matched evenly with the current search times it would have been even worse before that.
I know the problem is the playerbase playing 1v1 games (83% of all players DON'T play 1v1) and not the search algorithm
BUT
it is very frustrating to loose 10 matches in a row because of the matchmaking (on workday evenings) and afterwards win 10 matches in a row when playing at the weekend when a lot more casual players are online...
6 Jul 2016, 20:59 PM
#14
avatar of empirescurropt

Posts: 708 | Subs: 1

Siphon X. great work well done it the graphics really a "good read" !!
Thank you for the hard work congrats( keep it coming !!)
7 Jul 2016, 18:05 PM
#15
avatar of borysiatko

Posts: 2

Impressive. Thanks!
16 Jul 2016, 04:12 AM
#16
avatar of Cyanara

Posts: 769 | Subs: 1

Huh. I'm quite surprised that the top 5 ranks are absolute. I would have thought the whole thing would be on a bell curve. Very interesting, thanks.
20 Jul 2016, 22:34 PM
#17
avatar of Goldora
Donator 11

Posts: 22

This is impressive! I wish we could have an interview of whoever at Relic is in charge of matchmaking too.
Then the analysis would be complete :)
29 Jul 2016, 13:03 PM
#18
avatar of Cyanara

Posts: 769 | Subs: 1

Came here to find out if it included the fact that ranks 14 and 15 don't exist for some teams. Was not disappointed :p Sure makes the whole system a bit crappy though.
1 user is browsing this thread: 1 guest

Ladders Top 10

  • #
    Steam Alias
    W
    L
    %
    Streak
Data provided by Relic Relic Entertainment

Replay highlight

VS
  • U.S. Forces flag cblanco ★
  • The British Forces flag 보드카 중대
  • Oberkommando West flag VonManteuffel
  • Ostheer flag Heartless Jäger
uploaded by XXxxHeartlessxxXX

Board Info

600 users are online: 600 guests
0 post in the last 24h
13 posts in the last week
32 posts in the last month
Registered members: 50080
Welcome our newest member, hubetchat
Most online: 2043 users on 29 Oct 2023, 01:04 AM