User avatar
Greece BrookG
ESOC Media Team
Posts: 1177
ESO: BrookG
Location: not Amsterdam

11 Jun 2019, 08:01

Image

It is fairly hard to provide an interesting story that comes with the complicated statistics. So, we are skipping this step and letting the community to tell the story itself. Here we will be compiling spreadsheets with tournament data, regardless of size. The community can freely browse the sheets, and see various statistical facts like Civ Selection, Map Preferences, popular MatchUps and even aging times. Below I am including the links to the Google Sheets. This is a on-going project so if you have ideas, thoughts, feedback, or if you want to contribute yourselves, feel free to share it either here or personally to me :flowers:

Attachments
Autumn Championship 2018.xlsx
(157.51 KiB) Downloaded 20 times
New Year's Classic 2018-2019.xlsx
(13.91 KiB) Downloaded 11 times
New World Championship 2019.xlsx
(103.02 KiB) Downloaded 46 times
Underused Cup 2019.xlsx
(60.16 KiB) Downloaded 18 times
Correlation doesn't mean causation.
http://www.tylervigen.com/spurious-correlations

"mr.brookg go buy jeans and goto the club with somppuli" - Princeofkabul, July 2018
User avatar
United States of America Cometk
ESOC Media Team
Posts: 3667
ESO: DJ_Cometk
Location: California

12 Jun 2019, 01:14

look at what I stumbled upon :)

ESOC Best Maps Showdown 2019 Statistics
sebnan12 wrote:whenever i see a picture of siege elephants i question why they do 40~ dmg when they hit u. that phat cannons probably loading coconuts
User avatar
Serbia ShinkuroYukinari
Musketeer
Posts: 70
ESO: ShinkuroYukinari

14 Jun 2019, 13:05

No WoL Jamboree? :o
First get to Age 4, then complain about Bulgaria :hehe:
User avatar
Greece BrookG
ESOC Media Team
Posts: 1177
ESO: BrookG
Location: not Amsterdam

25 Jun 2019, 11:29

Two Towns Tournament added and few details edited in Underused sheet :bump:
Correlation doesn't mean causation.
http://www.tylervigen.com/spurious-correlations

"mr.brookg go buy jeans and goto the club with somppuli" - Princeofkabul, July 2018
User avatar
Greece BrookG
ESOC Media Team
Posts: 1177
ESO: BrookG
Location: not Amsterdam

29 Jul 2019, 16:37

ΝWC sheet was updated with LAN stats
Correlation doesn't mean causation.
http://www.tylervigen.com/spurious-correlations

"mr.brookg go buy jeans and goto the club with somppuli" - Princeofkabul, July 2018
User avatar
New Zealand zoom
EP Project Lead
Posts: 8594
ESO: Funnu
Location: New_Sweland

29 Jul 2019, 16:42

Excellent job, guys!
Effective ESOC Patch notes

"In general you shouldnt say gg before clicking resign IMO"

"♪You can't trust anyone, because you're untrustable. How can you trust someone you know can't trust you?♫"
User avatar
France [Armag] diarouga
Gendarme
NWC LAN Gold
Posts: 9564
ESO: diarouga
Location: France

29 Jul 2019, 17:26

A few analysis :

Maps

1) Players like to play on Kamchatka, Florida and Pampa which is why these maps weren't veto'd a lot.
Players don't like to play on Trash Desert, High Plains and New England.
My conclusion : Everybody likes to play on Standard maps, but it's likely at least one of the two players doesn't like to play on unstandard maps. Thus, I think that least half of the tourney maps should be standard.

2) India was overplayed on Deccan as the map is just too good for India.
My conclusion : Remove Deccan from the future map pools, as there's only one way to play it.

3) People almost only play India and Russia on no TP maps.
My conclusion : I've always been against no TP maps for that reason. You can't age (because no TPs), so you have to take the map and play the game in late colonial. I don't think that having only 2-3 viable civs with one playstyle on a map is healthy for a tournament. Thus, admins should be careful when making the map pools. I think that more than 1 TP maps for a BO7- is too much, and then it should be a good no TP maps (not Trash Desert :chinese: ).

MUs and civs :

1) Thanks to the rules, we didn't have a lot of mirrors, which is good imo. Furthemore, almost 90% of the possible MUs were played in the tourney, and more than 50% at the LAN (I'm actually surprised).
My conclusion : Keep the same picking rules.

I couldn't find the win rates, so can't discuss balance.

Meta :

1) In more than half of the games, one of the players went age 2 and never aged.
My conclusion : I'm not sure that there's an issue with the "age 3 meta". If half of the people play in colonial, it shows that they believe that playing age 2 is possible. Both players don't have to play in colo though, and age 2 vs age 3 is interesting.
User avatar
Greece BrookG
ESOC Media Team
Posts: 1177
ESO: BrookG
Location: not Amsterdam

29 Jul 2019, 17:47

Regarding winrates, it was an educated decision to exclude them, because the amount of matches would have been too small for providing any significant result. Any conclusion would have been biased. Consider, for example, the Spain situation. It was one of the least played civs (32 times), any winrate wouldn't be suggestive of the civ's actual status in the civ pool.

Touching upon the map issue, I will disagree on two points. Deccan was picked because of its special starting crates and its unique geography. Indians were highly chosen in LAN (8 times). In the rest of the tournament Indians were played 18 times, followed by British (17 times) and French (11 times). In my opinion, it is hard to deduce something about them just being too strong on that map. These civs were top picks in the whole tournament in the first place. Lastly, Deccan was only 6th out of 9 maps played in total and 1st in LAN. Thar Dessert, however, and New England were last, which might mean that non-TP maps aren't as popular with the players and New England honestly offers more strategic options and is a harder map for less experienced players.

Non-TP maps although unpopular are necessary because you would have to reiterate similar MUs and a tournament of such a scale requires to test the players in all possible situations and skills. And yes Indians and Russians is safer than your other options are maybe Dutch, Aztecs or Japanese among others. Is it a map's problem that they aren't picked or civ balance issue?

And about "age 3 meta" you may read again, because at least one player went to age3 in ~9/10 their games. One player staying in colonial, doesn't exclude the situation where the other aged to Fortress.
Correlation doesn't mean causation.
http://www.tylervigen.com/spurious-correlations

"mr.brookg go buy jeans and goto the club with somppuli" - Princeofkabul, July 2018
User avatar
Canada Mitoe
ESOC Media Team
EWTNWC LAN Top 8
Posts: 4098
ESO: Mitoe
GameRanger ID: 346407

29 Jul 2019, 18:11

I strongly disagree with Deccan being a limited map. I think pretty much every civ is playable there. Like NE, I don’t think the players really tried to figure things out as much.


I still think winrates should be included. It’s simply the most interesting statistic regardless of its implications.
Great Britain I_HaRRiiSoN_I
Skirmisher
Posts: 176
Location: United Kingdom

30 Jul 2019, 13:21

I've only briefly had a look through one of the sheets, It looks really in depth and excellent but a bit of a ball ache to keep on top of things. I think adding input data in terms of a massive table and using pivot tables and pivot charts (charts with filter options) to automatically read off and process the data is the next step to go.

I will have a longer look when I get home from work however excel spreadsheets and dashboards/reports and general Data stuff is my forte.
User avatar
Greece BrookG
ESOC Media Team
Posts: 1177
ESO: BrookG
Location: not Amsterdam

12 Aug 2019, 17:25

Updated Winrates on NWC 2019 Statistics. Please be careful when explaining them, since the games played aren't always enough to extract a definitive output.
Mirrors were exlcuded

spoiler
Correlation doesn't mean causation.
http://www.tylervigen.com/spurious-correlations

"mr.brookg go buy jeans and goto the club with somppuli" - Princeofkabul, July 2018
User avatar
France [Armag] diarouga
Gendarme
NWC LAN Gold
Posts: 9564
ESO: diarouga
Location: France

12 Aug 2019, 17:40

Tbh they're still more relevant than the pr40 vs pr20 games.
User avatar
Greece BrookG
ESOC Media Team
Posts: 1177
ESO: BrookG
Location: not Amsterdam

18 Aug 2019, 08:37

In the process of maintaining the same format for all sheets, a complementary sheet was added for Autumn 2018.
Correlation doesn't mean causation.
http://www.tylervigen.com/spurious-correlations

"mr.brookg go buy jeans and goto the club with somppuli" - Princeofkabul, July 2018
User avatar
Greece BrookG
ESOC Media Team
Posts: 1177
ESO: BrookG
Location: not Amsterdam

07 Sep 2019, 15:57

Watch the EPL stat sheet update, updated the thread
Correlation doesn't mean causation.
http://www.tylervigen.com/spurious-correlations

"mr.brookg go buy jeans and goto the club with somppuli" - Princeofkabul, July 2018

Forum Info

Return to “General”



Who is online

Users browsing this forum: kami_ryu and 1 guest