Hi everyone!

About half a year ago I did analysis on the 1v1 Ladder and created a large post showing various breakdowns like turn distributions/ player results by wins+win rate+boots/ players under different rating systems/ games by turns. I mentioned that I would take a look at the other ladders around September, and well, I decided to stop being lazy.

You can find the 1v1 ladder analysis here:
https://www.warzone.com/Forum/418612-1v1-ladder-statistics


Over the past month, I have been building a platform to track the ladders on warzone. Currently the platform has the entire Seasonal and I am thinking about adding in the 1v1 Ladder as well. It also updates every 2 hours so it will always be up to date (and hopefully stays like that). At the moment the platform houses every game from every season and tracks things such as player results by season (and overall), results by colour, games by day, and game/player counts. It is mainly intended on being an easier way to navigate the ladders as well as to see more in-depth stats on specific ladders and players.

Here is the link to the platform:
https://warzone-statistics.herokuapp.com/

At the moment I am thinking about maintaining the platform for a year, but that is dependent on how much time I have and if it is worth it. I do have to pay for the platform and I am pretty cheap, so not really sure how long it'll last (+ this was more of a side proj to practice React/Express). Expanding the platform to the main 3 ladders might be a challenge due to the volume of games and it might cause too much stress on the backend (hoping that this release doesn't crash anything :P ). If there are any additional things I should add, I'd be happy to hear it and you can mail me or ping me on discord.

Thanks again to Fizzer for allowing this to be possible with the API!

Note: The site isn't quite optimized for mobile at the moment and might look weird in portrait. It'll look best on PC, but at least use landscape on mobile :)

---

To share some of the more interesting data, here is some summarized data from the site:

Over the course of 42 seasons on the Seasonal, there have been 88 600 games played, whereupon 17 609 of those games involved boots (~20% of total games). There have been a total of 2 136 players across all seasons. The average turn length of all games is 8.71 turns.

Top 10 Players by Wins:
1.  Math Wolf (532 wins)
2.  Botanator (489 wins)
3.  ACL Tears (466 wins)
4.  tartan1314 (438 wins)
5.  Diabolicus (427 wins)
6.  master of desaster (385 wins)
7.  Timinator • apex (383 wins)
8.  konglaide (362 wins)
9.  JV (337 wins)
10. Muli (328 wins)

Top 10 Players by Seasons Played:
1.  Math Wolf (42 Seasons)
1.  Botanator (42 Seasons)
1.  majokun (42 Seasons)
1.  Diabolicus (42 Seasons)
5.  tartan1314 (39 Seasons)
6.  {101st} Grickface (36 Seasons)
6.  {101st}Blowfly (36 Seasons)
6.  DaBird (36 Seasons)
6.  jasdanmoo (36 Seasons)
10. Krzysztof (32 Seasons)

Of the 2 136 players that have played on the Seasonal, only 4 players have participated in every season (Math Wolf, Diabolicus, Botanator, majokun). The closest runner-up is tartan1314 at 39 seasons.


Number of Games by Season:

* Season X was on Four Castles FFA (and the only template to be a non-1v1)
** Season XLII is currently in progress and incomplete

Number of Players By Season:


Seasonal-Wide Colour Results:

* The only game played with Artichoke (Reserve Colour - unplayable) was by Semicedevine (Link: https://www.warzone.com/MultiPlayer?GameID=10501433 )
** Wild Strawberry is the runner-up for most underused colour on the ladder with only 8 games played (between 2 players)

Top 10 Longest Games by Turns:
1.  90 Turns - Season XXXVI - Oilspotter defeats Janine (https://www.warzone.com/MultiPlayer?GameID=18584499)
2.  86 Turns - Season X - ChrisCMU defeats Verzehrer, Lightningbolt, Bert Jordan (https://www.warzone.com/MultiPlayer?GameID=4512357)
3.  80 Turns - Season XII - The Left Dude defeats veitplatter (https://www.warzone.com/MultiPlayer?GameID=5264812)
4.  69 Turns - Season X - nameless defeats Random, Chewie, Bert Jordan (https://www.warzone.com/MultiPlayer?GameID=4540289)
5.  68 Turns - Season XI - AWESOMEGUY defeats Dobnarr (https://www.warzone.com/MultiPlayer?GameID=5003620)
6.  67 Turns - Season X - Gyver defeats amerlafrance, Galahad, Motocross5800 (https://www.warzone.com/MultiPlayer?GameID=4540288)
7.  67 Turns - Season X - Jackie Treehorn defeats Verzehrer, GW Bush, elbee (https://www.warzone.com/MultiPlayer?GameID=4524605)
8.  67 Turns - Season X - Phaeril defeats hedja, Master Ree, 00Dev (https://www.warzone.com/MultiPlayer?GameID=4571149) -- boots
9.  66 Turns - Season X - aper defeats Math Wolf, Verzehrer, Kenny Blankenship (https://www.warzone.com/MultiPlayer?GameID=4536116)
10. 65 Turns - Season X - Bert Jordan defeats Wenyun, tartan1314, Stephan (https://www.warzone.com/MultiPlayer?GameID=4512391)

Unsurprisingly, the majority of these games come from Season 10 which was played on four castles FFA.

Some really interesting stuff there, thanks for pulling together and for the lovely charts!

Personally I thought Season X being FFA was good fun, though I understand why some didn't like it (and hence the huge drop off in players)

Thanks!

Surprisingly there wasn't any drop in players for season X. The number of games also seems proportionate with the previous season given that the games were between 4 players instead of just 2 (and looking at game counts, the # of games in season X is a little under half that of season IX). Just did some quick queries on the database and seems like the average games per person for season X was 11.82 whereas season IX was 16.62 which might account for the rest of the difference.

Yeah I realised later that the number of games must have fallen due to there being more players per game!

One thing that might be a better indication on the popularity of each season could be looking at how many players did more than half of their games, instead of games played or players signed up, as it shows intent to play through at least most of the season. Unsure how the database is structured or easy that would be to make but could be interesting to see how it differs from players signed up.

Nice jump in players for the Strat ME season haha

Finally got around to doing some of this analysis. The first column of numbers is the number of players in the season; second column corresponds to the number of players that played at least 10 games (and 7 games if season X FFA); and last column is the percentage of people that played at least half the number of games vs total players.

Season		Players		w/ >= 10G / 7G		%
I		146		129			88.4
II		142		119			83.8
III		132		109			82.6
IV		119		82			68.9
V		123		98			78.0
VI		139		117			84.2
VII		143		117			81.8
VIII		163		134			82.2
IX		161		134			83.2
*X		179		134 / 161		74.9 / 89.9
XI		236		189			80.1
XII		250		194			77.6
XIII		256		211			82.4
XIV		306		250			81.7
XV		274		194			70.8
XVI		283		213			75.3
XVII		312		265			84.9
XVIII		323		265			82.0
XIX		313		242			77.3
XX		286		224			78.3
XXI		352		288			81.8
XXII		370		286			77.3
XXIII		362		293			80.9
XXIV		373		300			80.4
XXV		318		231			72.6
XXVI		317		258			81.4
XXVII		311		233			74.9
XXVIII		283		233			82.3
XXIX		296		223			75.3
XXX		291		227			78.0
XXXI		265		209			78.9
XXXII		291		243			83.5
XXXIII		269		229			85.1
XXXIV		267		187			70.0
XXXV		265		220			83.0
XXXVI		257		208			80.9
XXXVII		258		215			83.3
XXXVIII		276		237			85.9
XXXIX		279		228			81.7
XL		366		316			86.3
XLI		361		294			81.4
XLII		315		269			85.4
-------------------------------------------------------------------
μ		262.6		211.3			80.5
σ		74.2		60.0			 4.61

Min							68.9		
Q1							78.0
Median							81.7
Q3							83.3
Max							89.9

Season XLII is still in progress, so the proportion should only improve with over half a month remaining.

Surprisingly the proportion of people completing at least half of the games remain relatively consistent throughout the entire ladder. From what I read, the FFA season (X) was intended to have a max of 12 games, but many players got 14. Either way, the proportion of players that played at least half of season X was the highest of any season. On the opposite end, the worst season was season IV which was played on ME SR

The first and last few seasons appear to be consistently higher than the average %. The first few seasons would make sense with the introduction of a new ladder, but I am curious if COVID has lended to improving the proportions or some other factor accounting for the uptick. The dataset appears prime for running statistical tests, but unfortunately I don't have the time :(

Edited 12/14/2020 01:32:14

looks great, and interesting. takes a far smarter person than me to manage to make such a thing. i'll check in on it from time to time~

Edited 12/14/2020 03:06:27

The color results always fascinate me for some reason. Of course win rate is more normalized on seasonal after the first few games, since you draw opponents closer to your level.

I started playing mardi gras partway through georgia. Since then my results 9-5 (georgia), 13-7 (landria), 13-7 (yorkshire). That puts me at 35-19 (0.648) playing as mardi gras. Mardi gras as a whole 233-200 (0.538). Remove my contribution and mardi gras is 197-181 (0.521).

Cool to see I've added 1.7% to Mardi gras win rate.