Nice work!
i know how to import sql. problem is that i just copied/pasted all battles into excel and data is unfriendly for importing without prior arranging. my preferred choice would be to make some script to pull all data directly into sql, just that it required too much effort.
btw only manual thing that i did was click load more replays for some 100 times ( end+mouse click). then i copy/pasted whole page to excel, put formulas for counting specific battles/modes (countif) and formula for summing nth row. this was done relatively fast.
1 Like
1000? Looks like you’re overestimating it.
1 Like
Most likely lmfao
Also Pacific somehow more battles than Shitgrad and Shitnisia lol
i looked into importing data into sql, but it simply takes too much time. so if anyone is interested here is link where you can datamine all data https://www.enlisted.net/en/replays/load/page/1/?Filter[type]=general&action=search, page number goes up to 400 (that is ~10k matches)
and i made another analysis with more matches. i removed playerbase cause i am not certain that i am doing it right and bots pollute result anyways.
so anyways here are more results with 8.1k matches taken from some 15:00 CEST till 20:00 CEST:
not much difference between original data. normandy has few % less, while some campaigns have few % more.
there is change in average time, so average match is 1 minute less than original result.
here is breakdown campaign/mode
and in percentages
and finally average time played per campaign/mode
edit:
here are also longest battles
7 Likes
It was a great visualization of the question that we all somehow had, that there were hardly anyone playing in Tunisia and the Pacific. Thanks for your excellent research.
well this could be analyzed better if data was in friendlier format. maybe one day when i am not lazy i will make proper database that will pull week or month of battles alongside playerbase and bot analysis. btw bots are instantly recognized in replay data so we can finally get answer to playerbase question.
btw just did quick player/bot analysis of last 250 battles at around 21:15 CEST (so peak europe time). assuming perfect world and no MM stack bug, i assume perfect 20 players per battle for all 250. we get that around 16% players desert the battle. on average we enter games with 40% of bots and exit with 45% of the bots. some campaigns are more bot heavy, so take the results with grain of salt.
you can check bots on that replay link that i previously shared. they have “userid:0”. i just checked raw data from 10th page towards 1st (for no duplication of results) and searched userid":0 for bots and just userid for total players and wrote number of found results.
edit:
fixed player deserted with actual % of human players.
1 Like
was bored so i made quick script to download info from 10000 battles just to see number of bots in them. results are from 11:00 CEST for last 10k battles. also i counted number of console players and pc player in those battles (so it is not unique result).
when i have time i will try to fix data in the file so i can mine better info (like number of unique players), or data about playerbase per campaign as well as enlarge the database for more accurate info.
1 Like
Before the major update they should rename their game into Botslisted lol
3 Likes
have made database, need more data to make great analysis. but here are some data
next few data is taken from 1 whole day (yesterday) from all 3 servers
total unique players
unique players by platform
peak hours with unique players
peak hours by platform unique players
non unique players count
peak hours non unique (players that played multiple battles in that hour)
peak hours by platform non unique
this is average playerbase analysis for one day.
this analysis doesnt include MM stack bug (aka 10v13 bug) which is suspected to be around 3% of total battles.
4 Likes
btw i will need at least 6 players to DM me. doesnt matter if i get more. just include server and crossplay on/off.
1 player that only played NA server (PC player or crossplay off) yesterday or today
1 player that only played EEU server (PC player or crossplay off) yesterday or today
1 player that only played EU server (PC player or crossplay off) yesterday or today
1 player that only played EU server (console player crossplay on) yesterday or today
1 player that only played EEU server (console player crossplay on) yesterday or today
1 player that only played NA server (console player crossplay on) yesterday or today
data doesnt include server played, so i want to make table of players by server for better analysis. i just need at least one player from one of those groups to fetch other players from those servers.
here is some extra for same data analysis i did 2 days ago. data for hours 13 and 14 are incomplete cause i didnt pull all data (missing about hour of data in between). i am also missing about hour of data for 26th. didnt automate the process, so still have manually start the program.
so here is data from yesterday, it should be fairly accurate.
players unique per side
players non unique per side
playerbase by campaign (unique)
playerbase by campaign (non unique).
data is pretty shocking. normandy is most balanced campaign, pacific isnt one sided botfest like everyone is saying tunis has 2:1 ratio of players as expected.
number of unique players per campaign per team for all hours
number of non unique players (more relevant for seeing player disparity per battle) per campaign per team for all hours
i am missing servers and cant datamine this, so i need players to DM server that they play on if they only play on one server so i can accurately determine servers.
I highly doubt your stats about alive/bots, since you can be constantly playing with/against the same names in 3-4 games in the row.
Especially, in 8-12 hours out of prime time. And in campaigns like Stalingrad
If you datamined nicknames, you can try to search for those nicknames ingame. And, also, you can count average games per name. Pseudo-“players” will have their numbers really low, since there is a pool of names to fill the game when alive player is leaving.
idk what is so hard to believe about playin with same names 3-4 games in a row. you are in same cycle, you all finished same battles and overall you are more probable to play against same player if both of you press play at the same time.
you are not taking into account that there are 6 servers (3 crossplay on, 3 crossplay off) with crossplay off making around 22% of total battles. i will just quote myself here
Surely, since by means of extrapolation the actual daily online was like 25-30k alive people daily in December 22. Or even lower.
Counted via Xbox stats.
Another option to count the proportion is to have like 10 of series for 5 games in well-feed area, like Normandy, and in Stalingrad. The amount of same names found in those series could give the idea, how well is the playerbase in those campaigns.
I can’t believe the id-check, since you can enter the replay and validate the behavior (aim) of “players” from it. From what I’ve tried - up to 2/3 of team could be bots, sometimes even more of them.
Have you ever tried “search name” test?
you mean searching ingame for friends? and it really doesnt matter to me cause i datamine directly from replay site where i get name and id.
you want exact stats? here is for xbox
started taking data on 24th so it is incomplete, ended on 1st so it is not whole day and there is around 1 hour of gap on 25th and 26th where i didnt fetch data.
here is for psn
counting unique names doesnt give you insight in healthiness of the game. if you are playing one after another it is more likely for you to end up in same battles as them. also game will have same health if there is 1 player playing 10 games, or 10 players playing 1 game in regards to player count.
more relevant stat for healthiness of the game is player disparity between sides and player/bot ratio.
there are many problems with replay and you cant validate behavior of anything in them.
Can’t agree with that. Players with 1-5 games in month have zero skill and can’t be really distinguished from bots. They don’t build rallies, they evade friendly bombers/artillery raids.
I highly appreciate your efforts! Still, it would be really interesting to see the distribution of playerbase by number of games per month. Eg. how many players play 10+ per month
90 of 100 quitting the game after 3 plays - is a very bad sign for the game as a product