Hello guys
Big update! From 12 July 2022 We are now gathering ~90% -> ~95% all matches played.
This increased the amount of games for analysis 3x.
https://coh2stats.com/stats?range=week&statsSource=all&type=1v1&race=wermacht&timeStamp=1658102400
Old crawler process:
- We gathered top 600 positions in leaderboards
Problems:
- We were missing lower rank games (which is majority of the games)
- We were gathering about 40% of all games played
New crawler process:
- Every 5 minutes we check live games, we gather played games and based on that we get those games into the system
This brings:
- Much more games in analysis
- Way less requests to Relic API (Because we request only players which played games that day)
Problems:
- Some games under 5 minutes are not gathered (Shouldn't be a problem)
- Games longer than 6 hours are not gathered (Very small number of games) - this help us save resources
Keep in mind that - that number total games is not accurate - that's in fact the ~95% of games. We are getting that number from amount of live games we track. Aka we might miss some games under 5 minutes.
The reasons why the amount of analyzed games is less that the amount of played games could be various:
- Some games end with incorrect results - there is no winner (everyone loses) - we filter them out
- Some games have corrupted data in Relic Servers - we filter them out
Reasons why there is more games than total games:
- Should be the 5 minutes games
- Sometimes it's possible that live games gathering function crashes (error from Steam / error from Relic etc ) and the game is not tracked in total games. This happens rarely tho - should only affect the games under 10 minutes.
Overall I think the analysis should give us now much better overview of the game balance!
Cheers
pagep