Hey, a quick and (a bit sad) update.
I was offered a pretty decent summer job as a programmer, so I started to work there. Sadly, I'm afraid that I won't have enough time to focus on this update project at the moment :/
Nevertheless, I will restore the old coh2chart.com back as soon as Relic fixes their side. When I have time, I will do the update
I've noticed that the website has been down for sometime. Do you have plans of putting it up soon again (even the old version).
Alternatively, is there a way to get access to the raw input you are using?
If I had access to the raw input, I would compute matchup win ratios as follows:
- First, filter out all players that have been inactive (e.g., fewer than 15 games in a month for a given faction)
- Then, pick the top80 players of a given faction
- For all the games belonging to the top80 of a faction (e.g., OST) account for which matches were played vs which opponents (Soviets/USF/UKF), and whether this was a win or a defeat
Output all 12 results as follows:
- Soviets vs OST (Soviet player in Top80)
- Soviets vs OST (OST player in Top80)
- USF vs OST (USF player in Top80)
- USF vs OST (OST player in Top80)
- UKF vs OST (UKF player in Top80)
- UKF vs OST (OST player in Top80)
- Soviets vs OKW (Soviet player in Top80)
- Soviets vs OKW (OKW player in Top80)
- USF vs OKW (USF player in Top80)
- USF vs OKW (OKW player in Top80)
- UKF vs OKW (UKF player in Top80)
- UKF vs OKW (OKW player in Top80)
For 2v2+, I would compute faction win-rates as you have. However, even then, I would filter out inactive players, and I wouldn't go with buckets larger than 100 ranks each.
Hey, thank for a great suggestion!
The plan was to keep the old site running as long as possible. But, for some reason, the place where the server downloaded the data, haven't been updated for over a month now. Sometimes, Relic had some delays, but now it seems like Relic had stopped providing the data. I have contacted them, but I haven't got an answer yet. Maybe it is because of an accident, I don't know. Let's hope the best.
Anyway, the website can be made even though Relic would have stopped providing the data. The data would be gathered by using a different method -> it would only mean more work. But yeah, let's hope the best
Edit: Relic answered, it was only an accident. So, once the data is updated, the old site will be online
Thanks for feedback!
Nice to see you're back
I was never very happy with the rank ranges. 1-250 seems to wide, 250-500 was to small (and not overly useful) and 500+ much to large.
I'd suggest to use player or team levels instead. Really fancy would be when it would be possible to pick a range of player levels. Like, you could say I want the win percentages of OH players between levels 8 and 10. And then the output would be how many games happened in to total, how the win percentage was and what the average level of the opponent was (using some robust stack).
I personally feel also being able to select for the opponents level would be problematic, unless you only allow for long time bases (I would expect that one week is not enough data for that.).
By the way, is there a documentation how the level system works?
By the way, I will change the hosting server of the website in a near future, because the current contract ends. So, the site may be offline a couple of days. Fortunately, the serves are much more powerful nowadays, which means that I will have a lot more resource to use in the future
Previously, I couldn’t run the statistics gathering scripts often because of the limitations the server. After the server upgrade, there wouldn’t be that kind of problems. Scripts could gather the data from ‘the recent matches’ hourly, which means that the data could be sorted by the rank of each player. In other words, the true 'top X players versus the top X players' chart could be made. Also, it means that statistics could be sorted by map, game time etc. But which kind of data is beneficial?
I am not sure if someone remembers me, but I am the creator of the site coh2chart.com. I have been in Finland’s compulsory military service for almost a year now. The time has gone surprisingly fast, and next month I will be released from the duty. After that, I will have some free time. I have thought to start new projects and maybe continue older ones like coh2chart.com. But is it worth to further develop? I was wondering does anyone use the site anymore, but it seems like the site still has on average one and a half thousand unique users a month. Thank you for that!
So, the important question is: Does someone need a better statistics site?
If you think, that there is a need for a better version of the website: What kind of statistics/features would you like to see on the site? Which statistics/features are the most important?
Could you please discuss this elsewhere. Everyone read those discussions a 100 times already and any feedback someone gives on the amazing Coh2Charts will be lost under the usual discussion.
Anyway I have a bit of feedback.
I noticed that in the games-played-this-week diagramm there is an insane number of games played listed, especially for 4v4.
After a tiny bit of math I realized that the problem is that it counts every player playing 4v4 and thats why there are ATM 108950 games played for 4vs4 and 23755 for 1vs1.
If you divide it through 8 and 2(the number of the involved players) you get the more reasonable 13618 Games of 4vs4 and 11877 Games of 1vs1.
If you could include that in your code it would make comparing the numbers of games played per gamemode a lot easier.
The old diagramm should stay of course because you see how popular a certain gamemode is by the playercount.
I love your statistic thanks for all the effort.
Hey MADsquirrel, thanks for a feedback. Sorry for a super late answer, I didn't notice your reply. That is a great suggestion, but sadly I might not have time to do it. I will explain down below.
Hi Paid Player.
I like what you did with coh2chart, but especially now it would be great to also see the day by day stats you used to have (even if it's just for top 150 and even if its sometimes bugged). Is this possible? It would be helpful to understand what's going on with this patch and the bug fixes etc. A week averaged doesn't cut it in that regard. Of course I am only suggesting it in addition to what you already have.
thanks for a suggestion. I really agree with you. A week average is a kind of boring and it's difficult to notice the balance changes after a patch. I would like to add that feature but sadly I won't have time.
I live in Finland where we have a mandatory military service, conscription. I have to join the army next week and I have to serve in the army for next 11 months. That means that I am not able to add new features into the site
Thanks for the feedback
Just a "silly" question - that may have been asked before : how comes that the average win rate is not .5 ?
If a player has played under 10 matches, his/her stats are not included. So, that is why the average win rate is a bit higher than 0.5 when top 500+ is selected.
When top 250 is selected, win-ratio is clearly higher than 0.5 since all games played by top 250 players are counted. So, if a player (rank 150) won a player (rank 700), the victory would be counted. The lose wouldn't be counted because the ranking of the loser was worser than 250.
With current playerbase top 250 is too low skill level to make conclusion s about balance. It must be top 50-100 normally, 150 as before at least. Outside 250 is just no skill clusterfak there. So why bring it in?
Because many has request it. Or at least I thought so
I have finally managed to update coh2chart.com
. Now, it supports arranged teams and you are able to filter statistics by ranking. The data is also downloaded directly from Relic's server. (Yes, Relic gave the permission to do that
) So now, the charts are really precise!
I have been working on this project for a quite long time. I will try to update this site more often, so feel free to give suggestions. And also please feel free to donate, I would really appreciate it
Here is a picture of a random 2v2 500+ chart:
Another simple solution would be that every unit would run 10 seconds after their spawn.