Supreme Commander web stats return
After a much too long downtime, the Supreme Commander web stats have finally returned. They were broken at some point (hardware failure or something?) and took an awfully long time to return. To view the web stats for a player, visit that link and type the player’s account name into the search box in the upper right. That’s it. I don’t see any way to get web stats for clans, unfortunately. And I also don’t see any way to get data for 2v2 or 3v3 ranked games. So maybe it still needs some improvement?
July 19th, 2007 at 4:39 pm
Well, just getting the 1v1 stats back is awesome. It used to be one of the big “wow” factors when showing off the game (and support) to potential new players. Just keep them up this time gpg :)
July 19th, 2007 at 5:09 pm
also, you can no longer view individual game stats and units built per game. but at least we can see once again the total units built ever and results by map and race for analyzing
July 19th, 2007 at 5:54 pm
Breezy: So web stats is “back”, but it’s still missing some of the things that it had before, thus its overall trend in time is downwards? Ugh.
July 19th, 2007 at 8:17 pm
It appears its overall trend is indeed downwards, or perhaps, we should consider this as a trend upwards…. I wish i could draw a time graph here, it would start with webstats being very high quality, then trending down as more data was added, until finally it went down. This would be ground 0 for the graph… Now they are getting it back under control and perhaps it is on an upward curve. Hopefully they improve it even more by putting results per game back eventually.
a few other things i’ve noticed with the return of webstats:
1. If you click leaderboard and then the button to flip to the very last page, the lowest ranked player according to webstats is 33000 ranked ish. His rating is around 1565. I know for a fact that there are people with lower ratings than that. A few clan mates of mine for example. However, my clan mate has a rating of 1519 and is ranked around 22000. Both profiles when searched on gpg show the same result with incorrect ranking and with the 22000 ranked having a lower rating than the 33000 ranked.
This brings me to a line of thought: there is something else that has an effect on ranking other the players ranking. It may be something that has such a small effect that it only happens on the lower end of the ladder, but there is definitely something affecting ranking other than purely the rating. I’m interested if it involves the number of games, or perhaps the number of wins. or win %. The last ranked player has 0 wins, but a higher ranking than my clanmate who has wins, but many losses also. just something that sort of bothers me.
The second thing I noticed is not something that is wrong with the system (in my opinion)
2. Sir Loui is #1 by a considerable lead. Nearly 70 points at the time of this posting. This is great in my opinion, it makes people really fight to get to him. What I find very interesting is this: He has more games than anyone else in the top 30. By a lot. He also has 305 more games than anyone else in the top 10. His win percentage is the lowest. This shows how much GPG rewards players who play more games. To prevent myself from turning this reply into an entire post itself, I think I’ll just stop talking, you get the point.
Anyway… I’m glad webstats is back. I also wish it cateloged replay links next to the last 5 maps you played on that screen.
July 19th, 2007 at 11:20 pm
Breezy: It’s not GPG rewarding Sir_Loui for playing many games. That’s simply the Elo ranking system at work, and it has nothing to do with how many games you’ve played. Sir_Loui has likely gotten better over time. Remember, Elo is an instantaneous ranking. Sir_Loui very well could be #1 right now. He wasn’t in the past, hence the losses. So looking at win/loss ratio or total number of games is missing the point when it comes to ranking.
July 19th, 2007 at 11:32 pm
I actually support the Elo system 100% however, I’d like to argue that it does reward more games played. You get better as you play more games, and youre opponents get betteralso, so if you get better at the rate your opponents do (theoretically) and you always win 60% of your games (again this is just numbers for theoretical situations), then by the time you have 1 million games, you will have 600,000 wins and 400,000 losses (disregaurd ties, it makes things easier). These numbers would make it so you have many more points won than lost, as the numbers get bigger, the point difference grows.
i’m really bad at explaining this. but the math works i’m pretty sure. I’m too lazy to figure it out this late at night right now. maybe i’ll do research later to prove myself wrong =) that way it’ll save you from wasting time doing it to post why i’m wrong, which i probably am because i’m sure you know about this stuff=P
July 19th, 2007 at 11:37 pm
FINALY!!!
WELL AT LEAST now i can bost that im number 29 in australia!!!
that makes me feel better again…
now i need about 10 more wins to get into the top 10 in aus!!!
well getting all the stats back online would be nice….
so would the custom game avatar for GPG least then u can know if someone is a noobie or just plays ranked games all the time
July 20th, 2007 at 3:58 am
Elo does reward more games played, but this isn’t necessarily a bad thing. There’s a reason I’m ranked ~1300th with only six losses — I’ve only played 34 ranked games (and all but one of them being before v3251). If RandomJoe has 600 wins and 400 losses, there’s a good chance his wins will have been against lower-skilled players and thus not worth many points. Conversely, his losses were probably against better players, also not worth that many points. RandomJoe is likely not very high on the ladder, while Loui is because he actually wins against high-ranking players.
July 20th, 2007 at 8:27 am
Another thing about the site being “fixed”. Its obvious that the DBs are created/changed only when you are logged in, and not from the main DB. I have not played a ranked 1v1 match in a few weeks. I have since quit the clan I was in (just because I wasn’t playing much, nothing personal), and yet it still shows me in that clan. I will log in tonight get smoked in a 1v1 probably (ha) and then check the page tomorrow.
On a side note, my rating is 1801, I remember the day that was good enough for 2 stars and a top 800 rating! Now I’m at 1366 or such. Obviously though this is a good thing, ‘more better’ competition. I think it’s time to try to get back into the top 1000, which was always my goal as a very casual SupCom player.
July 20th, 2007 at 10:13 am
T2A`: You say that Elo rewards more ranked games, but then you go on to contradict yourself. Elo doesn’t reward more games. As you point out, it’s that someone with fewer games has played worse opponents. But the rating is caused solely by the wins and losses against opponents of various ratings and doesn’t factor “number of games played” into it anywhere.
July 20th, 2007 at 10:32 am
Due to the way the Elo system works, it is theoretically possible to gain rating even with a win/loss ratio of less than 50%. All it takes is that your opponents have sufficiently higher ranking than you (the opposite is obviously true as well).
July 20th, 2007 at 11:11 am
I stopped playing ranked games 3 months ago. My rating hasn’t changed as far as I can see but I’ve lost 500 ranking positions.
http://gpgnet.gaspowered.com/scstats/player/ByGametype.aspx?PlayerID=144814
July 20th, 2007 at 11:42 am
*cough
http://en.wikipedia.org/wiki/ELO_rating_system
*cough
July 20th, 2007 at 11:43 am
@Malloy. oh - thanks a lot. So to crack the top 1000 I need 44+ rating points.
July 20th, 2007 at 7:28 pm
Elo is a skill based system, and not a number of games system. No where in the formula does it care about number of games. It makes an adjustments to players ratings based on the skill difference. The only way number of games helped out Sir_Loui is that he got better at Supcom itself through a lot of practice. He is clearly the number 1 player right now and I can’t think of anybody better. Watch his replays for a while if somebody doesn’t beleive it.
July 25th, 2007 at 3:13 am
Nice to know that they are back. I am a bit skeptical on using them again, not sure if the same would not happen again.
I have shifted to GoStats.com , and happy with them.