BuzzerBeater Forums

i think the best way to do a analisys of game shape is to put the data like this one:

I think the original poster preferred the simple method, where it's easy to see how much it changed.

Adding the skill names just adds an extra layer of difficulty that is not necessary.

However, if the previous game shape is in the post, it may have relevance in weeks beyond this one where not all players have the same form coming into the week...

But I understand your point.

I think he wanted to do it this week because we know all players were starting at respectable game shape after being reset. He's trying to get a base of time played to game shape movement from the reset value of respectable.

I don't believe this will be continued after this week.

But now in this second week, how you know who player goes down and who players goes up, if i put to play someone 50 minutes that didn't play last week, and he is "atrocius" game shape and go "mediocre"? it can turn the name of the level to the number of the level, but to do a good analisys you have to check the last game shape and the new game shape, and it will help better to the user and the community.

why now?
because only before the last training all the players had respectable game shape, and when players start the training with different game shape:
a) some user could not remember previos game shape of his players (you know...for game shape no arrow is showed)
b) probably the passage from respectable to strong, or from strong to proficient is different rather the one from the medium value (respectable) to the others...so it could be harder to study that relationship

excuse me for my bad english...

anyway ....data and graphs are ready...i will link or post them as soon as possible!

Edited by Frescobaldo1 (09/02/2008 11.06.11 CET)

Last edited by Frescobaldo1 at 2/9/2008 11:06:11 AM

Thanks for taking the time to put this together.

*applause!

This is the one week where GS is the same, with no high or low sublevels.

In future weeks we wont have the same "control" data.

Results of the research.

You will find the link to the webpage with regrssion graphs and data that define the relation between game shape variation and minutes played in a week by a player.
The study was performed with the data of 54 users and their 743 players, the software used for the analisys is statview.

As you can see on the graph (simple regression) and on the sheets, the relationship between the two variables is not so high (0,42), but statistically significative (p<0,0001).

The best category reslts the one from 45 to 70 minutes, or better from 50 to 65; most of +1 and +2 is around the 60 minutes value. However, looking at the Gauss lowess you can see that mean best value for 60 minutes is under +1...so if your player take +2 you are quite lucky.

Categories from 30 to 45 minutes and from 75 to 90 minutes have different results: everything could happen in those categories, but the more players go toward 60 minutes value, the more they could increase their game shape.

Moreover data confirm that is no good for a player to play less than 24 minutes and overall more than 96 minutes. Under 5 minutes and over 105 -2 in game shape is sure!

Descriptive data and graphs are linked if you want...

Limitations of this research:
734 players are few for this kind of research ...sample should be larger.
Dependent variable (game shape variation) is not really continuuos (it can be only -2,-1,0,+1,+2)
I made my last stat exam 2 years ago....so... :-)

http://img341.imageshack.us/my.php?image=varformaxminutes...
http://img167.imageshack.us/my.php?image=relationbetweeng...
http://img177.imageshack.us/my.php?image=graficidescritti...
http://img232.imageshack.us/my.php?image=descriptivedataw...

bye!

Interesting data. Thanks!

Great work!

Hopefully you'll be willing to do this again next season, and will get even more data (I didn't find this thread until after my weekly stats had been reset. So I, for one, will add some data next season if you're willing to do this again =)