Subpage under development, new version coming soon!
Asunto: oSokker
If you have some other solution of calculating talent just let me know. AFAIK linear regression works the best.
I think the best solution is not to do any predictions in the first 8 or 10 weeks :) Just put in 'unknown' or something like that. Ofcourse after 8 or 10 weeks the predictions can still be way off, but at least then more predictions will be realistic. And from that moment you can use linear regression but then it's based on more data.
And I found something that isn't working but don't know if it is already posted, I can't rearrange juniors by skills (best or worst skill on top).
I think the best solution is not to do any predictions in the first 8 or 10 weeks :) Just put in 'unknown' or something like that. Ofcourse after 8 or 10 weeks the predictions can still be way off, but at least then more predictions will be realistic. And from that moment you can use linear regression but then it's based on more data.
And I found something that isn't working but don't know if it is already posted, I can't rearrange juniors by skills (best or worst skill on top).
Well you can ignore predictions int the first 8 to 10 weeks.
I find it usefull :-P.
Do you have to latest version? :-P
I find it usefull :-P.
Do you have to latest version? :-P
I find it usefull :-P.
You can't mean that, you know to much of this to say this and really mean it :) And to new managers this must be really confusing to see predictions of 20+ or even 30+ weeks in the first few weeks in juniorschool.
And yes I have the last version installed. The others all do work, age, weeks etc, but talent doesn't.
(editado)
You can't mean that, you know to much of this to say this and really mean it :) And to new managers this must be really confusing to see predictions of 20+ or even 30+ weeks in the first few weeks in juniorschool.
And yes I have the last version installed. The others all do work, age, weeks etc, but talent doesn't.
(editado)
maybe you can add the accuracy of the talent thingy (i dunno what the word for the calc is)? so that its more clear so for beginners etc..?
I doubt that there can be talent 20+ in 10 weeks. :-P
Any talent > 10 means for me that the player is thrown out :-P
You forgot to mention, that it was the talent what you cant sort :-P.
Any talent > 10 means for me that the player is thrown out :-P
You forgot to mention, that it was the talent what you cant sort :-P.
maybe you can add the accuracy of the talent thingy (i dunno what the word for the calc is)? so that its more clear so for beginners etc..?
yes, that could i do :-P
yes, that could i do :-P
like i posted you in my last pm i had a guy with an official talent of 44 once using the official linear regression formula.
needless to say there's no player with that amount of weeks in any youth squad so the formula might have been 'predicting' it as a result or something? XD
seriously: the result of this type of calculation fluctuates a lot when the number of weeks is limited. any deviating estimation (which is possible in any week, including the first and the last of the series) can have a very big impact.
needless to say there's no player with that amount of weeks in any youth squad so the formula might have been 'predicting' it as a result or something? XD
seriously: the result of this type of calculation fluctuates a lot when the number of weeks is limited. any deviating estimation (which is possible in any week, including the first and the last of the series) can have a very big impact.
I doubt that there can be talent 20+ in 10 weeks. :-P
Exactly, that's my point :) After 10 weeks this won't happen that much anymore, or not at all, but in the first few weeks it can and does. And that's what makes it so confusing to give predictions in the first 8 or 10 weeks. You and I understand it, but new managers won't and tools should also (or maybe even mostly) help those who don't understand it all.
For example, if a new manager sees talent 30+ after 5 weeks he will delete this player, but in the next 3 to 5 weeks the talent will be completely different and maybe this player was even worth keeping. And without money to buy/hire a good junior coach the talents will be even less predictable and more often 20+ or 30+ will happen in the first few weeks as I have.
You forgot to mention, that it was the talent what you cant sort :-P.
And indeed, I can't sort it out, but oSokker can't either because there just isn't enough data to make a realistic prediction in 5 or 6 weeks of juniorschool, but instead it gives impossible talents 20+ or 30+ predictions ;)
This prediction would make more sense, talent 0.00 in 6 weeks because of not enough data. I would prefer 'unknown' but 0.00 would also be ok :)
(editado)
Exactly, that's my point :) After 10 weeks this won't happen that much anymore, or not at all, but in the first few weeks it can and does. And that's what makes it so confusing to give predictions in the first 8 or 10 weeks. You and I understand it, but new managers won't and tools should also (or maybe even mostly) help those who don't understand it all.
For example, if a new manager sees talent 30+ after 5 weeks he will delete this player, but in the next 3 to 5 weeks the talent will be completely different and maybe this player was even worth keeping. And without money to buy/hire a good junior coach the talents will be even less predictable and more often 20+ or 30+ will happen in the first few weeks as I have.
You forgot to mention, that it was the talent what you cant sort :-P.
And indeed, I can't sort it out, but oSokker can't either because there just isn't enough data to make a realistic prediction in 5 or 6 weeks of juniorschool, but instead it gives impossible talents 20+ or 30+ predictions ;)
This prediction would make more sense, talent 0.00 in 6 weeks because of not enough data. I would prefer 'unknown' but 0.00 would also be ok :)
(editado)
well you should see the results with a lesser coach because they'll be a lot worse :p
and also: if you know you don't have enough players left to keep your youth team open if you fire all those with a bad talent then you might just settle for some of these less talented ones.
i've a few players which i would like to throw away even though i managed to get promoted to the second junior division with them ...
(editado)
and also: if you know you don't have enough players left to keep your youth team open if you fire all those with a bad talent then you might just settle for some of these less talented ones.
i've a few players which i would like to throw away even though i managed to get promoted to the second junior division with them ...
(editado)
New version fixes my NTDB problem. Thanks again for all your great work Bluezero. Legend.
I got your point.
I am unable to get in the mind of newbies. And i know what is the algorithm. Therefore i understand better the numbers.
0.00 is beter for sorting :-P.
Now we just have to come up with a constant of talent which will say that the talent is unknown.
The constant may depend on number of weeks in the YS.
My suggestion:
weeks < 10 && talent > 15 => talent unknown
weeks >= 10 && talent > 20 => talent unknown
I am unable to get in the mind of newbies. And i know what is the algorithm. Therefore i understand better the numbers.
0.00 is beter for sorting :-P.
Now we just have to come up with a constant of talent which will say that the talent is unknown.
The constant may depend on number of weeks in the YS.
My suggestion:
weeks < 10 && talent > 15 => talent unknown
weeks >= 10 && talent > 20 => talent unknown
with new local version on ff works fine, i have access to settings but still i cant watch matches, so im gonna be watching games on opera, and do other stuff on ff :D
I'm using FF 3.6.26 and oSokker works good... :)
Btw: I recommend Newgrounds Classic addon for this version of FF ;)
Btw: I recommend Newgrounds Classic addon for this version of FF ;)
for the weeks >10 you cud set when talent >20 => sack me
:P
or maybe for 15 anyway
:P
or maybe for 15 anyway