Azərbaycan dili Bahasa Indonesia Bosanski Català Čeština Dansk Deutsch Eesti English Español Français Galego Hrvatski Italiano Latviešu Lietuvių Magyar Malti Mакедонски Nederlands Norsk Polski Português Português BR Românã Slovenčina Srpski Suomi Svenska Tiếng Việt Türkçe Ελληνικά Български Русский Українська Հայերեն ქართული ენა 中文
Subpage under development, new version coming soon!

Asunto: NEWS Sokker- big changes are coming!

2023-11-23 12:39:40
Geston para juytt
First of all it is not possible to predict with 100, or even 95% accuracy.
It is still JUST estimation.
I've made "formula" which recounts level of junior (with decimal places) to a value. So in this place already, where input is also estimation - predicted sumskill can be wrong.

But still... When I take all juniors data I've collected (excluding GKs) an average estimation mistake is close to... 0 (-0,148 to be exact)! Which is nice result.

Here is graph - mistake vs sumskill, it doesn't show any trend.


And mistakes distribution - 65% of predictions is mistaken max by 1 point. Another 25% by 2 points. Just 10% is mistaken by 3 or more.
2023-11-23 14:50:08
juytt para Geston
What do you mean by "mistake" ? Diff between predicted sumkills and real sumskills ? and you find avg 0 ?
2023-11-23 15:47:58
Geston para juytt
Yep, thats what I see in my data.
2023-11-23 16:40:33
juytt para Geston
So, BlueZero #1 case, mistake = predicted sumskills 38 vs real sumskills 34 = 4
And If i am not wrong, 4 is in your 10% mistakes distrib.
2023-11-23 16:45:12
Geston para juytt
That's right. Im not saying that my formula is perfect, but still it is possible that he just had bad luck with predictions.
2023-11-23 17:09:20
juytt para Geston
it shouldn't have "bad luck with predictions" since its talent is highly correlated. right?
2023-11-23 17:33:09
Geston para juytt
This is not the only thing. Even if you are sure about talent = you know junior level with decimal places for 99% it is not enough.

In such case after skills distribution there is still alot to predict.
Lets say you have X points to distribute on skills where on average every of eight skills is at good [7] level (8x good [7] skills), lets call it A case.
But it is possible that you will get player with 7x 6,8 and 1x 8,2 - B case (it doesnt work that way, higher skill needs more points, but its just for example purpose).
Case A - visible sumskill = 56.
Case B - 50.

This is something you just can't predict.
(editado)
2023-11-23 18:30:27
Well, you can. If you have the right formula and exact skillnum at the time of promotion to A team.
But we do not have either, so we are trying to do only the estimation.
2023-11-23 18:33:58
Arcade matches did not start
2023-11-23 22:11:32
Arcade

24 November 2023, 01:30 (+ 751 min)
2023-11-23 23:12:04
@ Geston

But case (A) a perfect distribution 8* 7.0 is not the most probable case ? If we assume that we can lost 0 to 7 visible skills, it's an average of 4. Judging from your stats average mistakes is 0 and only 10% 3 mistakes or more.
(editado)
2023-11-24 16:39:41
This match engine totally a joke.
Yesterday I couldn't pass the opponent goalkeeper and score a goal twice.
In the first one, the opponent's goal was empty, my player did not shoot and the opponent's defense came and took the ball. In the second time, the goal was empty again, my striker hit the post.
2023-11-24 16:41:35
Now imagine you're the opponent...

;o) But yeah, know the feeling.
2023-11-24 16:45:15
The biggest problem in this game is the match engine, I'm not saying anything else.
2023-11-24 17:08:25




(editado)
2023-11-25 21:39:07
These are still acceptable ones, you should see in my past few games, when my striker is 2m away and firest the ball to the corner post :P Superdivine form and pretty decent 14 in striker + decent construction level too :PP that makes my eyes bleed