Asunto: NEWS Sokker- big changes are coming!

It is proportional in term of relative progression, I think I keep repeating this, It’s not in term of absolute gain.

I said exactly why it's not possible:
I gave you statistical evidence that youths can increase their gk skill from 0.x to 5.x (so at least 5 times) while increasing their overall sumskill by less than 3 times, meaning it's not proportional in this sense.

So why your data are relevant when you count 0 gk? Since one skill can start and ends at 0-1 it doesn’t say anything about other skills.
Do you even read what I write? I said that there are youths that start at 0 gk and end with 5 gk. That's EXTREMELY RELEVANT to your theory. If the growth was proportional and it was possible to get more than 5 times growth to the initial skill with high enough talent then it would be enough to have 5x [2] initial skill mid (which happens a lot of time) to produce 5x [10] youth at the end of the youth school. You would get 5x[10] mid from initial (good enough talent) 5x[3] mid in ~22 weeks. That's not what is happening with youths, and that's why your theory is wrong.

I just didn’t see anything about gk growing 0.x to 5.x. How did you find this ? You count number of gk at 0 and 5 skills (and sum skills 20+) presuming it can starts at 5 (only two weeks, ok) and so what ? This table doesn’t mean it can grow from 0 to 5 end it just mean it can start at 0 to 5 with a sumskills of 20+. or I miss something ?
(editado)

I looked at youths that were 2 weeks in youth school and had sumskill higher than 20 (to only look at youths that have potential to be at least magical if the spend a lot of time in youth school). About 86% of them had gk skill 0, so I conclude that at least 86%* of high initial level youths start with gk at 0. Maybe a bit lower if we take into consideration that my data is limited, let's make it 80% to account for that.

* most likely higher number, as some must have popped from 0 to 1 in these 2 weeks, but it doesn't really matter

Then I looked at youths that have high sumskill at the end (from screenshot with sumskill between 52 and 56). 61 out of 207 youths had gk at least 5, so ~31%.

But at most 20% of youths start with gk at least 1, so some part of that 31% that ended with at least 5 had to start at 0.x

So first you conclude that 80% juniors and sumskills 20+ start with 0.x gk ? And that’s all. Are we ok ? Note that’s just a little part of your data (less than 100 juniors) 80% it’s not really representative, but anyway.

How it’s relevant for a group of juniors ending sumskills 50+ ? They should start at sumskills 20+ ok but nothing tell us if they start at gk 0 or not. 100% of this juniors of sumskills 50+ could be in the group of 20% of juniors of 20+ starting at least 1 gk, it’s not statistically wrong. And Imo I'm inclined to think that most juniors with higher sumskills are part of this 20%.

Your data just tell you that 31% ending at 5 gk and nothing more, you can’t conclude anything about gk starting level.
(editado)

So first you conclude that 80% juniors (ending sumskills 20+) start with 0.x gk ? And that’s all. Are we ok ?
Yes, although in reality it's probably higher, and if we only look at field youths and take into account that in those 2 weeks some youths went from 0 to 1... I wouldn't be surprised if all field youths started with 0.x gk. And I've seen field youths that came out with gk as high as good...

How it’s relevant for a group of juniors (ending sumskills 50+) ? 100% of this juniors (50+) could be in the group of 20% of juniors (20+) starting at least 1 gk. Imo I'm inclined to think that most juniors with higher sumskills are part of this 20%.
I assume initial gk skill, talent, weeks at youth school are independent from each other, then distribution of initial skills for high sumskill youths should be similar to those who come out as 20+ sumskill after 2 weeks in YS.
When you get good youth for 34 weeks with top talent he is going go be that divine youth and he is going to have sumskill of that divine youth... it doesn't matter if his gk was at 0.x or 1.x initially. And if 80%+ of youths with initial level good start with level 0.x gk then I also expect that in the subgroup of these youths with top talent and 30+ weeks in ys there will also be 80%+ with initial 0.x gk skill.


We have this discussion backwards - it should start with sumskill being strongly correlated with youth level. For every level you get on average extra ~3.2 skills, bit more at the highest level. There are some small differences sometimes (usually with gk youths, they often have smaller sumskill, as was the case with that youth which started discussion), but in 90% cases one can predict the sumskill with +-2 skills accuracy, without looking at initial skill.

If your theory was true then it would completely break this correlation, because initial skills would have much bigger impact on final skills.
For example someone starting at level [2] and finishing at level [12] would need to have about ~half of the final sumskill of someone starting at level [4] and finishing at [14] (because initial skills of level [2] are ~half of that of level [4], so proportional increase with the same talent = factor would keep this proportion). And that's not happening, differences are never this big :)

in 90% cases one can predict the sumskill with +-2 skills accuracy, without looking at initial skill.

I take your word. Predict the sumskill of these juniors, if you will be right in 90% then i believe you:


Training juniors Age Skill level Formation Weeks
Michal Káník 19 outstanding [12] outfield 1
Richard Vavřička 18 excellent [10] GK 2
Prokop Mestl 18 formidable [11] outfield 3
Marcel Kudláček 18 formidable [11] outfield 3
Arnošt Heřmánek 17 good [7] outfield 3

magical coach.

In my experience the sumskill is very random. Most probably because of bad estimation. But you said you can predict it, so lets test your theory.

Just use Geston's tool on the days they come out and tell us how far away it was with it's prediction :P
Personally I would take 1 point from this prediction on high levels, I feel like it more often than not overestimates the sumskill by a little

I would like to throw in an estimated guess ;-) I will make a small range guess with a range of 3 skillpoints, since every general skill rise will contain approx 3 skillpoints...

Michal Káník 19 outstanding [12] outfield 1 = 36 to 39 skillsum
Richard Vavřička 18 excellent [10] GK 2 = 31 to 34 skillsum
Prokop Mestl 18 formidable [11] outfield 3 = 35 to 38 skillsum
Marcel Kudláček 18 formidable [11] outfield 3 = 35 to 38 skillsum
Arnošt Heřmánek 17 good [7] outfield 3 = 23 to 26 skillsum
(editado)

The point was for you to prove the point, not for me to find out a "prediction". We wanted to establish if your algorithm is 90% correct as you say it is.
I predict that max 1 player from the 5 will stay in the team other will be either sacked or sold for 1 $ - 1 000 $:-).
(editado)

Just took a look at the tool it uses linear regression, so basically the same i built in oSokker (R.I.P.).
While that worked "fairly" fine when the exact current level was displayed, it will work worse now.
Also the correct would be to use logarithmic regression, because the higher levels jump slower.

That makes sense. Max level of a skill seems to be limited by a logarithmic scale of the Junior's level too, so the proportional theory of the skills depending on the Junior's level could fit in a logarithmic scale
(editado)

The point was for you to prove the point, not for me to find out a "prediction". We wanted to establish if your algorithm is 90% correct as you say it is.

I've seen tens of screenshots made with this tool and it almost always predicts within that +-2 of real sumskill. Sometimes it's 3 skills lower, that's why I deduct 1 point on higher levels to have it more likely to stay in that +-2 range. I won't predict sumskill for your youths without knowing their historical data and before they get all the trainings in youth school... So you can check it yourself in that tool and treat it as "my" prediction and possibly prove me wrong this way :)

Also the correct would be to use logarithmic regression, because the higher levels jump slower.

Are you sure about that? From my experience it stays linear, and was linear in the old times when we had exact level known, but maybe I don't know about something.

While that worked "fairly" fine when the exact current level was displayed, it will work worse now.

When youths play in youth league it stays pretty close to the predicted line, either on the line or 1 level off. You can estimate the real final level decently knowing the history of such youth... maybe that's why I think the prediction is that good, because I usually see screenshots from users that have their youths playing in youth league :)

@Terrion
Max level of a skill seems to be limited by a logarithmic scale of the Junior's level too
You mean individual skills are limited? There are many examples of youths coming out of school and having 10 in one skill and less than 10 points total in all other skills. Doesn't look limited by overall level in my opinion :)

> You mean individual skills are limited?

That was my feeling. I never saw such a thing as an excellent skill with few other skills (maybe keepers, but they work differently so I wasn't thinking on them). That would definately break the proportional growing theory
(editado)

So you can check it yourself in that tool and treat it as "my" prediction and possibly prove me wrong this way :)

The tool "predicts" the end sumskill in function of final level. The linear regression only helps with the range (decimals).

Are you sure about that? From my experience it stays linear, and was linear in the old times when we had exact level known, but maybe I don't know about something.

Yes, i am sure its not linear. There is probably something about it in the rules too. But i think it uses the same logarithmic function as the skills of players in normal training (not junior training).

When youths play in youth league it stays pretty close to the predicted line.

Thats not what my experience is.

Hello All,
I know that if I write here, I will suffer the consequences of deleting my account.
Admin always said players must respect the game rules but admin did not respect it.
1- Admin locked my account without giving any proof.
2- My team has not been deleted yet, but the admin has sold my players, including players who are not members of NT.
I hope the community can speak up and explain to me.
My team: https://sokker.org/vi/app/team/85901/
Thanks alll...!
(editado)

Yeah, that part about putting your players on TL looks very fishy to me...