Subject: »[info]Dev Diary #82: Automatic release of NT-players from

seems like the problem is the high number of players with form 0 even with played matches

Keep in mind that on that chart there are also players who played like 1 or 2 matches in (over) the last month, so it's not a surprise that some of them tend to have the lowest form.

---

Other things I found:
For some reason players who mainly play as MID tend to have lower form then players on other positions, no idea why (here it's shown in % of players from that position having each form):


When I checked form changes on Thursday few months ago players who played a match on Wednesday had slightly higher chance to gain a form than players who didn't play - this explain why players who play matches tend to have higher form than player who don't play


That's all I have.

Any chance you could pull the same data about playing on sunday/form update on monday?

I'll add it to the end of my list "things I want to do with sokker data, but I have other things to do than working for free", so don't expect it soon :P

Keep in mind that on that chart there are also players who played like 1 or 2 matches in (over) the last month, so it's not a surprise that some of them tend to have the lowest form.

My top players had 2 games every week and they stayed on low for the entire time!
Really...matches don't have any influence on form on a standard base. It's all random. But I think the trend for players to drop is bigger than the other way around, regardless of matches played.
I have 2 players in my team that almost never play.

Tuur De Clerck, age: 36, height: 185 cm
value: 460 750 €, wage: 5 400 €
club: Racing Roeland, country: België
form [17](+3) divine, tactical discipline [12](+1) outstanding

stamina [11] formidable, keeper [0] tragic
pace [13] incredible, defender [4] weak
technique [10](-1) excellent, playmaker [3] poor
passing [3] poor, striker [11] formidable

Tom Jordan, age: 38, height: 182 cm
value: 311 500 €, wage: 4 100 €
club: Racing Roeland, country: England
form [16](+2) unearthly, tactical discipline [13](+1) incredible

stamina [11] formidable, keeper [2] unsatisfactory
pace [7](-1) good, defender [8] solid
technique [9] very good, playmaker [8](-1) solid
passing [11] formidable, striker [4] weak

You can check their matches...and you can see how high their form is...

And this one plays 2 matches each week...

Noël Vandenhoeck, age: 22, height: 170 cm
value: 522 500 €, wage: 5 850 €
club: Racing Roeland, country: België
form [4](+3) weak, tactical discipline [14] brilliant

stamina [8] solid, keeper [0] tragic
pace [15] magical, defender [9] very good
technique [13] incredible, playmaker [10] excellent
passing [6] adequate, striker [10] excellent

With all due respect of your theory, but I just proved you wrong here...

So @raul you admit of game played affects the form with many games per week that obviously change affect what was before. Any plans for fix form some how ?

For me form is one of huge flaws of the game even you had great players with skills much weaker players play better if they are in form..

With all due respect of your theory, but I just proved you wrong here...

You proved nothing, individual cases are not statistically significant. Of course there are players who play a lot and have a low form and also there are players who play rarely and have a high form, you can see it in my previous charts.

Here is a distribution of form (again in % so it's easy to compare between groups) for players who had 1-5 matches, 6-10 matches, 11-15 matches played in the last ~6 weeks.

Is it possible to have stat for avg form of teams and if players that do not play help those that play to maintain high form?

seems like the problem is the high number of players with form 0 even with played matches

Back to this: there is a good mathematical reason why "0" columns are higher than "1", "2" etc. When you do a random walk on a 1 dimensional surface with limits from both sides (which is basically what form does, 1 dimension = form, random walk = form changes, limits = 0 and 18) then probabilities of being in extreme intervals of some length are higher than probabilities of being in middle intervals of the same length. I could create a stochastic model using Markov chains showing that, but I don't think it's worth it :P

You can see it happening at tragic form, but you could ask why it's not visible at the highest form - that's because it's hidden. At the bottom you have interval 0.00 - 0.9999999, but at the top it's divided into two columns: 17.00-17.9999999 and exactly 18.00 value, if you group them together it would be visible as a one column twice as big. If there was a "supertragic" skill at exactly 0.00 then it would divide that big column similarly.

---

@ Raul:
It looked like a Gaus curve, when there was just 1 league match per week. Terrion had some nice stats.
After the league reform the Gauss curve moved to the right.
I am sure you haven't changed the algorithm about the form and that's the reason why the curve moved from the middle to the right.

@ Mikos:
You wrote: "For some reason players who mainly play as MID tend to have lower form then players on other positions"
I know the reason.
Mostly this affects players who play on winger position.
I do not say, that it affects all players who play on winger position, but some players.
(the real problem is that their managers are not aware, that they can affect the form of their players)

And some more pieces of information:
Me and 3 more managers tried to figure out the algorithm of the form and we find some things out. Some of them are:
- Things that work for one player to boost his form, do not work for other players. It is a complex algorithm, with differently weighted parameters for every player
(differently weighted parameters = factors of the parameters can be dependent on various data of a player)
So, some players e.g. like to play 2 matches per week, some not. Some playrs like to play on winger position, some not. Etc etc.
- One of the factors of the parameters is for sure the age of the player. So, things that work for one player to boost his form on the current season, may not work on the next season.
So, one player can have this season high form if he plays 2 matches, but on the next season this can change.

Lol Dtox, mikoos is pulling data from the whole of SK, you can't compete with statistical data of this magnitude.

exceptions prove the rule

Too bad we don't know the rules of form :p
(edited)

Does the winger part of your theory have to do with the percentage a player plays offensive vs defensive?

Yes. Some plaers prefer high def, some high off.

"Too bad we don't know the rules of form"
And that's why I love sokker. It has his secrets and I try to figure out.
(and I also know, that I will never know all secrets of sokker...)

;)

One more thing. I wrote "Some players prefer high def, some high off."
It's not completely truth, because the percentages show the percentage of defensive or offensive actions of the whole team (100 % is the sum of all players).
On the algorithm probably it's not about the percentage, but it must be about the total minutes with the ball etc.

It would be good to keep some player history and stats for users.

It would be nice:)

And what about list of winners of Champions cup, national cup and first level league of every country, world champions of NT and U21?

sokker has this information but it´s wasted