Subpage under development, new version coming soon!
Subject: Fix randoms in Sokker [bug]
Sure. Age and Number of draws are major random criteria.
Weeks, talent, levels are also important.... But at 18y old we can always fired (whatever talents, weeks...) and off course for 1 to 6 draws other criteria's doesn't matter for draws that don't exist.
So if we remove 18y old and set draws to minimum of 3 to 6 (or 2 to 4 whatever) it should removed at least very poor draws and reduce inequity without adding/touching anything else in code (even pseudo random). It's poor man solution but much better than current state.
(edited)
Weeks, talent, levels are also important.... But at 18y old we can always fired (whatever talents, weeks...) and off course for 1 to 6 draws other criteria's doesn't matter for draws that don't exist.
So if we remove 18y old and set draws to minimum of 3 to 6 (or 2 to 4 whatever) it should removed at least very poor draws and reduce inequity without adding/touching anything else in code (even pseudo random). It's poor man solution but much better than current state.
(edited)
Having zero 16y old for 13 consecutive draws is (2/3)^13 it's 0,33 % of chance. Very improbable.
First, it's 0.51%
Second, you can't just look at the probability of the last 13 draws now that it happened to you. It could have occurred at any time during the last seasons, there were more opportunities for this to happen than just in the most recent draws. For example, if we look at the probability of this streak occurring somewhere in the last two seasons (26 weeks), it would be ~4.85%. That's not a rare event - about one in every 20-21 teams (with a youth academy and 6 empty slots for youths each week) should experience such streak over the last two seasons. If you could show that it happens to significantly more teams than expected, then you'd have a case.
As it stands, you've proven nothing and all the nonsense you've written about prng doesn't matter.
First, it's 0.51%
Second, you can't just look at the probability of the last 13 draws now that it happened to you. It could have occurred at any time during the last seasons, there were more opportunities for this to happen than just in the most recent draws. For example, if we look at the probability of this streak occurring somewhere in the last two seasons (26 weeks), it would be ~4.85%. That's not a rare event - about one in every 20-21 teams (with a youth academy and 6 empty slots for youths each week) should experience such streak over the last two seasons. If you could show that it happens to significantly more teams than expected, then you'd have a case.
As it stands, you've proven nothing and all the nonsense you've written about prng doesn't matter.
My bad, 0,51%. Should i change to "very probable" ?
Glad to know this streak could happen in 26 weeks with a more descent probability, but what a surprise.
If anyone is having trouble figuring what is 0.5%, I've quickly write up a simulation for 3 seasons of draws :
https://www.sktables.org/rand.php
Yes, its base on my own view, maybe I just catch one rare streak, but I see this kind of patterns in forum so much.
Other examples ? 7 away games (junior league) in a row. Another one, just today : 5 magic trainers in a row.
That's off topic (whatever a "bug") but this few single series are already "absurd" by itself.
So yes, for sure it needs data compilation and study. I'll pass.
Please, what is "all the nonsense you've written about prng" exactly ?
Is it correct that pseudo-rand are by essence not random ? is it correct than for a given seed will can always produce the exact same sequence of “random” numbers ?
(edited)
Glad to know this streak could happen in 26 weeks with a more descent probability, but what a surprise.
If anyone is having trouble figuring what is 0.5%, I've quickly write up a simulation for 3 seasons of draws :
https://www.sktables.org/rand.php
Yes, its base on my own view, maybe I just catch one rare streak, but I see this kind of patterns in forum so much.
Other examples ? 7 away games (junior league) in a row. Another one, just today : 5 magic trainers in a row.
That's off topic (whatever a "bug") but this few single series are already "absurd" by itself.
So yes, for sure it needs data compilation and study. I'll pass.
Please, what is "all the nonsense you've written about prng" exactly ?
Is it correct that pseudo-rand are by essence not random ? is it correct than for a given seed will can always produce the exact same sequence of “random” numbers ?
(edited)
Imo lifting the ceilings of skills with age is the better way to solve things in youth.
16 very good
17 excellent
18 formidable
19 outstanding
20 incredible
21 brilliant
16 very good
17 excellent
18 formidable
19 outstanding
20 incredible
21 brilliant
Why not. It will produce only and much good draws.
For me, the main problem is not typically having good draws, but having very poor draws (!) and so few draws a week, when others could catch the exact opposite in the meantimes.
As an aside, it's more time consuming to change the code (to generate a min output level) than to simply change the random ranges.
For me, the main problem is not typically having good draws, but having very poor draws (!) and so few draws a week, when others could catch the exact opposite in the meantimes.
As an aside, it's more time consuming to change the code (to generate a min output level) than to simply change the random ranges.
My bad, 0,51%. Should i change to "very probable" ?
No, you should put correct numbers in your comments instead of some random ones; later you managed to get it wrong again by writing "1-(2/3)^13. Around 99,55 % of chance" ...how?
Glad to know this streak could happen in 26 weeks with a more descent probability, what a surprise.
If it happened to you few weeks earlier you would also write it here on forum claiming some 0.X% probability, when in reality the odds of this happening is much higher if you give it more chances. If I roll a dice until I get three sixes in a row should I say that the dice is broken because something with 0.46% probability happened?
Another recent streak I v see : 7 away games (junior league) in a row. | 7 aways in a row, really ?
Perfect example of you not understanding randomness. Actually something that can be easily checked, something that could prove your theory about pseudo-random generation being bad... but it doesn't.
3011 teams play all 12 youth school matches this season. If the match host is chosen randomly you would expect about 1/2^7 = 1/128 of these teams playing the last 7 matches away.
3011 * 1/128 = ~23.52
Real data: there are 21 teams with the last 7 youth matches away. Very reasonable number. It's the same with your "Zero 16y in 5 weeks", it's the same with "5 magic trainers". Things with <1% probability will happen if you give it thousands of tries. Nothing extraordinary is happening, that's how randomness should work. It has nothing to do with pseudo-random generation.
Look, here is a team with all 12 youth matches away this season! Something with 1/4096 (0.0244%) probability happened to one of the 3011 teams! Is it wrong that it happened?
https://sokker.org/app/season/schedule/56436/
So yes, for sure it needs data compilation and study. I'll pass.
You don't care about data, you don't care about math, you slap some inadequate probability number to your sentence and you think something is wrong, when in reality the thing with youths should randomly happen to 150+ teams in the last 2 seasons.
Please, what is "all the nonsense you've written about prng" exactly ?
So you can write more nonsense in the response? Thanks. I'll pass.
EOT from me.
No, you should put correct numbers in your comments instead of some random ones; later you managed to get it wrong again by writing "1-(2/3)^13. Around 99,55 % of chance" ...how?
Glad to know this streak could happen in 26 weeks with a more descent probability, what a surprise.
If it happened to you few weeks earlier you would also write it here on forum claiming some 0.X% probability, when in reality the odds of this happening is much higher if you give it more chances. If I roll a dice until I get three sixes in a row should I say that the dice is broken because something with 0.46% probability happened?
Another recent streak I v see : 7 away games (junior league) in a row. | 7 aways in a row, really ?
Perfect example of you not understanding randomness. Actually something that can be easily checked, something that could prove your theory about pseudo-random generation being bad... but it doesn't.
3011 teams play all 12 youth school matches this season. If the match host is chosen randomly you would expect about 1/2^7 = 1/128 of these teams playing the last 7 matches away.
3011 * 1/128 = ~23.52
Real data: there are 21 teams with the last 7 youth matches away. Very reasonable number. It's the same with your "Zero 16y in 5 weeks", it's the same with "5 magic trainers". Things with <1% probability will happen if you give it thousands of tries. Nothing extraordinary is happening, that's how randomness should work. It has nothing to do with pseudo-random generation.
Look, here is a team with all 12 youth matches away this season! Something with 1/4096 (0.0244%) probability happened to one of the 3011 teams! Is it wrong that it happened?
https://sokker.org/app/season/schedule/56436/
So yes, for sure it needs data compilation and study. I'll pass.
You don't care about data, you don't care about math, you slap some inadequate probability number to your sentence and you think something is wrong, when in reality the thing with youths should randomly happen to 150+ teams in the last 2 seasons.
Please, what is "all the nonsense you've written about prng" exactly ?
So you can write more nonsense in the response? Thanks. I'll pass.
EOT from me.
later you managed to get it wrong again by writing "1-(2/3)^13. Around 99,55 % of chance" ...how?
Probably approx. to 0.66 instead of exact 2/3.
Probably approx. to 0.66 instead of exact 2/3.
You r right, those stats make sense globally.
But even loaded dices can reproduce a normal random distribution globally, It doesn't prove that dices are loaded for series, or not. Ultimately, source code can tell.
Loaded or not, if this kind of events can happen statically to a "very reasonable number" of teams, well, in term of game design it's reasonably absurd.
If 0,5% is "not so bad" (compare to this 0.0244% I feel lucky) we can imagine than few other rare teams can get zero 16y during a whole season, when many others will get 16y each weeks.
Loaded or not, distribution of randoms all the way (injuries, juniors...) without any limit of absurd series its certainly the first thing that should be fixed.
(edited)
But even loaded dices can reproduce a normal random distribution globally, It doesn't prove that dices are loaded for series, or not. Ultimately, source code can tell.
Loaded or not, if this kind of events can happen statically to a "very reasonable number" of teams, well, in term of game design it's reasonably absurd.
If 0,5% is "not so bad" (compare to this 0.0244% I feel lucky) we can imagine than few other rare teams can get zero 16y during a whole season, when many others will get 16y each weeks.
Loaded or not, distribution of randoms all the way (injuries, juniors...) without any limit of absurd series its certainly the first thing that should be fixed.
(edited)
it was still a reference to 0,33% (which was my first, incorrect), I suppose ?
Anyway the reverse of very low prob is very high probs, so when I have something like 99% chance to get at least one 16y it seems fair of having none at 1%
(edited)
Anyway the reverse of very low prob is very high probs, so when I have something like 99% chance to get at least one 16y it seems fair of having none at 1%
(edited)
13 draws! I hope :-)
EDIT : :
This week again,
Luis Platini 18 6 (adequate)
Matteo Le Brun 17 6 (adequate)
15 draws and 6 weeks in a row 2/3)^15 :-D
(edited)
EDIT : :
This week again,
Luis Platini 18 6 (adequate)
Matteo Le Brun 17 6 (adequate)
15 draws and 6 weeks in a row 2/3)^15 :-D
(edited)
Well then the chance of getting someone isnt that small..... you were just unlucky....
No one wants anything older than 18, at any point so I see this particular suggestion as pointless. Players coming out at late 19 or later have so much real training loss they are useless.
If they are triple brilliant at age 21 on the striker skills... not so pointless I guess...
Just to make an example...
Just to make an example...
I've purchased a divine pull at early age 20 (week 3-4, I don't recall). He didn't amount to anything important when I sold him at 25.
There would have to be a massive improvement in youth pulls, and a total revamp of the youth rating scale for your scenario to possibly occur.
There would have to be a massive improvement in youth pulls, and a total revamp of the youth rating scale for your scenario to possibly occur.