 |
View Old Forum Thread
Old Forum Index » Sim Forums » Baseball Game & Site Questions » IC distribution
Frunobulax
I've asked this in the past but haven't got an answer so far. So I'll post this again, both in support and in the Questions section.
I must say that I'm pretty disappointed about the reaction from support. I posted this before in [url=http://www.simdynasty.com/oldforum-viewthread.jsp?tid=409561#pid2164689]Support/url], and [url=http://www.simdynasty.com/oldforum-viewthread.jsp?tid=399149#pid2117782]here[/url] - the promised investigation never happend, and one answer to my IC distrib question was "Keep in mind that the rulebook is editable by the users. If you'd like to make changes, go ahead." Yeah, thank you very much, if I know the answer to my questions then I wouldn't have to ask, and of course I would have edited the wiki.
Issue 1: The IC assignment depeding on the CPs is broken if less than the maximal number of CPs is assigned, and broken badly.
As for system 2, I have set up several trial leagues to test this, and came to the same conclusion in every leage: If I assign less than 20 CPs, then sometimes the CPs are distributed completely at random. I set up a league where I assigned just 10 CPs, 5 on a pitcher and 5 on a position player. After the season finished, all players had about the same amount of ICs - the player with 5 CPs had actually less ICs than the average player.
In system 5 I didn't experiment myself since I didn't want to mess up one of my teams, but saw other players do something similar (assign 13, 14 and 15 CPs and 0 CPs to other players), and it appeared they didn't get a lot of ICs either.
I would assume that the code wants to make sure that a player with X CPs can't get more ICs than he's "supposed to". So in system 2, a player with 5 CPs would not get more ICs by deliberatly assigning less CPs than available. However, there must be a bug in there. (And I'd gladly have a look at the code myself if I could get access to it.)
Issue 2: The description of the IC distribution is wrong in the documentation, and anybody with a pocket calculator can figure this out. I just don't know what the correct distribution is, otherwise I would fix the docs. Of course, maybe there's a bug in there too.
Let's start with System 5.
This is in the docs: "After each game, each team will receive 4 Improvement Chances (ICs). Three improvement chances are awarded based on the number of CPs you have on the player. The 4th IC has an equal chance of being awarded to any player on your minor league roster. "
Then follows a table that lists the ICs for a 15-CP player at 69, and a 1-CP player at 21. The fun thing about this table: If you add up all values you come out at 600 ICs per season, or 3.7 per game. And don't assume rounding problems, if we increase all values by exactly 0.999 then we would still have only 615 ICs, less than the true value of 624. So this table is wrong of course.
Now let's do the math. 162 ICs distributed randomly, that's 10.8 per player. 486 ICs distributed according to the CPs, 120 total CPs, that's about 4.05 ICs per CP. So we have a 1-CP player at 14.8 chances a season and a 15-CP player at 71.5 chances. That's not only not consistent with the table given, but I've played a lot of system 5 leagues (because that's my favorite minor league system) and have hardly ever seen a player get less than 20 ICs if he was in the minors for a year. [I'll assume of course throughout this article that there are 15 or less players in the minors.)
So what's the true value? I thought maybe 2 chances per game are assigned randomly, and 2 depending on CP. Same math, 324 chances randomly, 2.7 per CP. That would make the lower bound plausible (24.3 ICs for a 1-CP player), the upper bould would end up at 62.1 ICs for a 15-CP player. The latter seems a bit low, I've seen players accumulate over 75 ICs per season quite often, but this may be just random fluctuation. Of course, this is not at all consistent with the CP table given.
In System 2 I would guess the docs are simply wrong from the beginning to the end. They read "There is also a single CP automatically assigned by ABE to each player on the minor league roster.", and "Of the four possible ICs available per game, the first two ICs are available for all players in the minors; the third is only for players with one or no CPs assigned; the fourth is for position changes only.".
Now, let's do the math again. If this was true, we would have 35 total CPs (15 invisible), 324 ICs distributed according to the CPs, 162 ICs distributed on players with 0 or 1 CP, and 54 mystery ICs (we have an extra IC every in third game that we don't know how they are assigned).
So we'd have 324/35 = 9.25 ICs per CP, where CPs range to 1 to 6. Or more likely 378/35 10.8 ICs per CP if we assume the mystery ICs are assigned that way too. That would give a 5-CP player (6 total CPs) about 64 chances. So far, so good. However, a 3-CP player would have 43 ICs and a 2-CP player would have only 32 ICs. That meshes neither with the table listed (49 ICs for 3 CPs, 41 ICs for 2 CPs) nor with my experience.
My best guess is that 1 IC per game is assigned depending the 20 CPs (forget about the invisible stuff), and the remaining 378 ICs are assigned at random - that would give each player 25.4 ICs, plus 8.1 ICs per CP. That makes 65.7/49.5/41.4 ICs for 5/3/2 CPs. This is close enough to my experience that I would assume that this is the correct formula.
.f
IC distribution
November 08, 2015 at 02:22PM View BBCode
Hi,I've asked this in the past but haven't got an answer so far. So I'll post this again, both in support and in the Questions section.
I must say that I'm pretty disappointed about the reaction from support. I posted this before in [url=http://www.simdynasty.com/oldforum-viewthread.jsp?tid=409561#pid2164689]Support/url], and [url=http://www.simdynasty.com/oldforum-viewthread.jsp?tid=399149#pid2117782]here[/url] - the promised investigation never happend, and one answer to my IC distrib question was "Keep in mind that the rulebook is editable by the users. If you'd like to make changes, go ahead." Yeah, thank you very much, if I know the answer to my questions then I wouldn't have to ask, and of course I would have edited the wiki.
Issue 1: The IC assignment depeding on the CPs is broken if less than the maximal number of CPs is assigned, and broken badly.
As for system 2, I have set up several trial leagues to test this, and came to the same conclusion in every leage: If I assign less than 20 CPs, then sometimes the CPs are distributed completely at random. I set up a league where I assigned just 10 CPs, 5 on a pitcher and 5 on a position player. After the season finished, all players had about the same amount of ICs - the player with 5 CPs had actually less ICs than the average player.
In system 5 I didn't experiment myself since I didn't want to mess up one of my teams, but saw other players do something similar (assign 13, 14 and 15 CPs and 0 CPs to other players), and it appeared they didn't get a lot of ICs either.
I would assume that the code wants to make sure that a player with X CPs can't get more ICs than he's "supposed to". So in system 2, a player with 5 CPs would not get more ICs by deliberatly assigning less CPs than available. However, there must be a bug in there. (And I'd gladly have a look at the code myself if I could get access to it.)
Issue 2: The description of the IC distribution is wrong in the documentation, and anybody with a pocket calculator can figure this out. I just don't know what the correct distribution is, otherwise I would fix the docs. Of course, maybe there's a bug in there too.
Let's start with System 5.
This is in the docs: "After each game, each team will receive 4 Improvement Chances (ICs). Three improvement chances are awarded based on the number of CPs you have on the player. The 4th IC has an equal chance of being awarded to any player on your minor league roster. "
Then follows a table that lists the ICs for a 15-CP player at 69, and a 1-CP player at 21. The fun thing about this table: If you add up all values you come out at 600 ICs per season, or 3.7 per game. And don't assume rounding problems, if we increase all values by exactly 0.999 then we would still have only 615 ICs, less than the true value of 624. So this table is wrong of course.
Now let's do the math. 162 ICs distributed randomly, that's 10.8 per player. 486 ICs distributed according to the CPs, 120 total CPs, that's about 4.05 ICs per CP. So we have a 1-CP player at 14.8 chances a season and a 15-CP player at 71.5 chances. That's not only not consistent with the table given, but I've played a lot of system 5 leagues (because that's my favorite minor league system) and have hardly ever seen a player get less than 20 ICs if he was in the minors for a year. [I'll assume of course throughout this article that there are 15 or less players in the minors.)
So what's the true value? I thought maybe 2 chances per game are assigned randomly, and 2 depending on CP. Same math, 324 chances randomly, 2.7 per CP. That would make the lower bound plausible (24.3 ICs for a 1-CP player), the upper bould would end up at 62.1 ICs for a 15-CP player. The latter seems a bit low, I've seen players accumulate over 75 ICs per season quite often, but this may be just random fluctuation. Of course, this is not at all consistent with the CP table given.
In System 2 I would guess the docs are simply wrong from the beginning to the end. They read "There is also a single CP automatically assigned by ABE to each player on the minor league roster.", and "Of the four possible ICs available per game, the first two ICs are available for all players in the minors; the third is only for players with one or no CPs assigned; the fourth is for position changes only.".
Now, let's do the math again. If this was true, we would have 35 total CPs (15 invisible), 324 ICs distributed according to the CPs, 162 ICs distributed on players with 0 or 1 CP, and 54 mystery ICs (we have an extra IC every in third game that we don't know how they are assigned).
So we'd have 324/35 = 9.25 ICs per CP, where CPs range to 1 to 6. Or more likely 378/35 10.8 ICs per CP if we assume the mystery ICs are assigned that way too. That would give a 5-CP player (6 total CPs) about 64 chances. So far, so good. However, a 3-CP player would have 43 ICs and a 2-CP player would have only 32 ICs. That meshes neither with the table listed (49 ICs for 3 CPs, 41 ICs for 2 CPs) nor with my experience.
My best guess is that 1 IC per game is assigned depending the 20 CPs (forget about the invisible stuff), and the remaining 378 ICs are assigned at random - that would give each player 25.4 ICs, plus 8.1 ICs per CP. That makes 65.7/49.5/41.4 ICs for 5/3/2 CPs. This is close enough to my experience that I would assume that this is the correct formula.
.f
Admin
I just added up all of the numbers in that table and got 648 ICs per season. Here is the table from the rules:
CP ICs
15 69
14 67
13 65
12 61
11 56
10 51
9 46
8 41
7 37
6 33
5 29
4 27
3 25
2 23
1 21
648/162 games = 4 ICs per game. CPs are not linear so you can't really do a ICs per CP figure.
I'm still looking at the other parts of the post.
Chris
[Edited on 11-8-2015 by Admin]
November 08, 2015 at 08:42PM View BBCode
I am going to answer this in several pieces rather than try to research all of the answers in a single reply.Originally posted by Frunobulax
Issue 2: The description of the IC distribution is wrong in the documentation, and anybody with a pocket calculator can figure this out. I just don't know what the correct distribution is, otherwise I would fix the docs. Of course, maybe there's a bug in there too.
Let's start with System 5.
This is in the docs: "After each game, each team will receive 4 Improvement Chances (ICs). Three improvement chances are awarded based on the number of CPs you have on the player. The 4th IC has an equal chance of being awarded to any player on your minor league roster. "
Then follows a table that lists the ICs for a 15-CP player at 69, and a 1-CP player at 21. The fun thing about this table: If you add up all values you come out at 600 ICs per season, or 3.7 per game. And don't assume rounding problems, if we increase all values by exactly 0.999 then we would still have only 615 ICs, less than the true value of 624. So this table is wrong of course.
Now let's do the math. 162 ICs distributed randomly, that's 10.8 per player. 486 ICs distributed according to the CPs, 120 total CPs, that's about 4.05 ICs per CP. So we have a 1-CP player at 14.8 chances a season and a 15-CP player at 71.5 chances. That's not only not consistent with the table given, but I've played a lot of system 5 leagues (because that's my favorite minor league system) and have hardly ever seen a player get less than 20 ICs if he was in the minors for a year. [I'll assume of course throughout this article that there are 15 or less players in the minors.)
So what's the true value? I thought maybe 2 chances per game are assigned randomly, and 2 depending on CP. Same math, 324 chances randomly, 2.7 per CP. That would make the lower bound plausible (24.3 ICs for a 1-CP player), the upper bould would end up at 62.1 ICs for a 15-CP player. The latter seems a bit low, I've seen players accumulate over 75 ICs per season quite often, but this may be just random fluctuation. Of course, this is not at all consistent with the CP table given.
I just added up all of the numbers in that table and got 648 ICs per season. Here is the table from the rules:
CP ICs
15 69
14 67
13 65
12 61
11 56
10 51
9 46
8 41
7 37
6 33
5 29
4 27
3 25
2 23
1 21
648/162 games = 4 ICs per game. CPs are not linear so you can't really do a ICs per CP figure.
I'm still looking at the other parts of the post.
Chris
[Edited on 11-8-2015 by Admin]
Frunobulax
D'oh! Seems I added up wrong. (I end up at 651, but this can be according to rounding.)
Other than that, why can't I do an IC per CP figure? No idea how you assign the ICs, I figured it's just a weighted chance which would make that linear?
Again, if you send me the relevant part of the code then I'll be happy to have a look at it. I used to be a software developer, so I guess I should be able to make something out of it (and write a decent description for the Wiki).
Regards, F.
November 10, 2015 at 09:31AM View BBCode
Originally posted by Admin
I just added up all of the numbers in that table and got 648 ICs per season. Here is the table from the rules:
CP ICs
15 69
14 67
13 65
12 61
11 56
10 51
9 46
8 41
7 37
6 33
5 29
4 27
3 25
2 23
1 21
648/162 games = 4 ICs per game. CPs are not linear so you can't really do a ICs per CP figure.
D'oh! Seems I added up wrong. (I end up at 651, but this can be according to rounding.)
Other than that, why can't I do an IC per CP figure? No idea how you assign the ICs, I figured it's just a weighted chance which would make that linear?
Again, if you send me the relevant part of the code then I'll be happy to have a look at it. I used to be a software developer, so I guess I should be able to make something out of it (and write a decent description for the Wiki).
Regards, F.
Mongrel
That list is pretty much the way it works, think of a lottery drawing with 648 balls, 69 balls have the number 15, 67 balls have the number fourteen etc. The RNG picks a ball and rewards the player in that slot with an IC.
But we've seen evidence that if all players are not assigned CP's the entire thing breaks, so I'm more interested in how that will be addressed.
November 10, 2015 at 02:50PM View BBCode
Originally posted by Frunobulax
Originally posted by Admin
I just added up all of the numbers in that table and got 648 ICs per season. Here is the table from the rules:
CP ICs
15 69
14 67
13 65
12 61
11 56
10 51
9 46
8 41
7 37
6 33
5 29
4 27
3 25
2 23
1 21
648/162 games = 4 ICs per game. CPs are not linear so you can't really do a ICs per CP figure.
D'oh! Seems I added up wrong. (I end up at 651, but this can be according to rounding.)
Other than that, why can't I do an IC per CP figure? No idea how you assign the ICs, I figured it's just a weighted chance which would make that linear?
Again, if you send me the relevant part of the code then I'll be happy to have a look at it. I used to be a software developer, so I guess I should be able to make something out of it (and write a decent description for the Wiki).
Regards, F.
That list is pretty much the way it works, think of a lottery drawing with 648 balls, 69 balls have the number 15, 67 balls have the number fourteen etc. The RNG picks a ball and rewards the player in that slot with an IC.
But we've seen evidence that if all players are not assigned CP's the entire thing breaks, so I'm more interested in how that will be addressed.
cdunn3
what effect does this have ?
November 10, 2015 at 02:58PM View BBCode
Another question - if only 13 are assigned (3-15)what effect does this have ?
Frunobulax
Really? I assumed that it's a lottery dependent on the CPs. In System 2, a lottery with 20 balls and therefore a 5-CP player has a 25% to receive any given IC (that is assigned depending on the CPs, and not randomly). In system 5 a lottery with 120 balls with 15 balls having the number of the player with 15 CPs.
[Edited on 11-10-2015 by Frunobulax]
November 10, 2015 at 03:38PM View BBCode
Originally posted by Mongrel
That list is pretty much the way it works, think of a lottery drawing with 648 balls, 69 balls have the number 15, 67 balls have the number fourteen etc. The RNG picks a ball and rewards the player in that slot with an IC.
But we've seen evidence that if all players are not assigned CP's the entire thing breaks, so I'm more interested in how that will be addressed.
Really? I assumed that it's a lottery dependent on the CPs. In System 2, a lottery with 20 balls and therefore a 5-CP player has a 25% to receive any given IC (that is assigned depending on the CPs, and not randomly). In system 5 a lottery with 120 balls with 15 balls having the number of the player with 15 CPs.
[Edited on 11-10-2015 by Frunobulax]
Mongrel
My description was wrong, as I forgot the fourth point is supposed to have an equal weight. Yours would seem more accurate, but if it's a 1:1 ratio 15 CPs should average 71.55 IC and 1 CP should average 14.85 CP so it's not as simple as 120 balls.
So the fourth IC drawing in system 5 should give 10.8 CPs to each player in a season. So subtract that from the values in the tables and either we round or multiply the ball count by 10 to remove the decimal. .
If you multiply by 10 then there would be 582 balls numbered 15, 562 numbered 14 etc...
But again, we have seen that it breaks when not all the CP's are assigned. I've seen examples where the CP distribution seemed either totally random, or actually favoring players with 0 CP over 15 CP, So My biggest question is why is it broken, and when is it going to be fixed. If I have a roster move that leaves a CP slot open, does it reduce the chances of my best players getting CP's ?
November 10, 2015 at 04:17PM View BBCode
Originally posted by Frunobulax
Originally posted by Mongrel
That list is pretty much the way it works, think of a lottery drawing with 648 balls, 69 balls have the number 15, 67 balls have the number fourteen etc. The RNG picks a ball and rewards the player in that slot with an IC.
But we've seen evidence that if all players are not assigned CP's the entire thing breaks, so I'm more interested in how that will be addressed.
Really? I assumed that it's a lottery dependent on the CPs. In System 2, a lottery with 20 balls and therefore a 5-CP player has a 25% to receive any given IC (that is assigned depending on the CPs, and not randomly). In system 5 a lottery with 120 balls with 15 balls having the number of the player with 15 CPs.
[Edited on 11-10-2015 by Frunobulax]
My description was wrong, as I forgot the fourth point is supposed to have an equal weight. Yours would seem more accurate, but if it's a 1:1 ratio 15 CPs should average 71.55 IC and 1 CP should average 14.85 CP so it's not as simple as 120 balls.
So the fourth IC drawing in system 5 should give 10.8 CPs to each player in a season. So subtract that from the values in the tables and either we round or multiply the ball count by 10 to remove the decimal. .
If you multiply by 10 then there would be 582 balls numbered 15, 562 numbered 14 etc...
But again, we have seen that it breaks when not all the CP's are assigned. I've seen examples where the CP distribution seemed either totally random, or actually favoring players with 0 CP over 15 CP, So My biggest question is why is it broken, and when is it going to be fixed. If I have a roster move that leaves a CP slot open, does it reduce the chances of my best players getting CP's ?
Frunobulax
I prefer to split the ICs in randomly assigned and CP-assigned.
For an individual game it would work like this: For any IC there is a lottery on which player it goes. There are random ICs, with all players having the same chance, and weighted ICs (depending on CPs), where the chances are weighted by the CP weight. But I'm really not sure if this is the way it really works. For system 2 the numbers make sense. For system 5 too, if we assume that two chances per game are random and two are IC-based. But in all cases the numbers do not concur with the tables given in the docs.
This seems to be the case.
.f
November 10, 2015 at 06:42PM View BBCode
Originally posted by Mongrel
My description was wrong, as I forgot the fourth point is supposed to have an equal weight. Yours would seem more accurate, but if it's a 1:1 ratio 15 CPs should average 71.55 IC and 1 CP should average 14.85 CP so it's not as simple as 120 balls.
So the fourth IC drawing in system 5 should give 10.8 CPs to each player in a season. So subtract that from the values in the tables and either we round or multiply the ball count by 10 to remove the decimal. .
If you multiply by 10 then there would be 582 balls numbered 15, 562 numbered 14 etc...
I prefer to split the ICs in randomly assigned and CP-assigned.
For an individual game it would work like this: For any IC there is a lottery on which player it goes. There are random ICs, with all players having the same chance, and weighted ICs (depending on CPs), where the chances are weighted by the CP weight. But I'm really not sure if this is the way it really works. For system 2 the numbers make sense. For system 5 too, if we assume that two chances per game are random and two are IC-based. But in all cases the numbers do not concur with the tables given in the docs.
Originally posted by MongrelBut again, we have seen that it breaks when not all the CP's are assigned. I've seen examples where the CP distribution seemed either totally random, or actually favoring players with 0 CP over 15 CP, So My biggest question is why is it broken, and when is it going to be fixed. If I have a roster move that leaves a CP slot open, does it reduce the chances of my best players getting CP's ?
This seems to be the case.
.f
Mongrel
I made a spreadsheet of what I described above, and when it rounds off, it gives the same results as the guide:
[url]https://docs.google.com/spreadsheets/d/152EpuC43O7ekqBFr7Qync39GaF_WLvXkz0nJfH_B-SM/edit?usp=sharing[/url]
One thing to realize is as the 'balls' are removed it changes the odds on the next drawing. So I first calculated as if the first 3 IC chances were all drawn from a pool of 4890 balls, and it was off by quite a bit. So in my sheet I simulate three drawings, removing a weighted portion of the balls from each 'slot' and one ball total from the total pool for each step. So the first IC is drawn from the full pool. the second has fifteen less balls, the third has 30 less balls.
[Edited on 11-10-2015 by Mongrel]
[Edited on 11-10-2015 by Mongrel]
November 10, 2015 at 07:24PM View BBCode
Only one chance is random in system 5I made a spreadsheet of what I described above, and when it rounds off, it gives the same results as the guide:
[url]https://docs.google.com/spreadsheets/d/152EpuC43O7ekqBFr7Qync39GaF_WLvXkz0nJfH_B-SM/edit?usp=sharing[/url]
One thing to realize is as the 'balls' are removed it changes the odds on the next drawing. So I first calculated as if the first 3 IC chances were all drawn from a pool of 4890 balls, and it was off by quite a bit. So in my sheet I simulate three drawings, removing a weighted portion of the balls from each 'slot' and one ball total from the total pool for each step. So the first IC is drawn from the full pool. the second has fifteen less balls, the third has 30 less balls.
[Edited on 11-10-2015 by Mongrel]
[Edited on 11-10-2015 by Mongrel]
Frunobulax
However, I wonder if it really works this way.
And I suggest we all start pestering Admin if they don't check out this potentially serious bug. I've only reported it 3 times so far.
[Edited on 11-13-2015 by Frunobulax]
November 13, 2015 at 08:20PM View BBCode
Well, yeah, essentially you can get any distribution this way. It would be completely equivalent to have 69 balls for the player with 15 CPs, 21 balls for the player with 1 CPs. The final result is a bit below because the sum of the 15 values is 3 more than the "desired" sum of 648.However, I wonder if it really works this way.
And I suggest we all start pestering Admin if they don't check out this potentially serious bug. I've only reported it 3 times so far.
[Edited on 11-13-2015 by Frunobulax]
bahstonwedsawks
I agree it should be fixed, I just don't think it's a pressing concern.
November 13, 2015 at 08:36PM View BBCode
I understand your concern, Frunobulax. However, I doubt it's seen as a serious concern by the simple fact that the bug is rendered ineffective rather easily. An owner needs to assign all IC's available to them and keep the exact number of prospects that should be in the minors. The owners it is impacting (lack of effort to either not assign all IC's or have too many/too few players) likely aren't too hung up on the details of the game in the first place.I agree it should be fixed, I just don't think it's a pressing concern.
Frunobulax
I think about speed leagues with 15 games a day and 15 players with CPs (system 5). It happens rather often that players are traded or promoted as a consequence of a trade and up to 15 games pass before the owner makes a move.
November 13, 2015 at 08:50PM View BBCode
You think about Dynasty league with only 5 players with CPs.I think about speed leagues with 15 games a day and 15 players with CPs (system 5). It happens rather often that players are traded or promoted as a consequence of a trade and up to 15 games pass before the owner makes a move.
bahstonwedsawks
You think about Dynasty league with only 5 players with CPs.
I think about speed leagues with 15 games a day and 15 players with CPs (system 5). It happens rather often that players are traded or promoted as a consequence of a trade and up to 15 games pass before the owner makes a move. [/quote]
Good point. For the leagues with a ton of games per day this definitely needs to be fixed.
November 13, 2015 at 11:23PM View formatted
You are viewing the raw post code; this allows you to copy a message with BBCode formatting intact.
[quote][i]Originally posted by Frunobulax[/i]
You think about Dynasty league with only 5 players with CPs.
I think about speed leagues with 15 games a day and 15 players with CPs (system 5). It happens rather often that players are traded or promoted as a consequence of a trade and up to 15 games pass before the owner makes a move. [/quote]
Good point. For the leagues with a ton of games per day this definitely needs to be fixed.
Pages: 1