Matrix Games Forums

A new update for Piercing Fortress EuropaNew screenshots for War in the West!Pike & Shot is now available!Server Maintenance Battle Academy 2 gets updated!Deal of the Week: Advanced Tactics Gold Ask Buzz Aldrin!Pike & Shot gets Release Date and Twitch Session!Deal of the Week Espana 1936War in the West coming in December!
Forums  Register  Login  Photo Gallery  Member List  Search  Calendars  FAQ 

My Profile  Inbox  Address Book  My Subscription  My Forums  Log Out

better leader - combat results

 
View related threads: (in this forum | in all forums)

Logged in as: Guest
Users viewing this topic: none
  Printable Version
All Forums >> [Current Games From Matrix.] >> [Napoleonics] >> Empire in Arms the Napoleonic Wars of 1805 - 1815 >> better leader - combat results Page: [1] 2   next >   >>
Login
Message << Older Topic   Newer Topic >>
better leader - combat results - 4/3/2009 4:41:00 AM   
borner


Posts: 1485
Joined: 3/20/2005
From: Houston TX
Status: offline
I was playing a game against the AI, and have seen two battles now where France under Nappy attacks a Prussian and then Austrian stack with no leader. (Yes, this was intentional.) The Austrians rolled 3 zeros. 0(-1). Prussia never rolled above a 1. I also have seen a live game early on where Nappy hammered the Prussians with equal chits ( counter attack vs Esc Counterattack), in that one Prussia had a bad 1805 leader, and did not roll higher than a 1. France lost 5 factors, Prussia over 40. Again the dice were terrible. I am no stats expert, but what are the odds or rolling a 1 three times in a row?

the game is supposed to have a max +1/-1 mod to the total die rolls. However, this really seems like a trend. Has anyone else seen very out of balance die rolls where one leader was so much better than another?

Post #: 1
RE: better leader - combat results - 4/3/2009 7:30:08 AM   
mr.godo

 

Posts: 61
Joined: 4/19/2008
Status: offline
one in 216. but not impossible.

try posting all the results from a game.

(in reply to borner)
Post #: 2
RE: better leader - combat results - 4/3/2009 7:43:38 AM   
mr.godo

 

Posts: 61
Joined: 4/19/2008
Status: offline
open your logfile. parse for "roll". find all the numbers rolled for.

i just played a solo game with 2406 rolls from 1 to 20. This is interesting.

Count of Roll
Roll   1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Grand Total
Total 541 217 219 213 211 200 131 139 136 166 23 26 13 27 18 16 27 30 29 24 2406


This is the *basic* unmodified roll. Why does "1" show up so often???


< Message edited by mr.godo -- 4/3/2009 7:46:07 AM >

(in reply to mr.godo)
Post #: 3
RE: better leader - combat results - 4/3/2009 12:51:26 PM   
NeverMan

 

Posts: 1716
Joined: 2/24/2004
Status: offline
I'd be pretty surprised if Marshall used anything other than a simple "rand()%" function for this but then again, maybe I wouldn't be.

(in reply to mr.godo)
Post #: 4
RE: better leader - combat results - 4/3/2009 1:08:37 PM   
borner


Posts: 1485
Joined: 3/20/2005
From: Houston TX
Status: offline
one in 216, then how about 9 rolls all below 2 before modifications, to account for what I have seen up to this point in other games? this suuuure sounds like a bug; as in the system giving too much credit / traking too much away for leadership differences. by the way - I re-ran that turn and the same thing happened again.

(in reply to NeverMan)
Post #: 5
RE: better leader - combat results - 4/3/2009 5:25:02 PM   
obsidiandragon


Posts: 171
Joined: 3/22/2008
From: Michigan, USA
Status: offline
I have actually brought better leaders and 2 or 3 times the number of troops the AI will have and still LOST battles due to the fondness of the 1 on the dice...  So it is not tied to leaders or troops, my guess is it just likes to roll ones!!


(in reply to borner)
Post #: 6
RE: better leader - combat results - 4/3/2009 6:22:22 PM   
mr.godo

 

Posts: 61
Joined: 4/19/2008
Status: offline
Reviewing combat rolls would be problematic since you have to check and record each one. But if you parse through the game log, you can search each line for a forage roll. I have looked at three different games as well as two faked games and it looks like you will get a 1 twice as likely as any other number. I fail to understand why this would not be an even distribution unless the value of the roll is not 1d6 for foraging with modifications afterwards.

if there is a rolling problem with foraging, i would guess that battle rolls would be compromised as well. anyone not having a problem with die rolls?

(in reply to obsidiandragon)
Post #: 7
RE: better leader - combat results - 4/3/2009 7:43:22 PM   
Jimmer

 

Posts: 1968
Joined: 12/5/2007
Status: offline

quote:

ORIGINAL: mr.godo

open your logfile. parse for "roll". find all the numbers rolled for.

i just played a solo game with 2406 rolls from 1 to 20. This is interesting.

Count of Roll
Roll   1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Grand Total
Total 541 217 219 213 211 200 131 139 136 166 23 26 13 27 18 16 27 30 29 24 2406


This is the *basic* unmodified roll. Why does "1" show up so often???


Because the program's random function is not random. I reported this problem almost a year ago, and everybody who responded said it was because of a failing in the logging or my math. I still believe it is a problem, especially with forage rolls. 1s come up >10 times as often as they should.

_____________________________

At LAST! The greatest campaign board game of all time is finally available for the PC. Can my old heart stand the strain?

(in reply to mr.godo)
Post #: 8
RE: better leader - combat results - 4/3/2009 8:30:22 PM   
mr.godo

 

Posts: 61
Joined: 4/19/2008
Status: offline
there is obviously no math involved here.

(in reply to Jimmer)
Post #: 9
RE: better leader - combat results - 4/3/2009 8:40:52 PM   
NeverMan

 

Posts: 1716
Joined: 2/24/2004
Status: offline
It would be nice for Marshall to chime in since I'm not sure how this could be a problem.

If he's simply seeding a rand function and calling that function, there's no reason for it to come up 1s 10x more often than any other number.

If he tried to implement his own psuedo-random algorithm then yes, it's probably all jacked up.

(in reply to mr.godo)
Post #: 10
RE: better leader - combat results - 4/4/2009 1:06:17 AM   
easterner

 

Posts: 178
Joined: 10/9/2004
Status: offline
When does need for d10 & d20 come up?

< Message edited by easterner -- 4/4/2009 1:10:37 AM >

(in reply to NeverMan)
Post #: 11
RE: better leader - combat results - 4/4/2009 9:09:00 AM   
Ted1066


Posts: 214
Joined: 12/10/2007
From: Vancouver, BC, Canada
Status: offline
where is zero in all these rolls? Is it being lumped in with 1? I'm wondering if the log file is attributing a zero in combat as a 1 in the log. If there were an even split between 0 and 1 for the 541 1 rolls (i.e ~270 each), that is not an abnormal distribution (it has variance of roughly 5-10% and is less than 1 standard deviation - not bad by stats rules IF 0's are being counted as 1's in the log).

As to the 10 to 20, that shows up in reference to the minor diplomacy rules I think. You have to roll higher on d10 plus your bonuses vs the comp rolling a d20.

Cheers,

Ted

(in reply to easterner)
Post #: 12
RE: better leader - combat results - 4/4/2009 2:54:44 PM   
mr.godo

 

Posts: 61
Joined: 4/19/2008
Status: offline
that must be when the die falls off the table. i always roll into box lids to prevent that from happening.

0-6 is technically 1d7. i have seen 0-7 for combat. that would be 1d8. the rules make no mention of these varied dice substitutions.

i am sure there is an explanation somewhere for the inconsistency.


(in reply to Ted1066)
Post #: 13
RE: better leader - combat results - 4/4/2009 4:30:58 PM   
borner


Posts: 1485
Joined: 3/20/2005
From: Houston TX
Status: offline
this would be something I would suggest people keep an eye out for. the total adjustment to the roll is supposed to be +1/-1. sure seems that the system wants to take it further.

(in reply to mr.godo)
Post #: 14
RE: better leader - combat results - 4/4/2009 4:42:03 PM   
NeverMan

 

Posts: 1716
Joined: 2/24/2004
Status: offline
Zero should not be a problem.

The random function probably looks something like: (rand()%6)+1, with the rand() function being seeded by time something like: srand(time(NULL)).

I've very curious as to how Marshall decided to implement this.

(in reply to borner)
Post #: 15
RE: better leader - combat results - 4/5/2009 4:58:44 AM   
Ted1066


Posts: 214
Joined: 12/10/2007
From: Vancouver, BC, Canada
Status: offline

quote:

ORIGINAL: mr.godo

that must be when the die falls off the table. i always roll into box lids to prevent that from happening.

0-6 is technically 1d7. i have seen 0-7 for combat. that would be 1d8. the rules make no mention of these varied dice substitutions.

i am sure there is an explanation somewhere for the inconsistency.



mr.godo

0-7 occurs from a 1d6 die roll being modified by +/-1, depending on combat circumstances (good vs bad leaders mainly, but cav superiority can supply a side with +1).

Ted

(in reply to mr.godo)
Post #: 16
RE: better leader - combat results - 4/5/2009 2:49:41 PM   
borner


Posts: 1485
Joined: 3/20/2005
From: Houston TX
Status: offline
Ture, but the the max a roll is supposed to be able to be modified is +1/-1

(in reply to Ted1066)
Post #: 17
RE: better leader - combat results - 4/5/2009 5:16:35 PM   
mr.godo

 

Posts: 61
Joined: 4/19/2008
Status: offline
ted,
so the combat reports the modified die roll, not some funky 0-7 1d8? it is purely the reporting of a modded die roll. the original roll is 1d6 based. but you suggest the forage could have an unmodified 0 die roll? the numbers i pulled from the log are unmodified. why would there be a 0-6 1d7 roll?

if there was a log addition for the battle results, we could see for ourselves if there was a pattern in the die rolling.



(in reply to borner)
Post #: 18
RE: better leader - combat results - 4/5/2009 6:29:55 PM   
easterner

 

Posts: 178
Joined: 10/9/2004
Status: offline
I would assume the mod is ignored in calculation otherwise 7 would be skewered result. 2-6, 7-10, 11-20 are all within probability of one another with only '10' a little high  +30 over 7-9.  But nothing like '1' 

(in reply to mr.godo)
Post #: 19
RE: better leader - combat results - 4/6/2009 1:48:02 PM   
Mardonius


Posts: 653
Joined: 4/9/2007
From: Watertown, NY
Status: offline

quote:

ORIGINAL: Jimmer


quote:

ORIGINAL: mr.godo

open your logfile. parse for "roll". find all the numbers rolled for.

i just played a solo game with 2406 rolls from 1 to 20. This is interesting.

Count of Roll
Roll   1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Grand Total
Total 541 217 219 213 211 200 131 139 136 166 23 26 13 27 18 16 27 30 29 24 2406


This is the *basic* unmodified roll. Why does "1" show up so often???


Because the program's random function is not random. I reported this problem almost a year ago, and everybody who responded said it was because of a failing in the logging or my math. I still believe it is a problem, especially with forage rolls. 1s come up >10 times as often as they should.


There was an issue in earlier versions with autoforages being reported as "1"s. Moreover, I have recently seen forages that were manually done but assured being reported as "1"s. Therefore, to get a representative data set is cumbersome. One must extract all autoforages or automatic success forages. This data set will give a result much closer to the expected mean.

Also, please note that the above data set extracted from Mr. Godo's sample takes input from multiple variables (1d6, 1d10, 2d6, and 1d20) each with variable modifiers so can not return a statistically viable result unless parsed out. Time consuming, but doable.

(in reply to Jimmer)
Post #: 20
RE: better leader - combat results - 4/6/2009 2:01:34 PM   
Marshall Ellis


Posts: 5626
Joined: 10/2/2001
From: Dallas
Status: offline
FYI: Not all rolls are rolls. Mardonius is correct. Autoforages were always being reported as "1s" so tracking all rolls would be difficult and would reveal little since they are not all rolls.


_____________________________

Thank you

Marshall Ellis
Outflank Strategy War Games



(in reply to Mardonius)
Post #: 21
RE: better leader - combat results - 4/6/2009 3:46:07 PM   
NeverMan

 

Posts: 1716
Joined: 2/24/2004
Status: offline

quote:

ORIGINAL: Marshall Ellis

FYI: Not all rolls are rolls. Mardonius is correct. Autoforages were always being reported as "1s" so tracking all rolls would be difficult and would reveal little since they are not all rolls.



Well, this doesn't seem like a good idea since it makes possible bugs harder to decode. Why not make them "-1" or some value specific to "auto forage" so that we (your playtesters) can help you debug this game better?

(in reply to Marshall Ellis)
Post #: 22
RE: better leader - combat results - 4/6/2009 4:49:02 PM   
Marshall Ellis


Posts: 5626
Joined: 10/2/2001
From: Dallas
Status: offline
Actually, I believe they are being reported as "Autoforages" in 1.06 so this should go away.

_____________________________

Thank you

Marshall Ellis
Outflank Strategy War Games



(in reply to NeverMan)
Post #: 23
RE: better leader - combat results - 4/6/2009 9:15:48 PM   
mr.godo

 

Posts: 61
Joined: 4/19/2008
Status: offline
fine. the log data is crap and cannot be used for analysis. a die roll may be reported where no roll is made or it may be modified or unmodified. makes sense. one more place to avoid.

so, can anyone say definitively whether this game can roll dice fairly?

(in reply to Marshall Ellis)
Post #: 24
RE: better leader - combat results - 4/6/2009 10:48:33 PM   
Thresh

 

Posts: 393
Joined: 12/25/2006
From: KCMO
Status: offline
quote:

so, can anyone say definitively whether this game can roll dice fairly?


Seems to be par for the course where most computerized die rollers are concerned...

Thresh

(in reply to mr.godo)
Post #: 25
RE: better leader - combat results - 4/6/2009 11:32:31 PM   
no_dice

 

Posts: 41
Joined: 1/12/2007
Status: offline
Hmm, sounds like the "random" number system they use in Las Vegas. :)

(in reply to Thresh)
Post #: 26
RE: better leader - combat results - 4/7/2009 1:47:00 AM   
Marshall Ellis


Posts: 5626
Joined: 10/2/2001
From: Dallas
Status: offline

quote:

ORIGINAL: mr.godo

fine. the log data is crap and cannot be used for analysis. a die roll may be reported where no roll is made or it may be modified or unmodified. makes sense. one more place to avoid.

so, can anyone say definitively whether this game can roll dice fairly?



Yes. I can :-)


_____________________________

Thank you

Marshall Ellis
Outflank Strategy War Games



(in reply to mr.godo)
Post #: 27
RE: better leader - combat results - 4/7/2009 1:56:36 AM   
NeverMan

 

Posts: 1716
Joined: 2/24/2004
Status: offline
mr. godo,

You should have a better look at what's going on in 1.06 when the numbers aren't all skewed by the extra 1s. You can then take the log and find out the percentages for each number over a large dataset.

(in reply to Marshall Ellis)
Post #: 28
RE: better leader - combat results - 4/7/2009 3:22:11 PM   
Marshall Ellis


Posts: 5626
Joined: 10/2/2001
From: Dallas
Status: offline
That would reveal more accurate results.


_____________________________

Thank you

Marshall Ellis
Outflank Strategy War Games



(in reply to NeverMan)
Post #: 29
RE: better leader - combat results - 4/7/2009 6:10:27 PM   
mr.godo

 

Posts: 61
Joined: 4/19/2008
Status: offline
looking forward to it. am working strictly on solo games as the pace for multi is too slow for me and i enjoy shellacking the ai.

(in reply to Marshall Ellis)
Post #: 30
Page:   [1] 2   next >   >>
All Forums >> [Current Games From Matrix.] >> [Napoleonics] >> Empire in Arms the Napoleonic Wars of 1805 - 1815 >> better leader - combat results Page: [1] 2   next >   >>
Jump to:





New Messages No New Messages
Hot Topic w/ New Messages Hot Topic w/o New Messages
Locked w/ New Messages Locked w/o New Messages
 Post New Thread
 Reply to Message
 Post New Poll
 Submit Vote
 Delete My Own Post
 Delete My Own Thread
 Rate Posts


Forum Software © ASPPlayground.NET Advanced Edition 2.4.5 ANSI

0.117