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?
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.
Posts: 181
Joined: 3/22/2008 From: Massachusetts, 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!!
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?
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?
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.
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.
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, 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.
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'
Posts: 654
Joined: 4/9/2007 From: East Coast 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.
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.
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?
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?
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?
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.