Problems with Die Rolls disparity.

[RobertBr]

So you must have seen on a regular and I do mean regular basis, disparities of a 2.73 Average compared to a 4.93 Average over 15 die rolls, not 2 or 3 or even 5 rolls but 15 rolls...

Yes, because that is how independent probabilities work, and you have seen it as well if you have genuinely spent 35 years playing these games. Play a euro if this bothers you - this is how hex and counter wargames work.

And it is very hard to believe that you don't like posting about this when you won't listen to people who do understand the maths. This happens on the Backgammon forums as well, it is not helpful or useful to anybody else for you to make up (and the problem does not exist, these are random results) a problem and then persist in labouring the point even if you are told the problem doesn't exist.

Worse, you are focusing on the combat dice rolls (and it appears only ones that are bad for you) which are not even a serious game issue if you are not keen on independent probabilities (which some people aren't, which is why euro games exist) rather than the weather rolls, which do have game play issues.

[steve-perkins]

Yes I have seen disparities in my yes 35 years of wargaming, (Although you seem to think I'm lying about that) But they are very rare, they don't continue over more than half a dozen die rolls. You seem to think that the evidence I've presented of actual die rolls is irrelevant (I can give you the game number if you think I'm making them up as well)

The Random die roll system is perfect, it never needs checking, it always gives perfect results, that's your opinion, my experience of it, is it's flawed and can be made better, so it doesn't become the over riding focus of game like a referee who dishes out yellow and red cards like confetti.

All I'm saying is the random die roll generator can be programmed to recognise when it's gone terribly wrong and rectify it. It sounds like the same one is in use in Backgammon, with the same outcry from players there.

I do know what hex and counter games are about (Maybe you don't so I'll remind you) and they are about planning and strategy and tactical opportunities, what they are not about is a die roll system that gives one side a 2.93 average roll and the other side a 4.93 average roll. I can't begin to plan anything if I can't rely on the random die rolls being, oh I don't know, random.

It's so easy to fix, just program a flag into the system to self check and re-roll when things have gotten too one sided, we don't have to have games ruined. Games that last a long time that players have put hours of their time into. It's something that a game called Gettysburg 125th anniversary edition tried to do, by having a re-roll chit, once you used it you passed it to your opponent who then could use it in a future roll and pass it back to you.

My memory is not what it once was, but I think I'd remember a 2.93 to a 4.93 die roll advantage over 15 rolls and no, there as never been anything remotely like that. (But again I guessing you think I'm lying about that too)

I also guessing that you'll ignore message this too, your die roll system is so perfect, so flawless that no one should ever criticise it.

[Lunalol]

Like I said in a previous post:

What will I do to "fix" random problem ?
...hum... let me remember my answer...

Oh yes ! NOTHING.
You could write pages and pages an this forum, I will never do some WRONG mathematics to fix some WRONG feeling.

Sorry, but randomness makes life hard

[steve-perkins]

I'll make one final comment on this profoundly flawed random die roll system. The one that is so perfect and so flawless that nothing could be wrong with it. (King's new clothes)

My two opponents are fed up with getting all those successes on the dice, because they can see it makes the game meaningless. They'll easily win, but with no merit to show for it, I can't provide any opposition to them.

The latest 'random' rolls are as follows: Me 3,3,4,3,5,4 which is great, a total of 22 and an average of 3.6 However, my opponents rolls are: 6,5,5,6,5,6. a total of 33 for an average of 5.5! if this was a one off rare example, then fine, frustrating, but fine. It is unfortunately not a one off, it's happened in 3 of the 4 competitive games I'm currently involved in and happened more than once in those games.

Sadly once those games are finished (Because I won't give up in them) I'll not be playing this excellent game again online unless I hear that someone as looked into the randomness factor and perhaps done something to mitigate it.

[jonathanmcc]

Intuitively, at the end I felt I had suffered more from the rolls, the worst being the Germans had use of 6 fair weather turns in 1941.

But in general wargamers have lousy ability to judge probability and tend to hyper-focus on rolls that went against them. In this case you mention the 1941 fair weather turns (I presume you were playing a historical opening and the Soviets collapsed in 42; if the Germans took 90 NW in one summer offensive you don't get to blame the dice for that) but you don't mention how many fair weather turns you got as the allies in the endgame - suggesting you noticed the adverse results more strongly.

The mistake people make with probability, is they think past outcomes have an impact on the next one. "It will all balance up".
The chance of rolling 6 is still 1in 6 no matter how many '1s in a row you rolled prior to that.
Now if my die rolled ten 1's in a row I would start to suspect a rigged die, so in this case people are questioning the computerised roller.
That's human nature.

I followed your earlier thread about the high impact of the weather rolls. I have been debating it with my son (& opponent).

With 2 turns to go the Germans still had 41 will points left and we were both of the opinion that even with 2 fair turns (the last turn is always fair),
the Allies would not win. There was still a lot riding on that 2nd last weather roll though and it turned out Poor which meant not a lot could be achieved. So I had an 'unlucky' roll there also - I just didn't mention it above.

In the end I has a strong last turn on both fronts and reduced the Germans to 24 will by the end. So if the second last turn has been fair, it would have been much closer.

In our discussion, we didn't see any way to solve this 'do or die' weather roll issue. One idea I considered was to add an extra turn to the game if the overall weather rolls favoured the Germans past a certain threshold, but this doesn't seem practical.

So my approach is to be philosophical about it - if you lose on the the last meaningful weather roll, then you played a good close game.

[Romain672]

So it look like you are focusing too on a specific player doing a lot of rolls in a row.
Here is the maximum amount of 5 or 6 I got in a row in 600 rolls (which look like the duration of your game). I did 250 series:

Series of 6: 107 (42.8%)
Series of 7: 38 (15.2%)
Series of 8: 17 (6.8%)
Series of 9: 9 (3.6%)
Series of 10: 4 (1.6%)
Series of 11: 1 (0.4%)
Series of 12: 1 (0.4%)

What surprise me from your series is the amount of (5 or 6 in a row) is really low. 8 is still unlikely but possible, but I got none of those.
The weirdness would begin at 11+. We are really far from that.

And btw, you was really hard to understand on that spot. I have no idea if you are telling randomness is broken, or if it was a suggestion and wanted weird random to be changed.

[jonathanmcc]

Just wondering if you used the same random number generator that the game uses.

I think Lunalol said in an earlier thread that he was using the in built SQL function in the database.

(If there was a problem with that, it would affect presumably all the games on BGA, and all the other applications worldwide that use it ).

[Romain672]

I'm using microsoft excel and their alea().
I suppose I could calculate the probabilities, and find similar numbers

[Lunalol]

Just wondering if you used the same random number generator that the game uses.
I think Lunalol said in an earlier thread that he was using the in built SQL function in the database.
(If there was a problem with that, it would affect presumably all the games on BGA, and all the other applications worldwide that use it ).

Precision : in MySQL, "ORDER BY RAND" is just used when you want to pick a unit/marker at random.
All USE dice rolls use a framework build-in function:

Dice and bga_rand
bga_rand( min, max ) This is a BGA framework function that provides you a random number between "min" and "max" (inclusive), using the best available random method available on the system.
This is the preferred function you should use, because we are updating it when a better method is introduced.
As of now, bga_rand is based on the PHP function "random_int", which ensures a cryptographic level of randomness.
In particular, it is mandatory to use it for all dice throw (ie: games using other methods for dice throwing will be rejected by BGA during review).
Note: rand() and mt_rand() are deprecated on BGA and should not be used anymore, as their randomness is not as good as "bga_rand".

[jonathanmcc]

>> The latest 'random' rolls are as follows: Me 3,3,4,3,5,4 which is great, a total of 22 and an average of 3.6 However, my opponents rolls are: 6,5,5,6,5,6.

The other interesting thing to note is that these two sequences were probably interleaved as typically both sides roll once in an engagement.
This the sequence of results from the generator was probably more like:

3,6,3,5,4,5,3,6,4,5,4,6