I Have Mathematical Proof That The Random Number Generator Is RIGGED

[rah]

Experience leads me to believe that it's not totally fixed in BTS. But as always, I could be wrong, but I generally win at odds less that 50% when multiple first strikes are involved. With drill III it seems like I win more that half the time at 33% But it might just be my memory. I like to remember happy things.

[mkorin]

Originally posted by Wiglaf
Also, Algebra II deals with elementary probability, permutations and such. If there is a 5% chance of winning a lottery with one ticket and I buy 5 tickets, I have 5*5 = 25% chance of winning. This is called a permutation, which is a fancy word for multiplying, and I took a class on this in high school by Mr Stephen Wright who would be happy to sign up and respond to your insults personally.

Wig, One of my degrees is in Statistics and you frankly don't understand it if you think you should multiply 5 battles to prove the results. The difference between your example above and battles is that the battles are not dependent on each other. Battle 1 is X%, Battle 2 is X% etc. Battle 2 is not dependent upon whether Battle 1 was won or loss. Whereas a lottery ticket is. If one ticket is a loser than the next ticket has slightly better odds because one possible outcome was removed. It is the difference between permutations and combinations. I'm sure you teacher can explain it too you better than I'm willing to do it in this post. I hope that's clear.

As to being paranoid, remember even paranoids have enemies.

Lastly name one court jester where there is historical evidence that they were brilliant thinkers. They were most likely like the kid who doesn't fit in, but does stupid things to sit at the cool table of your high school. Look around tomorrow at lunch and I'm sure you'll see what I mean, unless it is you.

[Wiglaf]

I'm not in high school but have vivid memories or it

Please see: Mulla Nasrudin, the legendary Sufi mystic of the 14th century who was a jester of Tamerlane. Or Jeffrey Hudson.

Also, if I'm right, then there is no way for a first strike to be calculated in the percentage calculator without running thousands of trials which the game clearly does not do. Am I right on this?

The difference between your example above and battles is that the battles are not dependent on each other. Battle 1 is X%, Battle 2 is X% etc. Battle 2 is not dependent upon whether Battle 1 was won or loss

[mkorin]

Originally posted by Wiglaf
I'm not in high school but have vivid memories or it

Please see: Mulla Nasrudin, the legendary Sufi mystic of the 14th century who was a jester of Tamerlane. Or Jeffrey Hudson.

Also, if I'm right, then there is no way for a first strike to be calculated in the percentage calculator without running thousands of trials which the game clearly does not do. Am I right on this?

Sorry, you said fresh out of Algebra II, so I assumed.

2 Jesters in history doesn't make a trend. I'd bet that most were simpletons just trying to get along.

I'd also bet that the first strike has a value assigned to it which it uses for the check. This would also lead to slight inconsistencies with the percentage, but not enough to knock the long term percentages off by a significant amount.

[Krill]

[q=Wiglaf]Also, if I'm right, then there is no way for a first strike to be calculated in the percentage calculator without running thousands of trials which the game clearly does not do. Am I right on this?[/q]

If this i how it works, or unless the numbers of stuck on the HDD, yeah...

[Krill]

Please remember this post for the future...

[Harrier UK]

Originally posted by rah
Hmmm so you're saying that if there are 10 tickets sold for a drawing where If I buy 1 ticket (10% chance of winning) that if I buy 4 tickets my chances of winning aren't 40%, (10X4)

In the UK, this example would be called a raffle - not a lottery.

The chance of winning in a raffle depends on the total number of tickets sold.

Assuming there is one prize only, if 1000 tickets are purchased and you by 1 ticket. The odds are 999 -1.
If you buy 10 of those tickets the odds become 99-1. (Some may argue it should be 990-1)

In a lottery the odds of winning are based on the mathmatical probability of a sequence of numbers appearing from all the possibilities. It has nothing to do with the number of tickets sold or the sequences of numbers chosen by the players.

In the UK the chance of winning the lottery is 14.000.000 -1 (approx). But not that many tickets are bought for each draw.

[TriMiro]

To compute the FS influence on the chance of win/loose, then you only need to consider cases equal to the number of FS plus one. With 2 FS, consider 2 successful ones, 1 successful one and 0 successful one. Consider the odds for each of the above cases and then consider the regular odds of winning after each of the cases. Multiply and add.

It is perfectly doable on a modern CPU, unless there are 1000 first strikes, there would not be 1000 cases to consider.

[VoiceOfUnreason]

Originally posted by TriMiro
To compute the FS influence on the chance of win/loose, then you only need to consider cases equal to the number of FS plus one. With 2 FS, consider 2 successful ones, 1 successful one and 0 successful one. Consider the odds for each of the above cases and then consider the regular odds of winning after each of the cases. Multiply and add.

The last time I looked, the calculator code was doing precisely this, with the additional twist that it needs to deal with First Strike Chances as well.

[nbarclay]

Originally posted by rah
Hmmm so you're saying that if there are 10 tickets sold for a drawing where If I buy 1 ticket (10% chance of winning) that if I buy 4 tickets my chances of winning aren't 40%, (10X4)

There is an important difference between the mathematics used with independent events (where the outcome of one random event has no impact on the outcome of another) and the mathematics used with dependent events (where the outcome of one random event does have an impact on another). Consider the following illustration.

Suppose you get the ace through ten of diamonds out of a deck of cards and use them as a means of selecting a number between one and ten. If you draw a card, put it back, reshuffle, and draw another card, the two draws are independent of each other. The fact that you got a particular card in one draw makes no difference in the odds regarding what card you will get in the next draw.

But now suppose you draw two cards out of the deck at once. The two draws are not independent because the cards are guaranteed to be different from each other. That makes a significant difference in how the odds are calculated. For example, with two independent draws, the odds of getting the ace in both draws are one in a hundred. But drawing two cards at once, it is physically impossible to get the ace both times.

In a lottery where the mechanism for choosing the winning number is independent of the number of tickets sold (hence not guaranteeing that there will be a winner at all unless a ticket is sold for every possible outcome), it is proper to use the mathematics of dependent events. If you buy five lottery tickets with different numbers, your odds of winning are five times as good (albeit still lousy) because you've covered five times as many of the possible outcomes.

But if you buy tickets in five different lotteries, the mathematics are different and more complicated. The odds of winning at all are slightly less than if you bought five tickets in the same lottery. But in return, you get a tiny possibility of winning more than once.

In Civ IV, the random number generator is designed to produce numbers that act as if they were completely independent of each other. Thus, winning or losing one battle has no impact on the odds of winning or losing the next (except of course for the impact of injuries). That makes it inevitable that there will be occasional lucky streaks and unlucky streaks.

Also note that the problem of calculating combat odds in Civ IV is actually a much more serious challenge than the problem of producing good "random" numbers. There have been good algorithms for generating "random" numbers with computers for decades. In contrast, even right after I passed my graduate-level class in Combinatorics (the field of math that deals with calculating numbers of possible combinations of outcomes and the use of that in calculating probabilities), I would have viewed the problem of calculating combat odds in Civ IV as a very serious challenge. So if careful, extensive testing would demonstrate a clearly significant discrepancy between the combat odds we are given and the actual patterns of results, I would be much more inclined to suspect a flaw in the code for calculating the odds than a flaw in the random number generator. (Actually, just knowing the complexity of the mathematics involved makes me highly suspicious of whether the odds calculations are always reliable.)