I Have Mathematical Proof That The Random Number Generator Is RIGGED

[Dale]

Can't people just accept that "**** happens"? Or are you complaining because out of your 100 winning battles you had a single loss? I think the later is more the case with the OP.

Some battles go your way, some don't. Just accept it.

I've wasted 5 units on 1, and I've seen the AI waste 5 units on 1. So it's fair.

[Krill]

[q=TriMiro]Dumb Example: Pikemen vs Horse Archer, no promotions. Strengths 6 and 6, but P has +100% modifier, thus the odds are 12 vs 6 for P, or in other words there is a 66% chance that P will win the first round vs 33% that HA will win. If we get the 1/3 odds and HA does win, then the strength of P would be reduced to say 5.4 (I am not sure about the exact number). Then 5.4 would get modified by +100% and the odds for the second round would be 10.8 vs 6 for P. In other words there is a 64% chance that P would win vs 36% chance that HA would win. Repeat until one unit dies.[/q]

Actually no, the strength of the units are taken before the first round, and then are fixed for the rest of the battle, not altered after each individual round like you say.

[TriMiro]

OK so I don't know how combat works then. Since the strength is modified by the health if the unit is wounded in the beginning of the combat, it only seemed logical to recompute the strength each round, but I guess I am wrong.

In that case computing the win/loose odds becomes very simple, I don't see why they cannot incorporate the FS mechanics into it.

[couerdelion]

A few points about statistical tests.

a) Calculating something with a low probability of happening does not prove that the original assumption (eg that the chances are random) is false. It merely shows that it is unlikely. In effect you are “post priori” stating that the likelihood of this consecutive five losses is 1 in 3,200,000. It’s not really a proof even if we all agree that it might amount to something akin to proof.

b) When doing these sorts of statistical test, you have to set up the test before you measure it. If you have 5000 combats in each game and play a series of 20 games then you’ve already got a total of 100,000 combats (or 20,000 sequences of 5 combats). Even if these were random, there will be one set of combats that will be a 1 in 20,000 event (or 0.005%). However, if we test our randomness theory after that event we would EXPECT to get a statistical probability of around 0.005% which would be significant proof if it were an unbiased test.

Or to put another way, it is LIKELY that we will get an UNLIKELY result.

[dschur]

Originally posted by couerdelion
Or to put another way, it is LIKELY that we will get an UNLIKELY result.

This is exactly correct. One of the sure signs of "human tampering" in results where you expect random distributions is the lack of seemingly unexpected runs.

For example, if we have two sets of 10 coin tosses, recording heads or tails and record the results:

HHTHHHTTTT and HTHTHTHHTH , the first one is what you actualy expect from real results, the second is faked (full disclosure, both are faked, but I just wanted to make the point...most places I'm sure I could get away with not making this disclusure, but here someone might pull out SAS or SPSS on me :-)

[Seedle]

Originally posted by rah
Please note that first strike is not calcualted in the odds.
Quite noticable when you have 4th level drill units that consisitently win even attacking at suposed 33%.
I had one stack of 8 of them, with none showing more than 30-32% chance, and all 8 won. So you must keep first strike in mind when you read the %

This is not true, the odds given are affected by first strikes. I went into world builder, had 5 units, all with combat 1+2. Had (1)Longbowman with drill 1+2, (2)Longbowman with just combat, (3)Swordsman with just combat, (4)Swordsman with drill 1, (5)Swordsman with drill 1+2. Saw what odds I was given for attacking a spearman on grassland.

1=99.1, 2=98.2, 3=97.1, 4=97.6, 5=98.5

So first strikes and even chances at first strikes affected the odds. Whether they affected them correctly, I don't know, but first strike is certainly calculated with the odds.

Also, I don't know if this is correct or not, but I believe I read in a thread on here along time ago that combat strength IS recalculated after the entire first strike phase of combat. Not every round, not every first strike, just once after all first strikes are completed. E.G., a maceman attacking a maceman, all combats will have both units roll on whatever str they had at the beginning, a mace vs. a samurai, the first strikes will be rolled with beginning str, if the mace took any hits after all first strikes are calculated, the mace will roll with lowered str from that point on. Can anyone confirm or deny this?

[Krill]

IIRC there was a bug where calculations involving FS were way out in warlords that got corected in BtS, but I've never seen any proper tests to calculate if the FS calculations were working as intended.

[Robert]

Soldiers that have to fight against the odd get a 500% power bonus! Didn't you see 300?

[TriMiro]

Originally posted by CyberShy
Soldiers that have to fight against the odd get a 500% power bonus! Didn't you see 300?

I cannot resist to quote Terry Pratchet:

one in a million happens in nine out of ten times.

[mkorin]

Re: I Have Mathematical Proof That The Random Number Generator Is RIGGED

Originally posted by Wiglaf
Let me break this down for you, fresh from algebra II.

A 95% chance to win 5 times means a (95*5) chance to win = 475% chance to win. Therefore my stack of pikemen being destroyed by a single horse archer makes no sense and defies logic. My study is better than previous ones, which did not have a stack of the same unit with the same experience doing the same trial.

I believe whoever designed the RNG is also behind the apostolic palace. Stay tuned

Algebra is not what you use to prove statistics. That is a whole different branch of mathematics and obviously you don't understand it. Nor am I going to waste time trying to explain it to you. Back in Civ IV original, I put in a statistical proof that the RNG was accurate then. If you search for it, you should be able to find it. I will say in BTS the RNG looks not random because the attacker and defender are not nescecarily the ones that were used in the estimate when you hovered over the upcoming battle. If you use a one to one fight, it is clear the RNG is correct. If the designers would put back the odds during the fight, I would be happy to show proof that this RNG is correct.

Mike

[wodan11]

You guys realize that what Wiglaf did was drop his logic bomb and then sat back with popcorn to watch the fur fly? He's quite intelligent; there's no way this wasn't intentional.

Wodan

[Supr49er]

Originally posted by wodan11
You guys realize that what Wiglaf did was drop his logic bomb and then sat back with popcorn to watch the fur fly? He's quite intelligent; there's no way this wasn't intentional.

Wodan

[mkorin]

Originally posted by wodan11
You guys realize that what Wiglaf did was drop his logic bomb and then sat back with popcorn to watch the fur fly? He's quite intelligent; there's no way this wasn't intentional.

Wodan

Based on his plethora of idiotic posts, if you believe that I question your intelligence along with his.

[wodan11]

Originally posted by mkorin


Based on his plethora of idiotic posts, if you believe that I question your intelligence along with his.

You're doing him a disservice. I think he's fully aware of how idiotic he comes off.

Wodan

[mkorin]

Originally posted by wodan11

You're doing him a disservice. I think he's fully aware of how idiotic he comes off.

Wodan

It boggles my mind to think that someone would know how idiotic he sounds and continue to do so. Or, in fact do it purposely.

Mike