I've had it with the biased random number generator

[Jayce]

I play Classic Iron man, and I've lost countless missions due to skewed RNG!

6 team members, all with 55% hit chance, all miss.... and this is a regular occurrence. I play Ironman so don't/can't reload, so every experience of this is a unique scenario. Has anyone else felt that that they miss more than to be expected.

[lucio1019]

numberless posts already mentioned this odd.
The lovers are holding their line to defense.

[EAPidgeon]

This is why I value heavys extremely high, along with other classes with deployables, although snipers can be useful if they have the Archangel Armor.

[Jayce]

The only posts regarding RNG that I have seen have been about the seeded RNG. Something about the outcome of an attack is already determined. However when it comes to Ironman this does not apply as reloading is not possible

[Gerald]

Please read this before entering endless whining mode:
link

[Demyan]

I´m also under the impression that the game doesn´t roll a random number - with a new seed - every time some1 fires a shot.
I´ve had plenty of situations were all missed (with not exactly the same hit chance but quite close), so I reloaded, did the same thing and it was miss miss miss all over again. But the moment you change something in your turn - and get a new seed for the RNG, suddenly all start to hit their targets.

So yes, it seems to me the seed generation for the RNG is broken, because it´s not renewed every time some1 fires a shot during a turn. Because I can´t imagine the programmers planned to use the same seed for all shots in a turn.....

And those fanboys who claim everything is perfect with their beloved game, well just ignore them.

[Poor Dick]

Honestly, people complain about missing 6 shots at 55%, but forget about all the long shots they take and hit. In my last playthrough, I had one round in a bomb defusal mission where I had only two rounds left and everyone was in a bad position to take shots or defuse the bomb without running though an alien killzone. I don't think I had a shot option higher than 30% and my squad nailed every last shot (6). Random is random: both good and bad.

[lucio1019]

I read it through the whole link carefully.
The paper analyzed how XCOM generated random number. It is good till the last part when it draw conclusion.

"To the casual observer there should not be much difference between the histogram plot from the XCOM: EU PRNG and the one generated using a modern high quality PRNG."

This is the only statement the writer made to support the hypotheses "XCOM EU has a reasonable random number system."
Do you notice something wrong here? There is no static prove, just a "casual observer". These 2 things are totally different.
And if you click all the links in the paper, you will notice, the way XCOM use to generate random number is "oldest and well-known".

[chriskrum]

55% chance to hit? That's a 45% chance to miss. I think the whiners failed math.

[Poor Dick]

Messaggio originale di lucio1019:

I read it through the whole link carefully.
The paper analyzed how XCOM generated random number. It is good till the last part when it draw conclusion.

"To the casual observer there should not be much difference between the histogram plot from the XCOM: EU PRNG and the one generated using a modern high quality PRNG."

This is the only statement the writer made to support the hypotheses "XCOM EU has a reasonable random number system."
Do you notice something wrong here? There is no static prove, just a "casual observer". These 2 things are totally different.
And if you click all the links in the paper, you will notice, the way XCOM use to generate random number is "oldest and well-known".

Out of curiosity, do you have any statistical evidence to backup your claim? Anecdotal evidence is great as a starting point for and investigation, however, it's mostly worthless. From my memory of experience, I percieve me making way more long-shot shots than should have been statistically possible - however, since I don't have any actually data on the matter (other than my memory), its largely useless (as is most of the anti-RNG threads I've seen).

If you are making the case the system is borked, its up to you to provide evidence that it is borked.

[N00bular]

Look, you're going to have bad luck sometimes. Some people will have streaks of bad luck. That's just RNGs for you, pseudo or not. If you have a hundred sets of six shots at 55%, you'll get a streak of six misses once, on average. Sometimes you'll get more than one. Sometimes those streaks will happen close together. One of the programmers even spent hours upon hours trying to figure out what the problem was with the biased PRNG when he had the same sort of situation.

The result he wound up with was that the PRNG was working just fine, and was not, in fact, biased; he simply happened to have a run of bad luck.

A lot of people play this game. Some people will have 'unusually' bad luck; the same number of people will have unusually good luck. The people with bad luck will come to the forum and complain that the RNG is biased. Today you happen to be that person. Sorry, I think I got your good luck. I'll try to give it back to you.

[lucio1019]

Messaggio originale di Poor ♥♥♥♥:

Messaggio originale di lucio1019:

I read it through the whole link carefully.
The paper analyzed how XCOM generated random number. It is good till the last part when it draw conclusion.

"To the casual observer there should not be much difference between the histogram plot from the XCOM: EU PRNG and the one generated using a modern high quality PRNG."

This is the only statement the writer made to support the hypotheses "XCOM EU has a reasonable random number system."
Do you notice something wrong here? There is no static prove, just a "casual observer". These 2 things are totally different.
And if you click all the links in the paper, you will notice, the way XCOM use to generate random number is "oldest and well-known".

Out of curiosity, do you have any statistical evidence to backup your claim? Anecdotal evidence is great as a starting point for and investigation, however, it's mostly worthless. From my memory of experience, I percieve me making way more long-shot shots than should have been statistically possible - however, since I don't have any actually data on the matter (other than my memory), its largely useless (as is most of the anti-RNG threads I've seen).

If you are making the case the system is borked, its up to you to provide evidence that it is borked.

That's why I stated in my post "I think".
The paper proved one of my suspicious "the system they use to produce random number is age old".
I keep my suspicious "system is broken" due to 2 reasons:
1. Without S/L it is almost not possible to collect data in actual game (I am not a programmer so I have no means to do it from code).
2. The odd event in game happens more than reasonable.

[Revolucas]

Oh bullcrap the RNG is fine:

http://steamcommunity.com/sharedfiles/filedetails/?id=103668868

I had an average of almost 80% hit percentage.

[chriskrum]

I think it's even more silly. The complaint isn't just that the RNG is biased, it's that it's biased against the player and always biased in a negative manner. No one is complaining about a streak of improbable hits or a streak of improbable misses on behalf of the aliens.

Any series of random numbers that is long enough will have streaks. If it doesn't it's not random.

[Demyan]

Well I´m a bit rusty at that kind of math but I´m pretty sure I remember how to calculate the chances to miss 6 times in a row with a to hit chance of 55%:
so: n=0.55 the chance to hit
the binomial coefficient for 6 shots and 0 hits - 6 over 0 - is quite simple, namely 1.
So therefore we have 1 * 0.55^^0 * (1-0.55)^6 = 0.0083

Which means the chance to miss 6 shots in a row with a 55% hit chance is.... 0.8 %
Or the opposite, the chance to get AT LEAST one hit with 6 shots is 99.2%
In other words, if the RNG does not completley suck, of the 1000 times all 6 soldiers shot with a hit chance of 55%, you should exactly see 8 times, that they all miss !

And like the OP said, it happens way too often. Therefore like I posted before, I think the seed update isn´t working as it should.

[Thradar]

Messaggio originale di Revolucas:

Oh bullcrap the RNG is fine:

http://steamcommunity.com/sharedfiles/filedetails/?id=103668868

I had an average of almost 80% hit percentage.

This indicates nothing. It indicates you finished the game. I'm about to finish it as well on Classic. The RNG seeding is utter crap.