My save doesn't like men

[Nathair]

Why should you expect any one particular result?

Because that is the point of the exercise. The claim was made that "The only way to have genders even is to take the random function out." The expectation "to have even genders" is explicit.

Assuming each coin toss is non-dependent upon any of the others, 28-2 and 15-15 is equally likely, along with all the other possible combinations of 30 flips. It's only when you start looking at dependent results (ie, runs of heads in a row) do the odds start to vary.

And that is exactly what we are discussing here. This entire thread is about the dependent results and our expectations of them. We expect rescue kerbals of both genders more or less equally.

If we toss the coin thirty times, and repeat this set of flips 50 times over, and every set of 30 is around 15-15, then I would contend that is anything BUT random.

Then you would be flat wrong. Looking only at aggregate results, finding that a coin toss produces 50/50 odds is emphatically not evidence of bias. Despite your denial of expectation you are here expecting to see some of the more unlikely aggregations in the various runs and then intuiting bias if you don't find them. That's not how it works, the absence of evidence of bias is not evidence of bias.

So to answer your original post. Yes, "random" in computers means something different that in statistics. Most computers cannot create a random number. It just can't. And if something they did with the AD mod affected the seed generator (or more likely, fudged up a line of code somewhere), then the RNG may have a bias installed.

There's an enormous difference between saying that a given RNG may have a bias and saying that any and all random functions must have a large inherent bias.

And, in point of fact, the only difference in computers is that what we tend to call "random number generators" are actually "pseudo-random number generators". The pseudo-random NGs are deterministic in process but statistically random in output (and output is what we are discussing here, the genders of rescue kerbals.) When these pRNGs are randomly seeded at run time the results can, for the clearly defined purposes of this particular discussion, be accurately called random.

[March Unto Torment]

There is an astonishing lack of sexism in this thread.

Good job, KSP community

Anyhow, it sounds to me that while isolated cases of Kerbelle Space Program could occur, the sort of endemic instancing we're seeing here makes it seem statistically improbable. Especially the aforementioned 70:0 number.

[Pax Kerbana]

There is an astonishing lack of sexism in this thread.

Good job, KSP community

You can thank the mods for that!

Regarding the probability of ... ratios, all I can say is that for anyone to defend a 100% single ... result for any installation, it's a little ridiculous. But now we have many instances of that. With or without mods isn't really the issue. This shouldn't be happening, I cannot understand how a mod could cause this or that mods should be used as an excuse. When I first saw all the rescuees were female I thought, "meh, eventually some male rescue contracts will appears." Such is not the case.

While we're on the subject tho, while Squad is fixing this, can you please reduce the cost of new Kerbal hires to AT LEAST 10% and remove the cumulative hiring costs?! The current cost and structure is ridiculous. I propose having Tier 0 applicants be cheap and come with no skills. As the Astronaut Complex is upgraded, applicants start with 1 star for Tier 1 and 2 stars for Tier 2. Their cost to hire would increase respectively. That would make a lot more sense. Alternatively, or maybe in addition too, the hire cost could be based on stats such as stupid ones cost much less and badS ones might be 3x the average cost, because you know, they're cowboys (or gals)!

[tater]

Yeah, I was in the "you just flipped "heads" 50 times in a row, it's still random" camp when people first reported this. Then I had a career with ZERO male rescues, VIPs, or tourists. A career that filled the entire tech tree. I made a point after a while of rejecting nearly every single such contract (accepted several, just to make sure it was't related to the same contact being replaced with a new one with the same gender). After that did;t generate any males, I started a new career.Same thing. Then another. Same thing. It strains reason that 4 careers in the same setup with no career/contract mods aside from Asteroid day (if it has any) would all generate many hundreds of exclusively female contracts.

[Renegrade]

Anyhow, it sounds to me that while isolated cases of Kerbelle Space Program could occur, the sort of endemic instancing we're seeing here makes it seem statistically improbable. Especially the aforementioned 70:0 number.

I love me some rescue missions. Why upgrade the Astronaut Complex past the second tier and pay for kerbals when missions appear that pay YOU to receive a kerbal, which also bypasses the max kerbal limit

I've since played a stock game a bit (as in pure stock, no mods at all) and I'm getting males and females on rescue missions. Next, I'm going to install Asteroid Day and see what happens.

(I stand by my theory that it's a cyclical bias of some kind)

[Vanamonde]

I have never noticed if I get more females than males, or whatever. A kerbal's a kerbal.

[Nathair]

I have never noticed if I get more females than males, or whatever. A kerbal's a kerbal.

I only rescue the women, I call the program The Revenge of the Mercury 13.

[The_Rocketeer]

Unlike on Earth, where men are from Mars and women are from Venus, in KSP boy-kerbs are from Eve and girl-kerbs are from Duna. The disparity arises simply because reaching orbit from Eve is so damn hard.

[Gargamel]

And that is exactly what we are discussing here. This entire thread is about the dependent results and our expectations of them. We expect rescue kerbals of both genders more or less equally.

So.... since Kerbal A is female..... or the last coin flip was heads.... the next result is affected by the previous result?

[tater]

I have never noticed if I get more females than males, or whatever. A kerbal's a kerbal.

I never did, either, until I added the Asteroid Day mod, and my new career had zero males, ever, for rescues/ferry/tourist contracts.

- - - Updated - - -

So.... since Kerbal A is female..... or the last coin flip was heads.... the next result is affected by the previous result?

Each event is unrelated to the others. Normal distributions are not the issue. Some people have had multiple carers with zero males offered. Not low numbers, zero, and not just one career. I did, even after making a point of tiring my wrist by declining hundreds of rescue/ferry VIP/tourist contracts, without a single male.

[Wild Cobra]

My Minmus station is 5 girls and 1 guy.

He must be exhausted!

[Dunbaratu]

The problem with the arguments about how proper randomness will generate some cases that look skewed, is that it should generate cases that look skewed the OTHER way too, and we're not seeing that. Nobody's posing here saying "actually, in my game I'm getting all males...".

I noticed the skew as well, but shrugged and assumed I was one of the outliers. The problem with judging whether something is random or not is that you need lots more sample cases than one person will witness playing the game. Just because YOU notice all females doesn't mean it's not random - YOU might just be an outlier case.

But in this thread we're seeing Multiple players noticing it and it definitely seems related to the rescue missions. If you populate your space program from hiring you tend to get equal numbers, but if you populate your space program from rescues, you tend to get a lot more females. Something odd is happening.

Also, the fact that some people have reported kerbal s e x change randomly after being hired is really weird - that's probably a separate, second glitch, although it's not one I've seen.

Edited September 15, 2015 by Steven Mading
apparently there's some banned word filter in place making it impossible to type "s e x change" without the spaces.

[Parkaboy]

I can attest that in a save game I started with the Asteroid Day mod installed only female kerbals appeared on the rescue contracts. After I deleted the mod, I started getting male kerbals.

[Renegrade]

I've followed up with my testing - installing Asteroid Day on the stock install (which had proper-seeming distribution prior to Asteroid Day) definitely causes it to change to all-female.

Anyhow, my best-guess programmer-gut-feeling is that Asteroid Day is causing some bias to occur somehow (either that or clobbering a value).

[Tex_NL]

I've followed up with my testing - installing Asteroid Day on the stock install (which had proper-seeming distribution prior to Asteroid Day) definitely causes it to change to all-female.

Anyhow, my best-guess programmer-gut-feeling is that Asteroid Day is causing some bias to occur somehow (either that or clobbering a value).

No offence but didn't I already say this a week ago in post #3?

[Gargamel]

Quote Originally Posted by gargamel

So.... since Kerbal A is female..... or the last coin flip was heads.... the next result is affected by the previous result?

Each event is unrelated to the others. Normal distributions are not the issue. Some people have had multiple carers with zero males offered. Not low numbers, zero, and not just one career. I did, even after making a point of tiring my wrist by declining hundreds of rescue/ferry VIP/tourist contracts, without a single male.

My comment was in response to a comment that said the coin flips were dependent upon previous results.

I'm not doubting in the bias in the game, far from it, there is way too much evidence to deny it. There was just a separate discussion happening within the thread.

[Renegrade]

No offence but didn't I already say this a week ago in post #3?

You said, "as far as I know", which doesn't really tell us anything about testing or presence or absence of mods etc, so, no.

I've tested a pure stock install, and and now a pure stock + Asteroid Day install (I checked out about 70 rescue contracts, to match my prior 'heavily modded' sample set in both new cases).