# DNA Scientists Are All FOS

Discussion in 'BlackHat Lounge' started by BreaknBrix, Feb 8, 2015.

1. ### BreaknBrixPower Member

Joined:
Mar 25, 2014
Messages:
756
4,322
Location:
NE US
So check this. Math confuses the shit out of me. As some of you may have witnessed in my previous thread about matching b-days.

But I got a problem I don't think anyone can solve.

And that is >>>DNA forensics.

I was just watching a show where "scientists" said there was a 1 in 20 sextillion, or "1/20000000000000000000000 chance" that they matched the DNA of a killer to a female cop. Then they just threw the bitch in prison for life.

However... how does this NOT violate the laws of physics? We don't do this shit with lotteries but we do it in forensics when someones life is on the line?

There are 7 billion people on the planet. But only 2.4 million locked up in the US and 49 million in the entire world.

We only collect DNA on prisoners. So somehow, these scientists can predict the DNA of math my calculator can't even do.

Doing it manually, that's 20 sextillion - 49 million = 19,999,999,999,999,951,000,000 people we don't even have DNA records of. Not to mention that 19,999,999,999,993,000,000,000 of those people DON'T EVEN FUCKING EXIST.

That shit is ABSURD. Using their same exact math, I can say that there is a 19.9 sextillion chance they are completely FULL OF SHIT.

2. ### tony_dElite Member

Joined:
Jun 22, 2013
Messages:
2,582
3,171
Location:
1600 Amphitheatre Parkway, Mountain View CA
Probably what they're doing is simply taking all of the variables in a DNA sequence and multiplying them by each other.

Eg, if there were 15 positions (or, indicators) in a DNA sequence, and there are 6 possible variables for each position, the equation would simply be;
6 x 6 x 6 x 6 x 6 x 6 x 6 x 6 x 6 x 6 x 6 x 6 x 6 x 6 x 6.

That gives you a 1 in 470 billion chance of the same number (DNA) appearing twice. So add a few more positions, or a few more variables in existing positions, and then you're pretty quickly up to that number you posted above.
Which I guess is the same as number plates; if you have NNN-LLL, you have (10x10x10) x (26x26x26) possible number plate combinations, therefore a 1 in 17.576m chance of any given combination.

Of course, to do the DNA calculation properly, you'd probably need to calculate the likely incidence of any given variable, and other fancy things that I neither know about, nor how to calculate. But I can see shit getting pretty 'unlikely' by the time you consider all the variables, and multiples thereof.

The fact that there aren't that many people in existence is kinda irrelevant to the fact that that's the chance.

• Thanks x 1
Last edited: Feb 8, 2015
3. ### prab1996Elite Member

Joined:
Jan 8, 2013
Messages:
3,494
2,031
Occupation:
Location:
Prab1996.com
Don't worry they can't do it. Because at first they will not have dna record of female cop in their criminal DNA databse and even if they matched it somehow still female cop will have a proof that she was on duty/shoping/home. (there are cameras or other people who might have seen her somewhere)

Moreover, there are many other things which you can extract with dna. Like cells , list of diseases , etc and there is thin probability of matching them all.

Last edited: Feb 8, 2015
4. ### TrepanatedSupreme Member

Joined:
Sep 18, 2010
Messages:
1,395
5,376
It's the the opposite of what you are thinking BB.

What they are saying is there is a 1 in 20 sextillion chance that the match is random.

So if it was 1 in 10 sextillion - there is a 99.99999999999999999999 certainty that it's a DNA match.

5. ### pewepPower Member

Joined:
Nov 3, 2008
Messages:
654
2,615
Gender:
Male
Occupation:
Yogi-In-Training
Location:
U.S.
Math is the foundation of the world.

6. ### SimpleSiteSolutionsJr. VIPJr. VIP

Joined:
Oct 2, 2009
Messages:
607
1,016
Occupation:
Location:
Yes, most likely they meant it's a 1 in 20 sextillion chance that it's NOT a match

7. ### neverquittingRegular Member

Joined:
Nov 10, 2014
Messages:
410
327
Location:
United States
"Scientist" here. There are four bases in DNA. They are A, T, C, and G. So say I pick a random point in your DNA and say "I think that's going to be an A." Well, I have a 1 in 4 chance of being right. Let's say I pick two consecutive bases and say I think it'll be an A, followed by a T. Well since I have to get both right, my odds drop to (1/4)(1/4)=(1/4^2)=(1/16). In short, if you're trying to make a call about X number of consecutive bases, you have a 1 in 4^X chance of being right just by dumb luck. Turns out that 4^37 is roughly 20 sextillion. So they sequenced the DNA out to 37 bases and it was a match (one with an infinitesimally small chance of just being bad luck). Then they put her in jail.

Next time consider the possibility that the truth is a little over your head rather than calling people full of shit. Thanks.

Joined:
Jun 22, 2013
Messages:
2,582