Judge wants *everyone* in the UK on DNA database

[detective-boy]

...or equipment going to the police that lets them know who is in a house and what they are thinking

Are you suggesting that this equipment exists and is being issued? If so, please link to a source because it sounds like bollocks to me.

 

[TAE]

No. You haven't. You may have done something else, but you have not posted a correction.

Yes, I did, on several occasions in several threads.

 

[detective-boy]

Note to d-b: I know it's confusing dear, but we are on the right thread; we agreed to stop derailing the other long ago.

No. We're not. There is a thread specifically for the DNA statistics debate.

http://www.urban75.net/vbulletin/showthread.php?t=221386

My question is that you further explain your logic (to get to 1:17) as I do not follow it, and apply it to tossing a coin repeatedly.

 

[detective-boy]

Yes, I did, on several occasions in several threads.

Whatever.

 

[ymu]

No. We're not. There is a thread specifically for the DNA statistics debate.

http://www.urban75.net/vbulletin/showthread.php?t=221386

My question is that you further explain your logic (to get to 1:17) as I do not follow it, and apply it to tossing a coin repeatedly.

Jesus. I didn't even see that one! Too busy to be seeking out new vortexes right now.

I resurrected this one ages ago - wondered why it was so dead.

 

[YoursTruly]

That might be ... but is the capability of warning a gang kicking fuck out of someone that the police have been called, in the hope they may desist and not kill him? (That is all that many councils intend to use the facility for - a couple of fuckwit councils have bigger ideas and these have been turned into sensationalist headlines by the tabloids)

Erh no...one of the councils was quite happy to demonstrate to a BBC reporter that it was happy to bark orders out to tell people to pick up litter.

I'm not having it, because eventually the whole thing can be abused to the point when the rules are changed down the years to something more and more authoritarian.

"You stop....show me your ID Card..." * Camera zooms

"Move along....sunshine"

Then in 10 years time, you'll be sat here trying to persuade us all that we need RFID enabled ID cards so that we don't have to show them to security cameras.

 

[YoursTruly]

Are you suggesting that this equipment exists and is being issued? If so, please link to a source because it sounds like bollocks to me.

The equipment is being worked on. I never said it existed.

Would you want it issued?

You see I want to understand your logic here. I want to see where you draw the line and why you draw the line where you draw the line. Then we'll have some fun and give you some arguments back using your own logic.

 

[MikeMcc]

Perfect analogy!

It's ~250 4million-sided die right now, and 17 60million-sided die if DNA gets collected as part of the ID card scheme.

Now, before rolling first you must select a die.

That's why it's one in 17 unique DNA samples profiled and compared to the hypothetical pan-national database. Because there is a one in a billion chance of a fluke match on all profiled sites. And this is true every time you pick up a new sample of DNA to test in this fashion.

So I think from your arguement in this post that you are seriously saying that because there is a 1 in 17 chance of there being a FP match (within the limits of the test) that the test does not prove guilt beyond reasonable doubt. Yet that defense would have to try and show that it could be reasonably expected that the other matching person in the UK would have to be in the same area, have motive, etc that it could then be reasonably expected that the accused was innocent.

It also disputes your claim that 1 in every prosecution suffered from this, a claim that is clearly rubbish. If there's 4 million on the present DNA database, I'll make a WAG (wild arsed guess) that 2 million of those have been involved in a prosecution, that means that there would be 117600 FP cases. Whereas your also say that there's a 1 in 17 chance in there being just ONE. At the other extreme if only 100k out of those 4 million resulted in prosecution that still indicates by your arguement that there would be 5880 cases, still a hell of a lot more than a 1 in 17 chance of 1!

 

[detective-boy]

Would you want it issued?

No. If you bothered to read my posts you would know what my views are on a whole range of issues. But clearly you are (a) assuming you know what my views will be simply because I was a police officer and (b) hoping that that is what they are because you want a fight.

Bye.

 

[ymu]

So I think from your arguement in this post that you are seriously saying that because there is a 1 in 17 chance of there being a FP match (within the limits of the test) that the test does not prove guilt beyond reasonable doubt. Yet that defense would have to try and show that it could be reasonably expected that the other matching person in the UK would have to be in the same area, have motive, etc that it could then be reasonably expected that the accused was innocent.

It also disputes your claim that 1 in every prosecution suffered from this, a claim that is clearly rubbish. If there's 4 million on the present DNA database, I'll make a WAG (wild arsed guess) that 2 million of those have been involved in a prosecution, that means that there would be 117600 FP cases. Whereas your also say that there's a 1 in 17 chance in there being just ONE. At the other extreme if only 100k out of those 4 million resulted in prosecution that still indicates by your arguement that there would be 5880 cases, still a hell of a lot more than a 1 in 17 chance of 1![/quote]
You seem to be making it up now. We've addressed the other evidentiary issues, the fact that it's rarer currently with a smaller db (the whole point is to try to estimate how bad it could be with a pan-national database). And it still seems pretty clear that you have no idea what a false positive is or what the Prosecutor's Fallacy is.

There's just no point going round in circles with you until you can demonstrate that you understand the basic point; risk of a false positive is not the same as the chance of being a false positive having had a positive test result. The latter is positive predictive value and it depends on prevalence of true positives in the population you test (1/n where n is the total DNA records in the database).

This is covered on the other thread. I'm done repeating myself - it's not like you're reliant on the sources I've offered (or TAE, or LLB), it's an easy google. You don't even seem to understand the calcs I'm doing (which are very standard for calculating test accuracy, ie Bayes Theorem). You only need basic probability theory to work it out from first principles, but you do need to apply it correctly.

 

[TAE]

Experts call for DNA restrictions

http://news.bbc.co.uk/1/hi/uk/6999703.stm

 

[MikeMcc]

It also disputes your claim that 1 in every prosecution suffered from this, a claim that is clearly rubbish. If there's 4 million on the present DNA database, I'll make a WAG (wild arsed guess) that 2 million of those have been involved in a prosecution, that means that there would be 117600 FP cases. Whereas your also say that there's a 1 in 17 chance in there being just ONE. At the other extreme if only 100k out of those 4 million resulted in prosecution that still indicates by your arguement that there would be 5880 cases, still a hell of a lot more than a 1 in 17 chance of 1!

You seem to be making it up now. We've addressed the other evidentiary issues, the fact that it's rarer currently with a smaller db (the whole point is to try to estimate how bad it could be with a pan-national database). And it still seems pretty clear that you have no idea what a false positive is or what the Prosecutor's Fallacy is.

There's just no point going round in circles with you until you can demonstrate that you understand the basic point; risk of a false positive is not the same as the chance of being a false positive having had a positive test result. The latter is positive predictive value and it depends on prevalence of true positives in the population you test (1/n where n is the total DNA records in the database).

This is covered on the other thread. I'm done repeating myself - it's not like you're reliant on the sources I've offered (or TAE, or LLB), it's an easy google. You don't even seem to understand the calcs I'm doing (which are very standard for calculating test accuracy, ie Bayes Theorem). You only need basic probability theory to work it out from first principles, but you do need to apply it correctly.

But you seem to be missing the point that the results of a search against the database are not based on a completely new independent tests, but are actually fixed. Your calculations depend on 60 million retests for every investigation, only then dues your statements make sense. But that is not how it's done!

 

[YoursTruly]

No. If you bothered to read my posts you would know what my views are on a whole range of issues. But clearly you are (a) assuming you know what my views will be simply because I was a police officer and (b) hoping that that is what they are because you want a fight.

Bye.

What makes you think I want "a fight?"

I haven't been abusive, I've been on topic, I haven't been disrespectful to you. I even laid out my cards on the table about what my intentions are in this debate.

Just on case you don't understand, I'll repeat myself, hopefully making myself more clear this time around.

What I want to do, is to see where you draw the line with between the needs/authority of the state and the needs/rights of the individual.

I'm just trying to make you think in a different way concerning your support of a national DNA database and national. That's all. I just want to get the logic behind your reasoning.

And yes, you do make an interesting subject, because you are an ex copper, but I don't have any problems with the police, as a matter of fact it's the reverse, I have a lot of respect for the police.

So. Please tell me, why you draw the line at a machine that can be pointed at someone and can give away what they are thinking?

Oh and if I assumed your views, I wouldn't be asking you a question.
Cheers.

 

[ymu]

But you seem to be missing the point that the results of a search against the database are not based on a completely new independent tests, but are actually fixed. Your calculations depend on 60 million retests for every investigation, only then dues your statements make sense. But that is not how it's done!

Not true. These theoretically inevitable false-positives will be false positives for the same people no matter how often you rerun the test on them. The one billion tests are independent because they include 17 samples * 60 million innocents on a database. No comparison is run twice. They are all essentially random.

A Muslim is more likely to match with DNA from a Muslim, and slightly moreso because the Muslim population is fairly small and may have a slightly more concentrated gene pool than in Pakistan. The news is much worse for typical anglo-saxons because most crimes are committed by whites and there are more chances for finding a false-positive when most of the random innocents are white also.

 

[detective-boy]

So. Please tell me, why you draw the line at a machine that can be pointed at someone and can give away what they are thinking?

Which bit of the "No" in my previous post are you having difficulty with.

And what makes you think I support an national DNA database. I don't. And my comments about you actually reading my posts rather than assuming my views is largely based on the fact that I have repeatedly posted that I do not believe that is right for unconvicted persons DNA to be retained on the database as at present.

Your repeated assumptions about my views, and your apparent inability to understand (or maybe believe) the posts which I make, is extremely annoying. THAT is what makes me think you are simply spoiling for a fight.

 

[ymu]

True, dat.

To be fair, I think his actual point is that the current system is essentially random and unfair (due to the keeping of DNA from people arrested for a recordable offence but never charged, or charged and later acquitted). I have raised this a number of times and offered up two alternative ways of addressing the issue: (a) go back to only keeping DNA from convicted persons, destroying samples taken from people later acquitted or (b) take samples from everyone.

He seems to have favoured the latter. I favour the former and, listening to Tony McNulty and the ACPO lead (CC Tony Lake from Lincolnshire) it seems they do to (or at least something in that direction). Though I suspect the cost / logistics of a database with everyone on it is what is putting them off the idea most!

 

[YoursTruly]

Which bit of the "No" in my previous post are you having difficulty with.

And what makes you think I support an national DNA database. I don't. And my comments about you actually reading my posts rather than assuming my views is largely based on the fact that I have repeatedly posted that I do not believe that is right for unconvicted persons DNA to be retained on the database as at present.

Your repeated assumptions about my views, and your apparent inability to understand (or maybe believe) the posts which I make, is extremely annoying. THAT is what makes me think you are simply spoiling for a fight.

My sincere apologies, I obviously misread you. I don't know why I thought you supported a national DNA database. I try not to presume, but I could have sworn you posted something up in it's defence. Obviously I was wrong! Sorry! I am human you know!

Cheers.

 

[ymu]

 

[MikeMcc]

Not true. These theoretically inevitable false-positives will be false positives for the same people no matter how often you rerun the test on them. The one billion tests are independent because they include 17 samples * 60 million innocents on a database. No comparison is run twice. They are all essentially random.

A Muslim is more likely to match with DNA from a Muslim, and slightly moreso because the Muslim population is fairly small and may have a slightly more concentrated gene pool than in Pakistan. The news is much worse for typical anglo-saxons because most crimes are committed by whites and there are more chances for finding a false-positive when most of the random innocents are white also.

The comparisons are always run repeatedly, but the physical test of the DNA sample is only done once when a person is entered onto the database.

 

[ymu]

The comparisons are always run repeatedly, but the physical test of the DNA sample is only done once when a person is entered onto the database.

That's irrelevant. It's stored as a number. And there's a one in a billion chance that someone else out there has the same number from their DNA profile. That is the statistical limit on it's specificity; 999,999,999/1,000,000,000 times the sample will be unique, the other one time it won't be because someone else has the same alleles at their profiled STRs as the sample happens to have.

 

[Barking_Mad]

blah blah blah, numbers numbers numbers, blah blah blah.

We're human beings, not numbers. I dont give a rat's arse whether there is no chance at all of me being put in jail for a crime I didn't commit - the fact is my DNA is mine, and not the property of the state. I should not have to give them by force or threat something that does not belong to them.

I am not the property of the state
I am not the property of the state
I am not the property of the state
I am not the property of the state
I am not the property of the state
I am not the property of the state.

ta.

 

[Fruitloop]

Yeah I don't see how the debate has become so centered on efficacy. I mean it would greatly reduce crime if we all had an RFID chip inserted in our skulls at birth, but it still ain't a good idea.

 

[detective-boy]

I could have sworn you posted something up in it's defence.

Thank you for the apology. You may have remembered something out of context. It is rare that something has no advantages or no disadvantages. If I can see them, I will usually mention them regardless of my overall feeling about the concept as a whole. Sadly that means you have to take my posts in full context, something which regularly catches out some posters!

In relation to the DNA database I have mentioned that at least a compulsory one would have the advantage of fairness (as everyone would be treated the same) unlike the current random situation where some innocent people are on the database just because they got arrested on suspicion of something they did not do, something whihc is essentially random.

But overall, no. I don't think a compulsory national database is the way to go.

 

[MikeMcc]

That's irrelevant. It's stored as a number. And there's a one in a billion chance that someone else out there has the same number from their DNA profile. That is the statistical limit on it's specificity; 999,999,999/1,000,000,000 times the sample will be unique, the other one time it won't be because someone else has the same alleles at their profiled STRs as the sample happens to have.

Bloody hell how many times do we have to go back and forth on this point. It's nothing to do with the number of tests (ignoring the contribution from human error during the test), it's do do with the number of individuals. The DNA pattern of the individuals in the database does not change at all (births, deaths, migration, yadda, yadda...). The subsequently tested suspects are not checked against a completely new set of individuals.

Your arguement is suffering from a logical fault that would indicate that there would be many thousands of false positives. This is clearly false.

 

[ymu]

Bloody hell how many times do we have to go back and forth on this point. It's nothing to do with the number of tests (ignoring the contribution from human error during the test), it's do do with the number of individuals. The DNA pattern of the individuals in the database does not change at all (births, deaths, migration, yadda, yadda...). The subsequently tested suspects are not checked against a completely new set of individuals.

Your arguement is suffering from a logical fault that would indicate that there would be many thousands of false positives. This is clearly false.

Think about it Mike.

It's not the number of individuals, it's the number of individual comparisons. A match is found if the large random number belonging to one person is the same as that of another. If 60 million people are on the database and 6 billion people could have committed the crime, there are a hell of a lot more than 60 million possible comparisons!

Remember, this is not a one off "does this person have DNA which matches a single unique criminal pattern, or not". It's does this person have DNA which matches the stuff we found at crime scene A, or B, or C ...

Put it this way. If I tested Person A in the database against a billion suspect samples, there'd be one false match on average. Agree? So if I test a billion suspect samples against all 60 million in the database, there'd be 60 million false positives. It multiplies and adds. 1 sample is 60 million unique tests; 2 samples is 120 unique million tests; 17 samples is > billion unique tests.

Honestly, make an effort to get your head around it - you're wrong but you can't learn anything until you're willing to accept that possibility.