|
|
Thread Tools | Display Modes |
#21
|
|||
|
|||
Robert Chung wrote: Kurgan Gringioni wrote: Then those two get tested again. They both test positive *again*. The odds of that happening are 250^4 You're assuming independence. Dumbass - Let's say that it's not, that one false positive test will always yield a second positive on the B sample. Then the odds are one in 499 that it TH's roomate will also test positive. That 1 in 499 chance goes way, way up, however, when one considers that *no one else* had a "false positive". I had a probability class a long time ago and don't remember how to calculate it. Someone who has a better brain than I step in here. Regardless, it's a big number - something on the order of at least 10,000 to 1(very, very conservatively). thanks, K. Gringioni. |
Ads |
#22
|
|||
|
|||
Kurgan Gringioni wrote:
Robert Chung wrote: Kurgan Gringioni wrote: Then those two get tested again. They both test positive *again*. The odds of that happening are 250^4 You're assuming independence. Dumbass - Let's say that it's not, that one false positive test will always yield a second positive on the B sample. Then the odds are one in 499 that it TH's roomate will also test positive. That 1 in 499 chance goes way, way up, however, when one considers that *no one else* had a "false positive". You're still assuming independence. This is exactly why the entire chain of events has to be vetted. That was not done. And, anticipating your question, I remind you of this: http://groups-beta.google.com/group/...c35a0e1644d31c |
#23
|
|||
|
|||
Robert Chung wrote: Kurgan Gringioni wrote: Robert Chung wrote: Kurgan Gringioni wrote: Then those two get tested again. They both test positive *again*. The odds of that happening are 250^4 You're assuming independence. Dumbass - Let's say that it's not, that one false positive test will always yield a second positive on the B sample. Then the odds are one in 499 that it TH's roomate will also test positive. That 1 in 499 chance goes way, way up, however, when one considers that *no one else* had a "false positive". You're still assuming independence. This is exactly why the entire chain of events has to be vetted. That was not done. And, anticipating your question, I remind you of this: http://groups-beta.google.com/group/...c35a0e1644d31c Dumbass - How would Hamilton's test be dependent upon Perez's test? thanks, K. Gringioni. |
#24
|
|||
|
|||
Kurgan Gringioni wrote:
How would Hamilton's test be dependent upon Perez's test? I didn't say that Hamilton's test is dependent on Perez's. I think it likely that they're both dependent on the same systematic things. |
#25
|
|||
|
|||
"Robert Chung" wrote in message
... Kurgan Gringioni wrote: Robert Chung wrote: Kurgan Gringioni wrote: Then those two get tested again. They both test positive *again*. The odds of that happening are 250^4 You're assuming independence. Dumbass - Let's say that it's not, that one false positive test will always yield a second positive on the B sample. Then the odds are one in 499 that it TH's roomate will also test positive. That 1 in 499 chance goes way, way up, however, when one considers that *no one else* had a "false positive". You're still assuming independence. This is exactly why the entire chain of events has to be vetted. That was not done. And, anticipating your question, I remind you of this: http://groups-beta.google.com/group/...c35a0e1644d31c Robert, you're wasting your time arguing with Henry. |
#26
|
|||
|
|||
Robert Chung wrote: Kurgan Gringioni wrote: How would Hamilton's test be dependent upon Perez's test? I didn't say that Hamilton's test is dependent on Perez's. I think it likely that they're both dependent on the same systematic things. Dumbass - Such as? |
#27
|
|||
|
|||
Kurgan Gringioni wrote:
Such as? Well, the obvious one is the one you're thinking of: that there was an organized blood doping program. That's the one I lean towards, too. But, remember, I get paid high in the three figures to worry about non-obvious sources of systematic bias. That's part of what goes into proper vetting. |
#28
|
|||
|
|||
Robert Chung wrote: Kurgan Gringioni wrote: Such as? Well, the obvious one is the one you're thinking of: that there was an organized blood doping program. That's the one I lean towards, too. But, remember, I get paid high in the three figures to worry about non-obvious sources of systematic bias. That's part of what goes into proper vetting. Dumbass - Seriously, what's a potential systemic bias in this case? I'm curious what that may be. thanks, K. Gringioni. |
#29
|
|||
|
|||
Kurgan Gringioni wrote:
Seriously, what's a potential systemic bias in this case? I'm curious what that may be. I emphasize that I don't know about this particular case but, in other particular cases I've examined, systematic bias can arise because tests were done in the same place by the same people with the same equipment. This type of bias sometimes happens when there are a very limited number of places where something is performed, or a limited number of examiners who were all trained in the same place. Sometimes people see what they expect to see (we could name some rbr'ers who fall into that category); sometimes this happens innocently. Here's a kinda innocent way that things can get biased: suppose, for example, 1) blood doping is pretty evenly spread across the peloton; 2) experts at the technique are better at recognizing blood doping than novices; and 3) when a case happens to be sensitive it gets shunted up to an expert (this appears to be what happened at the Olympics, and Dick Pound suggested that Hamilton's Vuelta tests were scrutinized exactly because of the Oly results). Now suppose samples from teammates of the suspected blood doper are considered to be sensitive, and thus also get shunted to experts (it needn't be the same expert, just to another expert who has a higher detection rate than a novice). You would tend to see more doping than the overall rate based on the experts' analyses, but you'd also see less doping than the overall rate based on the novices' analyses. I think it probable that Hamilton blood doped. I think it probable that Perez blood doped. But I have no way of evaluating exactly how probable since the test hasn't been properly vetted, and I wouldn't build on the fact that only two people out those tested so far have been announced as positive. And, as for the A and B samples on each rider, they are done to eliminate random error in testing, not systematic error. Since Magilla claims to have seen the evidence, I'm hoping he'll tell us the proportion of foreign cells that were detected in Hamilton's blood in the Olympics A sample, the Vuelta A sample, and the Vuelta B sample. Two samples taken from the same person at the same time, like the Vuelta A and B samples, would give a very rough (n=2) idea of the precision of the test. In conjunction with a sample taken at a different point in time, like the Oly A sample, they might be able to give a rough idea of how much doping occurred. I think both sides in the dispute already know all of this. |
#30
|
|||
|
|||
Kurgan Gringioni wrote:
Ewoud Dronkert wrote: On Tue, 01 Mar 2005 15:48:16 GMT, B Lafferty wrote: Don't forget Santi Perez's two positives. If they're all false positives, that makes five from two guys on one team. What are the odds? We don't know, because there is nothing published on false positive (or false negative for that matter) rates for this test. Let's say that all the tests are false positives. What are the odds that the *only two* riders to get the false positive are teamates and roomates? Let's say 500 riders were tested. Two false positives make the rate 1 in 250. Then those two get tested again. They both test positive *again*. The odds of that happening are 250^4 = 3,906,250,000 about one in 4 billion. Then factor in the fact that they're roomates. Multiply the 4 billion by 499. 1 in 2 trillion. Poor Tugboat. K. Gringioni. Ahh. nope..the false positive rate is nothing of that magnitude if it's linked to specific types of anitgens that only Perez and Hamilton have in their blood. They would always test positive. Magilla |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Tyler in long BBC interview | David George | Racing | 0 | November 29th 04 11:29 AM |
Marital discord - Hamilton | crit pro | Racing | 9 | September 27th 04 12:42 PM |
Tyler Not Cleared, Lab Blunder | never_doped | Racing | 2 | September 25th 04 06:33 AM |
Olympic Pick Contest: finaler | Dan Connelly | Racing | 2 | August 19th 04 04:44 AM |
Tyler Hamilton Foundation Kick-Off | Richard Adams | Racing | 0 | November 10th 03 08:25 PM |