Quote:
Originally Posted by Ten5x
You might be right that it's hard to convince a jury, but statistical analysis should be perfectly valid.
It is valid, as far as it goes. SA is just not evidence in and of itself. IOW, if I flip a coin 100 times and get 100 heads, SA would tell us the odds of that happening are 1 in 30 million. But it just happened. You might strongly suspect cheating, but the mere odds themselves are not definitive proof of cheating.
The fact that Postle's results are highly improbable is not definitive proof of cheating.
Quote:
DNA evidence never means "this person 100% did it." DNA evidence is often given like "the odds that this dna doesn't match this person is 1 in 200 billion" or something similar. DNA is using the very same types of techniques as statistical analysis. In a normal trial, DNA is almost never used as the sole piece of evidence. Even cases with almost no evidence, it's usually DNA + the fact that you lived near the crime scene at the time (so you're pulling from a small population instead of global population) means it's basically 100% accurate.
DNA is physical evidence, recovered from the crime scene. While it's true that statistical analysis is used to match a sample to a person, the math itself is meaningless without the physicality of the DNA itself.
Quote:
So in a criminal prosecution, it should be "statistically absurd winrate"
"Statistically absurd" is not really a thing, though. There's no precise definition of what would be absurd and absurd things happen all the time.
Quote:
+ "never bluffing into nuts out of X attempts when the best in the world are only like 60% accurate or whatever"
Good luck getting a jury to understand why that's significant. And if I'm the defense attorney, I'm going to be pressing the "hot streaks are common in poker, his just lasted longer," angle. I can get a math professor to testify that over millions of poker hands, you are going to see streaks that seem anomalous.
Quote:
+ "plays that make 0 sense from a winning players perspective"
If this were considered direct evidence of cheating, it would be chaos in poker rooms. Players make plays that make 0 sense all the time.
Quote:
+ "unique insider knowledge of how these broadcasting systems work"
This is something concrete. I think it's the best actual evidence against him.
Quote:
+ "suspicious use of phone during hands"
Meh.
Everyone uses their phones and there is no direct evidence that anything untoward was going on with the phone.
Quote:
+ "one of the biggest win rates of all time not choosing to move up in stakes and instead only play lowstakes televised games"
It could be argued that he was doing so well at this game that he didn't feel he needed to play at higher stakes or at any other game.
Quote:
and I don't see why it's not "beyond a reasonable doubt." I think a competent lawyer can explain how statistical analysis is actually very similar to dna in terms or probabilities.
"Beyond a reasonable doubt" doesn't mean 0% doubt, it means would a reasonable person think there's an alternate set of facts that seem plausible to explain what happened? And if you polled 100 poker experts and presented the facts of the case, I'd say 100% would think he's cheating beyond a reasonable doubt.
But the jury is 12 random people. The alternative set of facts is rather simple: Postle found a game that fit his wild style of play and hit an incredible -extremely unlikely but not impossible- streak of luck. Now, maybe a jury of poker experts wouldn't buy that story, but absent definitive proof, I can see how a jury of 12 regular people might.