Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
0.975 | 125 |
0.026 | 133 |
0.975 | 141 |
0.55 | 149 |
0.075 | 157 |
0.6 | 165 |
0.3 | 174 |
0.1 | 184 |
0.075 | 197 |
0.825 | 285 |
4.501 | Expected AAs - Take the under |
I don’t think that math makes sense?Using the A&A likelihood table in the bracket release thread, we should expect this:
0.975 125 0.026 133 0.975 141 0.55 149 0.075 157 0.6 165 0.3 174 0.1 184 0.075 197 0.825 285 4.501Expected AAs - Take the under
He just took the decimals for each percentage and then added them together. Although it is admittedly a very flawed system, it coincidentally came out to 4.5. Which is EXACTLY what the over/under should be!!!I don’t think that math makes sense?
The way you calculated it, if Iowa had Spencer Lee and 9 28 seeds, it would give a very high probability of 0 All-Americans. The reality would be a .975% chance of 1 All-American.
I say that, and still think, yep, that’s probably about right
Just curious MSU158, do you see a flaw in the method other than it's using a general person with that seed vs. the specific person with that seed?He just took the decimals for each percentage and then added them together. Although it is admittedly a very flawed system, it coincidentally came out to 4.5. Which is EXACTLY what the over/under should be!!!
So, I wholeheartedly endorse his method in this case.
Yeah, percentages are simply a "likelihood" of 1 guy AA'ing. You really can't add percentages together to get to a number of AA's. Each guy is his own specific calculation. In reality 5 guys at a 20% chance do not add up to 1 guy AAing. It simply isn't a case of rolling a 5 sided dice 5 times, expecting to get your result 1 in 5 tries...Just curious MSU158, do you see a flaw in the method other than it's using a general person with that seed vs. the specific person with that seed?
I think you can argue that this guy or that guy is better than the "average X seed", but totaling up the individual chances to get an overall expected value is sound math. Five guys at a 20% does add up to 1 guy AA'ing, on average.Yeah, percentages are simply a "likelihood" of 1 guy AA'ing. You really can't add percentages together to get to a number of AA's. Each guy is his own specific calculation. In reality 5 guys at a 20% chance do not add up to 1 guy AAing. It simply isn't a case of rolling a 5 sided dice 5 times, expecting to get your result 1 in 5 tries...
But, it really doesn't. Each percentage is based off it's own 33 person bracket in a "closed" setting". Combining percentages really doesn't work here.I think you can argue that this guy or that guy is better than the "average X seed", but totaling up the individual chances to get an overall expected value is sound math. Five guys at a 20% does add up to 1 guy AA'ing, on average.
Your math is the same math that I would use to say Michael Jordan and I have 6 NBA Championships.Just curious MSU158, do you see a flaw in the method other than it's using a general person with that seed vs. the specific person with that seed?
Don't they teach statistics at MSU? I kid.But, it really doesn't. Each percentage is based off it's own 33 person bracket in a "closed" setting". Combining percentages really doesn't work here.
Would you really expect 1 AA from 20 wrestlers with a 5% chance spread across 20 different brackets? Now, if you had all 20 in the SAME bracket, yes that would hold up. But, across 20 brackets, the odds are much more likely zero AA than 1....
I get the math. I actually DID take a statistics class. I just don't think that method works well in this situation.Don't they teach statistics at MSU? I kid.
I won't go back and forth on this more, but if someone else wants to carry the expected value torch for me, be my guest.
Me thinks your math is right (just not intuitive), assuming the results at each weight class are independent (e.g., if Spencer Lee wins, does that really pump up Woods or Murin and change their 'expected value'?).Don't they teach statistics at MSU? I kid.
I won't go back and forth on this more, but if someone else wants to carry the expected value torch for me, be my guest.
Since we’re killing time for a couple more hours, I’ll try to explain it in a way that might make sense.Don't they teach statistics at MSU? I kid.
I won't go back and forth on this more, but if someone else wants to carry the expected value torch for me, be my guest.
Take the over. I think we end with 6, outside shot at 7Iowa number of All-Americans:
over 5.5 -150
under 5.5 +120
Which side you coming in on this one?
Not that I really give a shit, but huh? His math would say the expected number of AAs in your example would be 1.09.Since we’re killing time for a couple more hours, I’ll try to explain it in a way that might make sense.
If you have 10 wrestlers; one with 100% chance to AA, and 9 with a 1% chance, the equation you used would suggest that it is highly likely you would have zero AAs. Even though there is a 100% chance of getting at least one.
You added and divided perfectly. It just doesn’t work for this particular exercise.
I think you can argue that this guy or that guy is better than the "average X seed", but totaling up the individual chances to get an overall expected value is sound math. Five guys at a 20% does add up to 1 guy AA'ing, on average.
Thanks...was wondering if you or Spooner was going to help me out here.It’s been decades since my Quantitative Analysis courses, but I believe this is correct. On the other hand, the chances of having no AA would be .2 ^5, or .00032; meaning the chances of at least one would be .99968.
To be clear, I really wasn't arguing the math itself, just how accurate it would be when trying to calculate an over/under.Thanks...was wondering if you or Spooner was going to help me out here.
No AA would be .8^5, or 0.32768 (i.e., 32.768%), so at least 1 AA would be 67.232%.
You did argue against the math, in multiple ways and multiple postings. I could cite several examples, but will cite just one:To be clear, I really wasn't arguing the math itself, just how accurate it would be when trying to calculate an over/under.
Expected Value is a great formula for investment portfolios, not so much when trying to take 10 brackets across 10 different weight classes and accurately try to get results based off seeds!
Ugh, did I not clarify what I meant after? Multiple times? Did I not concede the actual math? multiple times as well? My argument has simply been its a flawed system for BETTING.You did argue against the math, in multiple ways and multiple postings. I could cite several examples, but will cite just one:
you wrote: "Would you really expect 1 AA from 20 wrestlers with a 5% chance spread across 20 different brackets?"
If you don't know the answer to that, we can't really discuss statistics.