20210613, 12:58  #1 
"University student"
May 2021
Beijing, China
3·67 Posts 
Is the "Combined" factoring probability actually wrong on mersenne.ca?
The "Combined" probability just add the trial factoring probability and the P1 factoring probability, but that makes no sense.
Take 108071849 for example: It has been trial factored to 2^77, with 64.9351% chance of finding a factor. Also, the chance of finding a factor in P1 with b1=755000, b2=21551000 assuming no factor below 2^77 is 3.5775% (a conditional probability). Let A denote the event of "finding a factor below 2^77" B denote the event of "finding a factor in P1 with b1=755000, b2=21551000" then P(A) = 64.9351%, P(B(!A)) = P(!A && B) / P(!A) = 3.5775% So the total probability of finding a factor should be P(A  B) = P(A) +P(!A && B) = P(A) + P(B(!A)) * P(!A) = P(A) + P(B(!A)) * (1P(!A)) = 0.0649351 + 0.035775 * (10.649351) = 66.1895% Which makes sense. This formula can also avoid having over 113% probability (???) on exponents such as 1277. (PS: I'm only a freshman student studying probability. If my formula is wrong, please point out :) Last fiddled with by Zhangrc on 20210613 at 12:59 
20210614, 00:52  #2  
"James Heinrich"
May 2004
exNorthern Ontario
13×277 Posts 
Quote:
If others agree that your implementation makes sense I'm happy to revise the site accordingly. But I'll need a little more explanation in simple terms what you mean with the logical combinations of probabilities, such as P(!A && B) or P(B(!A)) 

20210614, 03:25  #3  
"University student"
May 2021
Beijing, China
3×67 Posts 
Quote:
Last fiddled with by Zhangrc on 20210614 at 03:52 

20210614, 03:53  #4  
"University student"
May 2021
Beijing, China
3·67 Posts 
Quote:


20210614, 04:01  #5 
"James Heinrich"
May 2004
exNorthern Ontario
7021_{8} Posts 
Unfortunately that means about the same to me as this.

20210614, 04:03  #6 
"James Heinrich"
May 2004
exNorthern Ontario
7021_{8} Posts 

20210614, 04:07  #7 
"University student"
May 2021
Beijing, China
11001001_{2} Posts 
It seems so. But you'd better wait for another one who really knows the stuff and states that the formula is correct.

20210614, 06:03  #8  
"Mihai Preda"
Apr 2015
10101011111_{2} Posts 
Quote:
A) the chances of being eaten by the tiger is: "60% of 10%" == 0.6 * 0.1 == 6% (because, if the aligator gets him first, the tiger is out of luck). Overall being eaten is: 90% + 6%, 96%. B) considering the complement: surviving the whole trip means: not being eaten by the crocodile (10%), AND not being eaten by the tiger (40%). Surviving = 10% * 40% == 4%. The complement of surviving thus is 1  4% == 96%, same as above. Last fiddled with by preda on 20210614 at 06:10 Reason: spelling 

20210614, 12:52  #9 
"James Heinrich"
May 2004
exNorthern Ontario
E11_{16} Posts 

20210614, 14:44  #10 
If I May
"Chris Halsall"
Sep 2002
Barbados
2^{3}·19·67 Posts 

20210614, 15:42  #11  
Sep 2009
2×11×101 Posts 
Quote:
In the normal case where both TF and P1 have only a few% chance of finding a factor adding the probabilities would be nearly right. Which is probably why it's not been noticed until now. Chris 

Thread Tools  
Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
What's wrong with PrimeGrid? I keep getting "The connection was reset"...  Stargate38  Riesel Prime Search  3  20170814 18:14 
"factoring" vs "factorizing"  ixfd64  Factoring  4  20121016 04:07 
Official "Benford was wrong!" thread  MiniGeek  Aliquot Sequences  15  20090618 19:09 
P1 factoring != "Mersenne numbers to factor"?  James Heinrich  Marin's Mersennearies  8  20040517 11:09 
trial factoring of "small" mersenne numbers  antiroach  Lone Mersenne Hunters  6  20030716 23:35 