--- In

primenumbers@yahoogroups.com, "djbroadhurst" <d.broadhurst@...> wrote:

>

> --- In primenumbers@yahoogroups.com,

> "djbroadhurst" <d.broadhurst@> wrote:

>

> > Slightly unlucky. The project has found about 2660 primes,

> > each with 3 associated targets sieved to depth 41T.

>

> Sorry: 41T was the range of k that was sieved; the depth was 200T.

> It matters little. For

>

> > exp(-3*2660*exp(Euler)*log(41*10^12)/(666666*log(2))) =~ 38%.

>

> please read:

>

> exp(-3*2660*exp(Euler)*log(2*10^14)/(666666*log(2))) =~ 36%

>

> for the probability of not finding a pair, noting the rather

> small change of 2%, for spending 5 times longer on sieving.

>

http://www.primegrid.com/stats_sgs_llr.php shows 2554 primes. Adjusting David's figure:

exp(-3*2554*exp(Euler)*log(2*10^14)/(666666*log(2))) =~ 38%

I hope PG will continue with their search window for twin/SGs when the primes are too small the top5000, and that DB finds some more AP3s thereof. This is assuming they don't soon strike gold,

Paul