🚀 go-pugleaf

> > But what about parallel processing?  If you are examining 1e9 positions
a
> > second, your numbers dwindle real quickly.
>
>
> How long will it take you to search 10^50 positions at 10^9 positions
> per second?
>
> hint:  10^41 seconds.  _big_ number.

It has been my experience that trying to explain that it's simply not
possible for a chess program to search every possible sequence of moves is a
frustrating, and most of the time, futile task to get the person to
understand.

The poster who said something about how 10^48 _appears_ "close" to 10^50
said it best.  10^48 is not anywhere close to 10^50.  Here is a rough
estimate: 10^50 - 10^48 = 9.9^49.  Once you have searched 10^48 positions
(which isn't going to happen), you still have another 9.9^49 to go
(roughly).  With each new level of the depth of the search you are going to
multiply the time it took to do the previous search by the branching factor.
With a branching factor of 35-38 (average number of chess moves in a given
position), that means that once you've searched 10^15 positions, it's going
to take you 35 to 38 times longer to search all of the positions from 10^15
up to 10^16.  Even if you could reach 10^15 positions in your lifetime,
10^16 isn't likely to happen, and 10^17 is sure not to occur, now think
about 10^50!!  Now do you see why it's simply not possible?

Here's a real life example:

Let's say it takes you 1 year to search 10^15 positions in 1 second (which
is nowhere close to a good estimate, I'm just using 1 second to make a
point).  If it took you 1 second to search 10^15 positions, it will take 35
seconds to search 10^16 positions.  You can continue to multiply the time by
35 to get an estimate of how long it will take to search the next exponent
power.  Here are the numbers:

20 minutes to search 10^17 positions
12 hours to search 10^18 positions
17 days to search 10^19 positions
1.6 years to search 10^20 positions
58 years to search 10^21 positions
2,040 years to search 10^22 positions
71,406 years to search 10^23 positions
2,499,227 years to search 10^24 positions
87,472,962 years to serach 10^25 positions
and so on...

AND that is assuming that you can calculate 10^15 positions in 1 second,
which you can not do, or even come close to doing.