Simon Krahnke wrote:
* Noah Roberts (19:18) schrieb:


I could be wrong, but I don't believe iterative deepening actually works
unless there is a transposition table to record results of previous
iterations. Iterative deepening is meant to increase the likelihood of
perfect move ordering for the search by filling the TTable with good
moves. Unless you are recording results somehow, and the ttable is the
most direct method, you are simply wasting time.


Actually the asymptotic complexity of iterative deepeing is the same as
of fixed depth, it's b^d, b being the branching factor and d being the
final/fixed depth.

That's because the non-final depth calculations are far less expensive.

Certainly. But unless you are recording the results they are not
informative.

I did it the same way, when i started my own lame engine: I started with
fixed depth and then went to iterative deepening. There was no slowdown
at all.

My engines both actually sped up, but I recorded the results of my
former searches in the ttable.

--
"I'm a war president. I make decisions here in the Oval Office
in foreign policy matters with war on my mind." - Bush

Dr. David Kirkby wrote:


Assuming you don't want to put a GUI on it (why waste time when they
exist), would it not be better to write it in standard C++ (if such a
thing exists),

It does.

rather than a Microsoft version of it ?? Personally I
think standard C is better , as its much more portable than C++, but I
guess I'm showing my age in prefering C.

C++ has become pretty portable. I write things in Linux and can also
compile them on windows with no additional stuff. I have written chess
engines that way. You can do more with the C++ stdlib than you can with
the C stdlib.

[snip several truths]

--
"I'm a war president. I make decisions here in the Oval Office
in foreign policy matters with war on my mind." - Bush

"Noah Roberts" wrote

I could be wrong, but I don't believe iterative deepening actually works
unless there is a transposition table to record results of previous
iterations. Iterative deepening is meant to increase the likelihood of
perfect move ordering for the search by filling the TTable with good
moves. Unless you are recording results somehow, and the ttable is the
most direct method, you are simply wasting time.

I do not agree! (you can still apply some move ordering (at the root) and
then iterative deepning allows a depth-first approach like minimax to be
stopped after a certain amount of time).

Tommy

* Noah Roberts (18:23) schrieb:

Actually the asymptotic complexity of iterative deepeing is the same as
of fixed depth, it's b^d, b being the branching factor and d being the
final/fixed depth.
That's because the non-final depth calculations are far less expensive.

Certainly. But unless you are recording the results they are not
informative.

As I mentioned it's informative for timing and guessing which depth to go
for.

I did it the same way, when i started my own lame engine: I started
with fixed depth and then went to iterative deepening. There was no
slowdown at all.

My engines both actually sped up, but I recorded the results of my
former searches in the ttable.

Of course, recodirng the best move improves iterative deepening a lot
(if the search is stable), but at the time i implemented iterative
deepeing there wasn't any transposition table yet.

Trying to use mtd(f) without ttable is suicide, but iterative deepening
isn't.

mfg, simon .... l