/LMAR=0/XLINE=3/FONT#0=BASL30/FONT#1=BASI30/FONT#2=BASB30/FONT#10=BAXM30/FONT#11=ZERO30/FONT#3=STA200/FONT#4=NGB25
␈↓ ↓H␈↓␈↓βS␈↓∧ Artificial Intelligence Laboratory, STANFORD UNIVERSITY, Stanford, California 94305

␈↓ ↓H␈↓∧Telephone 415 497-4430␈↓ 	YSeptember 3, 1980 




␈↓ ↓H␈↓Professor M. M. Newborn
␈↓ ↓H␈↓School of Computer Science McGill University
␈↓ ↓H␈↓Montreal, Quebec H3A 2K6
␈↓ ↓H␈↓CANADA

␈↓ ↓H␈↓Dear Monte:

␈↓ ↓H␈↓        Many thanks for the invitation to give away the awards.

␈↓ ↓H␈↓        Let␈α∂me␈α∂take␈α⊂this␈α∂occasion␈α∂to␈α⊂make␈α∂some␈α∂suggestions␈α⊂concerning␈α∂the␈α∂tournament.␈α⊂ I␈α∂don't
␈↓ ↓H␈↓think they can be applied to the this year's, but I would like to urge them for the future.

␈↓ ↓H␈↓        It␈α∞has␈α∞always␈α∞bothered␈α∞me␈α
that␈α∞the␈α∞scores␈α∞of␈α∞computer␈α
chess␈α∞games␈α∞are␈α∞annotated␈α∞just␈α
as
␈↓ ↓H␈↓though␈α∩they␈α⊃were␈α∩games␈α∩between␈α⊃humans,␈α∩and␈α∩there␈α⊃were␈α∩no␈α∩way␈α⊃of␈α∩determining␈α∩what␈α⊃the
␈↓ ↓H␈↓program␈αwas␈αthinking␈αabout.␈α This␈αis␈αparticularly␈αbothersome␈αwhen␈αthe␈αsponsor␈αof␈αthe␈α
contest␈αis
␈↓ ↓H␈↓the ACM, which should be alert to opportunities to advance computer science.

␈↓ ↓H␈↓        Therefore,␈αI␈αwould␈αlike␈αto␈αpropose␈αthat␈αit␈αbe␈αa␈αcondition␈αof␈αentry␈αto␈αthe␈α
ACM␈αtournament
␈↓ ↓H␈↓in␈α
1980␈αthat␈α
programs␈α
have␈αsome␈α
minimal␈α
facilities␈αfor␈α
printing␈α
what␈αthey␈α
were␈α
thinking␈αabout.␈α
 I
␈↓ ↓H␈↓don't␈αhave␈αrecent␈αexperience␈αwith␈αchess␈αprograms,␈αbut␈αI␈αwould␈αlike␈αto␈αsuggest␈αsome␈αfacilities␈αthat
␈↓ ↓H␈↓we␈αused␈αin␈αan␈αancient␈αprogram␈αfor␈αthe␈αgame␈αof␈αKalah.␈α The␈αprogram␈αran␈αon␈αa␈αPDP-1␈αcomputer
␈↓ ↓H␈↓that␈α⊂had␈α∂only␈α⊂4096␈α∂18␈α⊂bit␈α∂words,␈α⊂so␈α∂costly␈α⊂analysis␈α∂features␈α⊂could␈α∂not␈α⊂be␈α∂used.␈α⊂ I␈α⊂think␈α∂these
␈↓ ↓H␈↓proposed facilities are well within the capability of any of the recent programmers of chess.

␈↓ ↓H␈↓        After each move the program printed the following:

␈↓ ↓H␈↓1. The ≡rst two plies of the move tree examined.

␈↓ ↓H␈↓2.␈αFrom␈αeach␈αend␈αpoint␈αof␈αthe␈α≡rst␈αtwo␈αplies,␈αthe␈αsubsequent␈αprincipal␈αvariation␈αto␈αthe␈αmaximum
␈↓ ↓H␈↓depth of search.  (This was sometimes distorted by alpha-beta cuto≥s).

␈↓ ↓H␈↓3.␈αAt␈α
the␈αend␈α
of␈αeach␈α
such␈αvariation,␈α
the␈αvalue␈α
of␈αthe␈α
endpoint,␈αalso␈α
sometime␈αdistorted␈αby␈α
cuto≥s,
␈↓ ↓H␈↓and␈α∂the␈α∂amount␈α⊂of␈α∂e≥ort␈α∂that␈α⊂went␈α∂into␈α∂the␈α⊂subtree.␈α∂ The␈α∂e≥ort␈α⊂was␈α∂the␈α∂number␈α⊂of␈α∂positions
␈↓ ↓H␈↓examined,␈α
but␈αonly␈α
relative␈α
values␈αare␈α
wanted,␈α
so␈αany␈α
such␈α
measure,␈αe.g.␈α
computer␈α
time␈αwould␈α
also
␈↓ ↓H␈↓be acceptable.

␈↓ ↓H␈↓4.␈α∂It␈α∂would␈α∂also␈α⊂be␈α∂worthwhile␈α∂to␈α∂print␈α⊂the␈α∂values␈α∂of␈α∂␈↓↓alpha␈↓␈α∂and␈α⊂␈↓↓beta␈↓␈α∂with␈α∂which␈α∂each␈α⊂of␈α∂the
␈↓ ↓H␈↓variations was entered.

␈↓ ↓H␈↓5.␈α∞It␈α∞would␈α∞be␈α∞further␈α∂worthwhile␈α∞if␈α∞the␈α∞programs␈α∞stored␈α∂the␈α∞values␈α∞of␈α∞the␈α∞parameters␈α∂used␈α∞to
␈↓ ↓H␈↓αProfessor M. M. Newborn␈↓ ¬RSeptember 3, 1980␈↓ 
nPage 2␈↓ 

␈↓ ↓H␈↓make␈α
each␈α
move,␈α
so␈α
that␈α
the␈α
move␈α
could␈α
be␈α
re-evaluated␈α
for␈α
the␈α
bene≡t␈α
of␈α
a␈α
commentator,␈α
who
␈↓ ↓H␈↓could, for example, ask whether a particular variation was considered.

␈↓ ↓H␈↓        The␈α
ACM␈α
might␈α
not␈α
require␈α
that␈α
this␈α
information␈α
be␈α
printed␈α
with␈α
the␈α
move␈αalthough␈α
there
␈↓ ↓H␈↓would probably be time to print it while the opposing program was thinking about its next move.

␈↓ ↓H␈↓        In␈αmy␈αopinion,␈αimposing␈αsome␈αsuch␈αrequirements␈αon␈αentries␈αto␈αthe␈αACM␈αtournament␈αwould
␈↓ ↓H␈↓raise␈αthe␈α
scienti≡c␈αlevel␈αof␈α
the␈αwhole␈αenterprise,␈α
would␈αeducate␈αthe␈α
public␈αto␈αthe␈α
fact␈αthat␈αwe␈α
really
␈↓ ↓H␈↓do␈αknow␈αsomething␈αabout␈α
how␈αthese␈αthings␈αwork␈αand␈α
to␈αthe␈αdi≥erences␈αbetween␈α
human␈αthought
␈↓ ↓H␈↓and␈α
current␈α
programs,␈α
and␈α
would␈α
lead␈α
to␈α
better␈α
programs␈α
by␈α
leading␈α
to␈α
more␈α
informed␈α
criticism␈α
of
␈↓ ↓H␈↓existing programs.


␈↓ ↓H␈↓Sincerely,



␈↓ ↓H␈↓John McCarthy
␈↓ ↓H␈↓Director
␈↓ ↓H␈↓Professor of Computer Science

␈↓ ↓H␈↓cc: Professor Ben Mittman, Computation Center, Northwestern University, David Levy.