<s>
In	O
computer	B-Application
chess	I-Application
programs	O
,	O
the	O
null-move	B-Algorithm
heuristic	I-Algorithm
is	O
a	O
heuristic	B-Algorithm
technique	O
used	O
to	O
enhance	O
the	O
speed	O
of	O
the	O
alpha-beta	B-Algorithm
pruning	I-Algorithm
algorithm	B-Application
.	O
</s>
<s>
Alpha-beta	B-Algorithm
pruning	I-Algorithm
speeds	O
the	O
minimax	B-Algorithm
algorithm	I-Algorithm
by	O
identifying	O
cutoffs	O
,	O
points	O
in	O
the	O
game	O
tree	O
where	O
the	O
current	O
position	O
is	O
so	O
good	O
for	O
the	O
side	O
to	O
move	O
that	O
best	O
play	O
by	O
the	O
other	O
side	O
would	O
have	O
avoided	O
it	O
.	O
</s>
<s>
The	O
null-move	B-Algorithm
heuristic	I-Algorithm
is	O
designed	O
to	O
guess	O
cutoffs	O
with	O
less	O
effort	O
than	O
would	O
otherwise	O
be	O
required	O
,	O
whilst	O
retaining	O
a	O
reasonable	O
level	O
of	O
accuracy	O
.	O
</s>
<s>
The	O
null-move	B-Algorithm
heuristic	I-Algorithm
is	O
based	O
on	O
the	O
fact	O
that	O
most	O
reasonable	O
chess	B-Application
moves	O
improve	O
the	O
position	O
for	O
the	O
side	O
that	O
played	O
them	O
.	O
</s>
<s>
So	O
,	O
if	O
the	O
player	O
whose	O
turn	O
it	O
is	O
to	O
move	O
can	O
forfeit	O
the	O
right	O
to	O
move	O
(	O
or	O
make	O
a	O
null	O
move	O
-	O
an	O
illegal	O
action	O
in	O
chess	B-Application
)	O
and	O
still	O
have	O
a	O
position	O
strong	O
enough	O
to	O
produce	O
a	O
cutoff	O
,	O
then	O
the	O
current	O
position	O
would	O
almost	O
certainly	O
produce	O
a	O
cutoff	O
if	O
the	O
current	O
player	O
actually	O
moved	O
.	O
</s>
<s>
In	O
employing	O
the	O
null-move	B-Algorithm
heuristic	I-Algorithm
,	O
the	O
computer	O
program	O
first	O
forfeits	O
the	O
turn	O
of	O
the	O
side	O
whose	O
turn	O
it	O
is	O
to	O
move	O
,	O
and	O
then	O
performs	O
an	O
alpha-beta	B-Algorithm
search	I-Algorithm
on	O
the	O
resulting	O
position	O
to	O
a	O
shallower	O
depth	O
than	O
it	O
would	O
have	O
searched	O
the	O
current	O
position	O
had	O
it	O
not	O
used	O
the	O
null	O
move	O
heuristic	B-Algorithm
.	O
</s>
<s>
Because	O
a	O
shallow	O
search	O
is	O
faster	O
than	O
deeper	O
search	O
,	O
the	O
cutoff	O
is	O
found	O
faster	O
,	O
accelerating	O
the	O
computer	B-Application
chess	I-Application
program	O
.	O
</s>
<s>
There	O
are	O
a	O
class	O
of	O
chess	B-Application
positions	O
where	O
employing	O
the	O
null-move	B-Algorithm
heuristic	I-Algorithm
can	O
result	O
in	O
severe	O
tactical	O
blunders	O
.	O
</s>
<s>
In	O
these	O
positions	O
,	O
the	O
null-move	B-Algorithm
heuristic	I-Algorithm
may	O
produce	O
a	O
cutoff	O
where	O
a	O
full	O
search	O
would	O
not	O
have	O
found	O
one	O
,	O
causing	O
the	O
program	O
to	O
assume	O
the	O
position	O
is	O
very	O
good	O
for	O
a	O
side	O
when	O
it	O
may	O
in	O
fact	O
be	O
very	O
bad	O
for	O
them	O
.	O
</s>
<s>
To	O
avoid	O
using	O
the	O
null-move	B-Algorithm
heuristic	I-Algorithm
in	O
zugzwang	O
positions	O
,	O
most	O
chess-playing	B-Application
programs	I-Application
that	O
use	O
the	O
null-move	B-Algorithm
heuristic	I-Algorithm
put	O
restrictions	O
on	O
its	O
use	O
.	O
</s>
<s>
Another	O
heuristic	B-Algorithm
for	O
dealing	O
with	O
the	O
zugzwang	O
problem	O
is	O
Omid	O
David	O
and	O
Nathan	O
Netanyahu	O
's	O
verified	B-Algorithm
null-move	I-Algorithm
pruning	I-Algorithm
.	O
</s>
<s>
In	O
verified	B-Algorithm
null-move	I-Algorithm
pruning	I-Algorithm
,	O
whenever	O
the	O
shallow	O
null-move	O
search	O
indicates	O
a	O
fail-high	O
,	O
instead	O
of	O
cutting	O
off	O
the	O
search	O
from	O
the	O
current	O
node	O
,	O
the	O
search	O
is	O
continued	O
with	O
reduced	O
depth	O
.	O
</s>
