<s>
The	O
computer	B-Algorithm
game	I-Algorithm
bot	I-Algorithm
Turing	I-Algorithm
test	I-Algorithm
is	O
a	O
variant	O
of	O
the	O
Turing	O
test	O
,	O
where	O
a	O
human	O
judge	O
viewing	O
and	O
interacting	O
with	O
a	O
virtual	B-Application
world	I-Application
must	O
distinguish	O
between	O
other	O
humans	O
and	O
video	O
game	O
bots	B-Protocol
,	O
both	O
interacting	O
with	O
the	O
same	O
virtual	B-Application
world	I-Application
.	O
</s>
<s>
The	O
computer	B-Algorithm
game	I-Algorithm
bot	I-Algorithm
Turing	I-Algorithm
test	I-Algorithm
was	O
proposed	O
to	O
advance	O
the	O
fields	O
of	O
artificial	B-Application
intelligence	I-Application
(	O
AI	B-Application
)	O
and	O
computational	B-General_Concept
intelligence	I-General_Concept
with	O
respect	O
to	O
video	O
games	O
.	O
</s>
<s>
It	O
was	O
considered	O
that	O
a	O
poorly	O
implemented	O
bot	B-Protocol
implied	O
a	O
subpar	O
game	O
,	O
so	O
a	O
bot	B-Protocol
that	O
would	O
be	O
capable	O
of	O
passing	O
this	O
test	O
,	O
and	O
therefore	O
might	O
be	O
indistinguishable	O
from	O
a	O
human	O
player	O
,	O
would	O
directly	O
improve	O
the	O
quality	O
of	O
a	O
game	O
.	O
</s>
<s>
It	O
also	O
served	O
to	O
debunk	O
a	O
flawed	O
notion	O
that	O
"	O
game	O
AI	B-Application
is	O
a	O
solved	O
problem.	O
"	O
</s>
<s>
Emphasis	O
is	O
placed	O
on	O
a	O
game	O
bot	B-Protocol
that	O
interacts	O
with	O
other	O
players	O
in	O
a	O
multiplayer	O
environment	O
.	O
</s>
<s>
Unlike	O
a	O
bot	B-Protocol
that	O
simply	O
needs	O
to	O
make	O
optimal	O
human-like	O
decisions	O
to	O
play	O
or	O
beat	O
a	O
game	O
,	O
this	O
bot	B-Protocol
must	O
make	O
the	O
same	O
decisions	O
while	O
also	O
convincing	O
another	O
in-game	B-Application
player	O
of	O
its	O
human-likeness	O
.	O
</s>
<s>
The	O
computer	B-Algorithm
game	I-Algorithm
bot	I-Algorithm
Turing	I-Algorithm
test	I-Algorithm
was	O
designed	O
to	O
test	O
a	O
bot	B-Protocol
's	O
ability	O
to	O
interact	O
with	O
a	O
game	O
environment	O
in	O
comparison	O
with	O
a	O
human	O
player	O
,	O
simply	O
'	O
winning	O
 '	O
was	O
insufficient	O
.	O
</s>
<s>
There	O
are	O
three	O
participants	O
:	O
a	O
human	O
player	O
,	O
a	O
computer-game	O
bot	B-Protocol
,	O
and	O
a	O
judge	O
.	O
</s>
<s>
The	O
bot	B-Protocol
needs	O
to	O
appear	O
more	O
human-like	O
than	O
the	O
human	O
player	O
.	O
</s>
<s>
Judge	O
scores	O
are	O
not	O
bipolar	O
—	O
both	O
human	O
and	O
bot	B-Protocol
can	O
be	O
scored	O
anywhere	O
on	O
a	O
scale	O
from	O
1	O
to	O
5	O
(	O
1	O
=	O
not	O
humanlike	O
,	O
5	O
=	O
human	O
)	O
.	O
</s>
<s>
Bots	B-Protocol
were	O
not	O
given	O
omniscient	O
powers	O
as	O
they	O
may	O
be	O
in	O
other	O
games	O
.	O
</s>
<s>
Bots	B-Protocol
must	O
react	O
only	O
to	O
the	O
data	O
that	O
might	O
be	O
reasonably	O
available	O
to	O
a	O
human	O
player	O
.	O
</s>
<s>
The	O
contest	O
was	O
held	O
with	O
the	O
game	O
Unreal	B-Application
Tournament	I-Application
2004	I-Application
as	O
the	O
platform	O
.	O
</s>
<s>
Contestants	O
created	O
their	O
bots	B-Protocol
in	O
advance	O
using	O
the	O
GameBots	O
interface	O
.	O
</s>
<s>
GameBots	O
had	O
some	O
modifications	O
made	O
so	O
as	O
to	O
adhere	O
to	O
the	O
above	O
conditions	O
,	O
such	O
as	O
removing	O
data	O
about	O
vantage	O
points	O
or	O
weapon	O
damage	O
that	O
unfairly	O
informed	O
the	O
bots	B-Protocol
of	O
relevant	O
strengths/weakness	O
that	O
a	O
human	O
would	O
otherwise	O
need	O
to	O
learn	O
.	O
</s>
<s>
The	O
first	O
BotPrize	O
Tournament	O
was	O
held	O
on	O
17	O
December	O
2008	O
,	O
as	O
part	O
of	O
the	O
2008	O
IEEE	O
Symposium	O
on	O
Computational	B-General_Concept
Intelligence	I-General_Concept
and	O
Games	O
in	O
Australia	O
.	O
</s>
<s>
Each	O
competing	O
team	O
was	O
given	O
time	O
to	O
set	O
up	O
and	O
adjust	O
their	O
bots	B-Protocol
to	O
the	O
modified	O
game	O
client	O
,	O
although	O
no	O
coding	O
changes	O
were	O
allowed	O
at	O
that	O
point	O
.	O
</s>
<s>
Judges	O
were	O
the	O
last	O
to	O
join	O
the	O
server	O
and	O
every	O
judge	O
observed	O
every	O
player	O
and	O
every	O
bot	B-Protocol
exactly	O
once	O
,	O
although	O
the	O
pairing	O
of	O
players	O
and	O
bots	B-Protocol
did	O
change	O
.	O
</s>
<s>
When	O
the	O
tournament	O
ended	O
,	O
no	O
bot	B-Protocol
was	O
rated	O
as	O
more	O
human	O
than	O
any	O
player	O
.	O
</s>
<s>
In	O
subsequent	O
tournaments	O
,	O
run	O
during	O
2009	O
–	O
2011	O
,	O
bots	B-Protocol
achieved	O
scores	O
that	O
were	O
increasingly	O
human-like	O
,	O
but	O
no	O
contestant	O
had	O
won	O
the	O
BotPrize	O
in	O
any	O
of	O
these	O
contests	O
.	O
</s>
<s>
In	O
2012	O
,	O
the	O
2K	O
BotPrize	O
was	O
held	O
once	O
again	O
,	O
and	O
two	O
teams	O
programmed	O
bots	B-Protocol
that	O
achieved	O
scores	O
greater	O
than	O
those	O
of	O
human	O
players	O
.	O
</s>
<s>
To	O
date	O
,	O
there	O
have	O
been	O
two	O
successfully	O
programmed	O
bots	B-Protocol
that	O
passed	O
the	O
computer	B-Algorithm
game	I-Algorithm
bot	I-Algorithm
Turing	I-Algorithm
test	I-Algorithm
:	O
</s>
<s>
UT^2	O
,	O
a	O
team	O
from	O
the	O
University	O
of	O
Texas	O
at	O
Austin	O
,	O
emphasized	O
a	O
bot	B-Protocol
that	O
adjusted	O
its	O
behaviour	O
based	O
on	O
previously	O
observed	O
human	O
behaviour	O
and	O
neuroevolution	B-Algorithm
.	O
</s>
<s>
The	O
team	O
has	O
made	O
their	O
bot	B-Protocol
available	O
,	O
although	O
a	O
copy	O
of	O
Unreal	B-Application
Tournament	I-Application
2004	I-Application
is	O
required	O
.	O
</s>
<s>
Mihai	O
Polceanu	O
,	O
a	O
doctoral	O
student	O
from	O
Romania	O
,	O
focused	O
on	O
creating	O
a	O
bot	B-Protocol
that	O
would	O
mimic	O
opponent	O
reactions	O
,	O
in	O
a	O
sense	O
'	O
borrowing	O
 '	O
the	O
human-like	O
nature	O
of	O
the	O
opponent	O
.	O
</s>
<s>
The	O
outcome	O
of	O
a	O
bot	B-Protocol
that	O
appears	O
more	O
human-like	O
than	O
a	O
human	O
player	O
is	O
possibly	O
overstated	O
,	O
since	O
in	O
the	O
tournament	O
in	O
which	O
the	O
bots	B-Protocol
succeeded	O
,	O
the	O
average	O
'	O
humanness	O
 '	O
rating	O
of	O
the	O
human	O
players	O
was	O
only	O
41.4	O
%	O
.	O
</s>
<s>
In	O
light	O
of	O
this	O
,	O
the	O
BotPrize	O
competition	O
organizers	O
will	O
increase	O
the	O
difficulty	O
in	O
upcoming	O
years	O
with	O
new	O
challenges	O
,	O
forcing	O
competitors	O
to	O
improve	O
their	O
bots	B-Protocol
.	O
</s>
<s>
It	O
is	O
also	O
believed	O
that	O
methods	O
and	O
techniques	O
developed	O
for	O
the	O
computer	B-Algorithm
game	I-Algorithm
bot	I-Algorithm
Turing	I-Algorithm
test	I-Algorithm
will	O
be	O
useful	O
in	O
fields	O
other	O
than	O
video	O
games	O
,	O
such	O
as	O
virtual	O
training	O
environments	O
and	O
in	O
improving	O
Human	O
–	O
robot	O
interaction	O
.	O
</s>
<s>
The	O
computer	B-Algorithm
game	I-Algorithm
bot	I-Algorithm
Turing	I-Algorithm
test	I-Algorithm
differs	O
from	O
the	O
traditional	O
or	O
generic	O
Turing	O
test	O
in	O
a	O
number	O
of	O
ways	O
:	O
</s>
<s>
Unlike	O
the	O
traditional	O
Turing	O
test	O
,	O
for	O
example	O
the	O
Chatterbot-style	O
contest	O
held	O
annually	O
by	O
the	O
Loebner	O
Prize	O
competition	O
,	O
the	O
humans	O
who	O
played	O
against	O
the	O
Computer	O
Game	O
Bots	B-Protocol
are	O
not	O
trying	O
to	O
convince	O
judges	O
they	O
are	O
the	O
human	O
;	O
rather	O
,	O
they	O
want	O
to	O
win	O
the	O
game	O
(	O
i.e.	O
,	O
by	O
achieving	O
the	O
highest	O
kill	O
score	O
)	O
.	O
</s>
<s>
With	O
regards	O
to	O
a	O
successful	O
video	O
game	O
bot	B-Protocol
,	O
this	O
is	O
not	O
to	O
be	O
confused	O
with	O
a	O
claim	O
that	O
the	O
bot	B-Protocol
is	O
'	O
intelligent	O
,	O
'	O
whereas	O
a	O
machine	O
that	O
'	O
passed	O
 '	O
the	O
Turing	O
test	O
would	O
arguably	O
have	O
some	O
evidence	O
for	O
its	O
Chatterbot	B-Protocol
's	O
'	O
intelligence.	O
'	O
</s>
<s>
The	O
game	O
Unreal	B-Application
Tournament	I-Application
2004	I-Application
was	O
chosen	O
for	O
its	O
commercial	O
availability	O
and	O
its	O
interface	O
for	O
creating	O
bots	B-Protocol
,	O
GameBots	O
.	O
</s>
<s>
The	O
available	O
information	O
to	O
the	O
participants	O
,	O
humans	O
and	O
bots	B-Protocol
,	O
is	O
not	O
equal	O
.	O
</s>
<s>
Humans	O
interact	O
through	O
vision	O
and	O
sound	O
,	O
whereas	O
bots	B-Protocol
interact	O
with	O
data	O
and	O
events	O
.	O
</s>
<s>
The	O
judges	O
cannot	O
introduce	O
new	O
events	O
(	O
e.g.	O
,	O
a	O
lava	O
pit	O
)	O
to	O
aid	O
in	O
differentiating	O
between	O
human	O
and	O
bot	B-Protocol
,	O
whereas	O
in	O
a	O
Chatterbot	B-Protocol
designed	O
system	O
,	O
judges	O
may	O
theoretically	O
ask	O
any	O
question	O
in	O
any	O
manner	O
.	O
</s>
