something fishy about jellyfish
helmet
i am around 1750 on fibs. downloaded jellyfish and just dont trust it.
is it that good at bg or does it cheat ?
always seems to get good roll after double by either side and
throws especially well in crucial games . i played it solid for about 3 weeks
and i too have come to conclution they named it jellyfish because there is
something fishy about it.is programmer having joke at our expence
Michael J Zehr
A. It's that good.
B. It never gets tired or steamed. (Are you *sure* you are still
playing your best game after the first time it gets a "lucky" roll?)
C. It's all too easy to underestimate it. Are you playing with the
same intensity and concentration you'd use if Kit Woolsey or Hal
Heinrich were sitting on the other side of the table. Do you change
your natural pace of play to match its pace?
D. Try inputting your own dice and see if it continues to get just as
"lucky."
-michael j zehr
Mashadi
Could you please tell me where I can download jellyfish.
Thank you
Mashadi
In article <acummin.5...@es.co.nz> helmet wrote:
>Date: Tue, 10 Jun 1997 14:45:59 +1200
>From: acu...@es.co.nz (helmet)
>Newsgroups: rec.games.backgammon
>Subject: something fishy about jellyfish
dave magyar
>
>D. Try inputting your own dice and see if it continues to get just as
>"lucky."
>
i would just like to add my 2.5 cents worth... when i first started
playing jellyfish i encountered an abnormal amount of successful rolls
for the computer.
a friend of mine had the same experience. he was getting beat left
and right.
he decided to roll the dice by himself and guess what? as the rolls
seems to be crazy still (haha) he seemed to start winning more. now i
do not know what this means, and i do not plan to speculate (my
rating is so bad that i just as soon divulge my credit card number and
exp date for the world to have a little fun with - utiil th limit gets
hit of course :) than tell it to the world) but i do find the results
odd! i like jellyfish and i think it is fun!
fruit for thought...
i am no different than anyone else - i get super pissed off when my
opponent rolls a ton of better rolls than i do. for the longest time
i challenged the best of them in here and via email. i emailed ken,
stephen and a few others with my complaints. you want to know what i
realized through all of this NOW PAY ATTENTION PLEASE... i seemed to
forget the good rolls i got moreover than the bad rolls. in short, my
memory for kicking @$$ with killer rolls seemed to slip my mind when
it came for my turn at getting bad rolls.
every now and Zen (no typo) i get aggravated when it doesn't go my
way, but then again, if it went my way i would expect more and more
and more and more until i could tell bill gates what to do with his
low salary .. but then i would still want more.
so, my advice is (if you're still listening) is to learn that the dice
are going to be good, bad, and of course ugly at times (thx clint e.)
and getting mad about it will only aggravate one person, take a guess
who that could possibly be?
try to enjot the game for what it is... there are a ton of great
people to meet in the backgammon community and being angry about the
rolls of dice will not make you a ton of friends
as my 3 year old says...
hasta lasagna
dave magyar
dm on gamesgrid
dmagyar on fibs
Morten Wang
[Mashadi <mas...@erols.com>]
> Could you please tell me where I can download jellyfish.
http://www.effect.no/jelly.htm
Morten!
--
"God does not deduct from our allotted
lifespan the time spent playing backgammon"
-> Morty on FIBS
--> Backgammon homepage: http://home.sn.no/~warnckew/gammon/
Colin Johnson
On Tue, 10 Jun 1997 20:11:39 GMT, da...@usga.org (dave magyar) wrote:
>>
>>D. Try inputting your own dice and see if it continues to get just as
>>"lucky."
>>
>
>i would just like to add my 2.5 cents worth... when i first started
>playing jellyfish i encountered an abnormal amount of successful rolls
>for the computer.
>
>a friend of mine had the same experience. he was getting beat left
>and right.
>
>he decided to roll the dice by himself and guess what? as the rolls
>seems to be crazy still (haha) he seemed to start winning more. now i
>do not know what this means, and i do not plan to speculate (my
>rating is so bad that i just as soon divulge my credit card number and
>exp date for the world to have a little fun with - utiil th limit gets
>hit of course :) than tell it to the world) but i do find the results
>odd! i like jellyfish and i think it is fun!
Hi,
I'm new to this newsgroup and new to Jelly fish but both myself and a
friend of mine who consider ourselves pretty solid bg players think
it is a cheating son-of-a-bitch.
I've yet to verify this by the manual dice rolling test you mention
but it CANNOT possibly get some of the rolls it comes up with.
In addition some of its tactics are just plain crazy??? like hitting
me in it's home base when it is winning and it cannot cover the
counter? Why do this - no human opponent would even consider it,
espescially when I had four out of six of my home blocks made.
It is apparent that it cheats if you play it on a fast machine. If you
double it for no reason it takes offence and starts cheating. The
cheating manifests itself by a noticeable delay before it's
dice-from-heaven appears.
What does it do? - roll all combinations until it finds one it likes?
All respect to the creator though, it has always been an idea of mine
to create a neural net enabled backgammon game for windows.
Who knows, maybe I'll do it now, make one that doesn't cheat,
can play the game like a human and is free!!!1
I would suggest however that the author re-evaluates the net topology
or the training algorithm or whatever is making it behave the way it
does. It is ANYTHING but natural.
Colin Johnson.
Stephen Hubbard
colin_...@globalvillage.com (Colin Johnson) wrote:
>On Tue, 10 Jun 1997 20:11:39 GMT, da...@usga.org (dave magyar) wrote:
>>>
>>>D. Try inputting your own dice and see if it continues to get just as
>>>"lucky."
>>>
>>
>>i would just like to add my 2.5 cents worth... when i first started
>>playing jellyfish i encountered an abnormal amount of successful rolls
>>for the computer.
>>
>>a friend of mine had the same experience. he was getting beat left
>>and right.
>>
>>he decided to roll the dice by himself and guess what? as the rolls
>>seems to be crazy still (haha) he seemed to start winning more. now i
>>do not know what this means, and i do not plan to speculate (my
>>rating is so bad that i just as soon divulge my credit card number and
>>exp date for the world to have a little fun with - utiil th limit gets
>>hit of course :) than tell it to the world) but i do find the results
>>odd! i like jellyfish and i think it is fun!
>Hi,
>I'm new to this newsgroup and new to Jelly fish but both myself and a
>friend of mine who consider ourselves pretty solid bg players think
>it is a cheating son-of-a-bitch.
You might want to reconsider!
>I've yet to verify this by the manual dice rolling test you mention
>but it CANNOT possibly get some of the rolls it comes up with.
Why not?
>In addition some of its tactics are just plain crazy??? like hitting
>me in it's home base when it is winning and it cannot cover the
>counter? Why do this - no human opponent would even consider it,
>espescially when I had four out of six of my home blocks made.
Sometimes this might be the best move. Never say never. Open yourself
to some other possibilities and maybe your game will improve.
>It is apparent that it cheats if you play it on a fast machine. If you
>double it for no reason it takes offence and starts cheating. The
>cheating manifests itself by a noticeable delay before it's
>dice-from-heaven appears.
Even more likely it is deciding whether to accept your double for no
reason.
>What does it do? - roll all combinations until it finds one it likes?
More likely sets it self up by "solid" play so that more of the
possible 36 rolls are "good"rolls.
>All respect to the creator though, it has always been an idea of mine
>to create a neural net enabled backgammon game for windows.
>Who knows, maybe I'll do it now, make one that doesn't cheat,
>can play the game like a human and is free!!!1
Please let us know when you do!
>I would suggest however that the author re-evaluates the net topology
>or the training algorithm or whatever is making it behave the way it
>does. It is ANYTHING but natural.
What would seem natural to you? A program you could beat with your
same old tired natural human moves?
>Colin Johnson.
A satisfied customer whose game has improved a lot by playing JF and
who thinks the price is a bargain and no I have no financial interest
involved!
--
Stephen Hubbard
the...@mindspring.com
jag
In article <33b9a7e5...@news.globalvillage.com>,
colin_...@globalvillage.com (Colin Johnson) wrote:
[snip]
>Hi,
>
>I'm new to this newsgroup and new to Jelly fish but both myself and a
>friend of mine who consider ourselves pretty solid bg players think
>it is a cheating son-of-a-bitch.
>
>I've yet to verify this by the manual dice rolling test you mention
>but it CANNOT possibly get some of the rolls it comes up with.
>
Its hitting you from behind a 6point prime with a 1 7? During the race its
swapping the cube for the dice and suddenly gaining 64 pips a roll? Please
do tell us all its impossible rolls.
I've been playing JammyFish for quite some time now, I consider myself a
'pretty solid bg player' and I've come to the following conclusion:
JammyFish comes up with some realy improbable rolls
These improbable rolls favour me as often as JammyFish
These improbable rolls are no more improbable than some of the rolls
experienced in my (usually) twice weekly chouette evenings.
You've already admitted you haven't tried playing it with manual dice, go
away and play with manual dice WITHOUT cheating and let us know if you
really are as solid a bg player as you think you are.
As for hitting loose when my opponent has only made 4 points in his home
board, well I've done it plenty of times in the past and I'll probably be
doing it again tonight.
jag on FIBS
James Grenier
Director, Information Systems, The HLT Group Ltd
http://www.hltpublications.co.uk
http://www.holborncollege.ac.uk
David Montgomery
In article <33b9a7e5...@news.globalvillage.com> colin_...@globalvillage.com (Colin Johnson) writes:
>On Tue, 10 Jun 1997 20:11:39 GMT, da...@usga.org (dave magyar) wrote:
>but it CANNOT possibly get some of the rolls it comes up with.
Well, let's see. It CANNOT possibly get these rolls... so I guess
it wasn't anything like 3-1 or 6-3, which come up 2/36 of the time.
Nor anything like 1-1 or 6-6, which come up 1/36 of the time. No,
it must have been something entirely different -- 7,7? -1,12?
32767,1296? Now I'm really curious what it rolled! (And how it
displayed the pips on those little dice!)
My Jellyfish always rolls one of the 21 legal rolls. But I probably
just haven't stumbled across the right seed to get the rolls that
CANNOT possibly be rolled. :-)
David Montgomery
mo...@cs.umd.edu
monty on FIBS
Graham Trevor Price
In <33b9a7e5...@news.globalvillage.com>
colin_...@globalvillage.com (Colin Johnson) writes:
>it is a cheating son-of-a-bitch.
>Colin Johnson.
Where can I get this cheat version? The version I have just plays
a strong and aggressive game but I can't get it to cheat.
Have I been cheated?
Graham.
Fredrik Dahl
Colin Johnson wrote:
>
> Hi,
>
> I'm new to this newsgroup and new to Jelly fish but both myself and a
> friend of mine who consider ourselves pretty solid bg players think
>
We've been through this several times over in here, so I'll be brief
this time. By experimenting with the dice generator you can prove to
yourself that it doesn't cheat. This will show that it's not who's on
roll, what the pos looks like etc that decides the roll. Read the help
file (or the manual if you register) for details.
--
- Fredrik Dahl
Kate McCollough
OK...my turn to toss in my $0.02...Nothing new here, but....
Just as with all the hootin' and a hollerin' about FIBS dice, all I
have seen is anecdotal evidence, which is really no evidence at all.
When Patti and then Stephen actually did statistic-based studies of
the FIBS dice, all was well. Instead of telling stories about your
and your buddy's experiences with Jelly, do some valid studies of the
program before you claim cheating. Take a look at Jelly's FIBS
rating. Jelly doesn't roll its own dice there. I know it's a
fruitless argument,though, cause I can hear all the FIBS/Jelly
collusion theories now....*sigh*
When I care to pay attention, I learn alot from Jelly. Try studying
the moves you believe "no human" would make.. ya might be surprised.
You also might want to check out Woolsey's & Heinrich's "New Ideas in
Backgammon" for discussions of moves in which Jelly taught the human
experts a thing or two. I bet ya won't though......
Kate
colin_...@globalvillage.com (Colin Johnson) wrote:
>On Tue, 10 Jun 1997 20:11:39 GMT, da...@usga.org (dave magyar) wrote:
(snip)
>>i would just like to add my 2.5 cents worth... when i first started
>>playing jellyfish i encountered an abnormal amount of successful rolls
>>for the computer.
>>
>>a friend of mine had the same experience. he was getting beat left
>>and right.
>Hi,
>I'm new to this newsgroup and new to Jelly fish but both myself and a
>friend of mine who consider ourselves pretty solid bg players think
>it is a cheating son-of-a-bitch.
>I've yet to verify this by the manual dice rolling test you mention
>but it CANNOT possibly get some of the rolls it comes up with.
>In addition some of its tactics are just plain crazy??? like hitting
>me in it's home base when it is winning and it cannot cover the
>counter? Why do this - no human opponent would even consider it,
>espescially when I had four out of six of my home blocks made.
(snip)
"Fate laughs at probabilities." "I dwell in Possibility-"
-Eugene Aram -Emily Dickinson
Kate McCollough e-mail: mcc...@northcoast.com FIBS: McCool
WWW: http://www.northcoast.com/~mccool
Brian Sheppard
Colin Johnson <colin_...@globalvillage.com> wrote in article
<33b9a7e5...@news.globalvillage.com>...
> On Tue, 10 Jun 1997 20:11:39 GMT, da...@usga.org (dave magyar) wrote:
>
> I'm new to this newsgroup and new to Jelly fish but both myself and a
> friend of mine who consider ourselves pretty solid bg players think
> it is a cheating son-of-a-bitch.
One thing I have noticed is that JF wins a lot more come-from-behind
races than I do. This is particularly galling... After I close out
JF, or have it primed, it then gets the big doublets just as I break,
and then it wins the race.
I thought this was peculiar, until I realized that I win many more
games with last-minute shots than JF does. And then I realized what
was going on.
I have concluded that this is a manifestation of JF's style and my
style. JF deliberately plays to win the race in such situations,
whereas I play to hit a shot.
Every backgammon player has his own way of playing to benefit from
certain possible dice throws. JF clearly understands how to benefit
if big doublets are ever thrown. I understand how to benefit from
fluky shots.
> I've yet to verify this by the manual dice rolling test you mention
> but it CANNOT possibly get some of the rolls it comes up with.
You must agree that it is actually possible that the dice are random.
By the way, JellyFish is the strongest player on FIBS, where it does
not roll its own dice. Think about it: JF has a higher rating than Kit
Woolsey...
> In addition some of its tactics are just plain crazy??? like hitting
> me in it's home base when it is winning and it cannot cover the
> counter? Why do this - no human opponent would even consider it,
> espescially when I had four out of six of my home blocks made.
JF's understanding of blitz tactics are exceptional. Human players now
understand much more about attacking play (and "tempo hits") as a
result of JellyFish and other computer programs.
I believe that JF's checker play in attacking positions is superior
to any human player's.
> I would suggest however that the author re-evaluates the net topology
> or the training algorithm or whatever is making it behave the way it
> does. It is ANYTHING but natural.
I suggest that the author has done a great job, and I would strongly
discourage him from changing anything without a great deal of thought.
Warm Regards,
Brian
Mark Damish
helmet (acu...@es.co.nz) wrote:
: i am around 1750 on fibs. downloaded jellyfish and just dont trust it.
: is it that good at bg or does it cheat ?
: always seems to get good roll after double by either side and
: throws especially well in crucial games . i played it solid for about 3 weeks
Helmet,
It is quite good. If anybody really believes that it cheats, please
stop by Boston, we'll go out and purchase precision dice and a baffle
box, and I'll back the program in $$ games.
--
...Mark Damish mda...@bbn.com
Stephen Turner
Martin Lee wrote:
>
> PS Maybe if the adherents to the JF Cheat Conspiracy included their
> current FIBS rating a lot could be revealed. I wonder why Kit Woolsey
> has not complained about this cheating ?! :)
>
It's obvious why, isn't it? He's involved with Fredrik and marvin in the
conspiracy.
--
Stephen Turner sr...@cam.ac.uk http://www.statslab.cam.ac.uk/~sret1/
Statistical Laboratory, 16 Mill Lane, Cambridge, CB2 1SB, England
"As always, it's considered good practice to temporarily disable any
virus detection software prior to installing new software." (Netscape)