A Reply to Mason Malmuth ( censored at 2+2 )
Mason Malmuth
back. He is welcome to come to Las Vegas and sit down with me and discuss
any topic that he wishes. He can then report back to this group.
Mason Malmuth
HitTheFlop wrote:
> D Taylor <ev...@home.com> writes:
>
> >I previously tried to post the following reply to Mason Malmuth on his
> >2+2 Forum, unsuccessfully.
> >
> >The issue: Mason denies having stated that the Sklansky Hold'em hand
> >rankings have been changed.
>
> Few people I know take criticism worse than Malmuth. He holds
> his opinions as beyond debate and his great work suffers from
> the damage done by his colossal ego.
>
> Many folks believe that the original hand ranking found in HFAP
> were generated using seat of the pants intuition and perhaps a
> pinch of eye of newt. These hand rankings are not far off in the
> main but I'm certain they suffer in certain areas.
>
> Dick Taylor has done some of the finest modern work on starting
> hand selection. Anyone not acquainted with it is likely giving away
> money at every session. The essence of it is that small suited
> connectors are not playable in most situations. I know this to
> be gospel. Can I prove it? No, but I don't need to be convinced
> at this point.
>
> S & M will eventually come out with starting hand standards that
> include more scientific research. Dick Taylor will get zero credit.
> The new standards will not be because of previous errors but
> because "game conditions have changed". You heard it here
> first, well, second anyhow.
>
> Best Luck,
> Ed (no, not that Ed!)
>
> I drive way too fast to worry about cholesterol.
Heldar
"I previously tried to post the following reply to Mason Malmuth on his
2+2 Forum, unsuccessfully."
If this is true, it should serve as a warning to all who want a moderated
group. It is extremely difficult for anyone to be totally objective, and
therefore it is extremely difficult for any moderator to be fair all the
time, and therefore it is almost inevitable that valuable material will be
excluded by any moderator, at least occasionally. I have read Dick Taylor's
work, and I think I understand it, but I am not qualified to comment on its
correctness. Certainly it is plausible enough that it should be considered
by qualified people, rather than excluded because it contradicts the gospel.
(Oh, my. I'm starting to sound like the Grantor...)
Bill
HitTheFlop <hitth...@aol.com> wrote in message
news:19990112234049...@ngol08.aol.com...
>
>Few people I know take criticism worse than Malmuth. He holds
>his opinions as beyond debate and his great work suffers from
>the damage done by his colossal ego.
>
>Many folks believe that the original hand ranking found in HFAP
>were generated using seat of the pants intuition and perhaps a
>pinch of eye of newt. These hand rankings are not far off in the
>main but I'm certain they suffer in certain areas.
.........etc....
Relax, I doubt that many people take S&M as gospel. It's just
a good starting guide to newbies and the mathematically
naive. Let's face it, the math behind poker "analysis" is just
freshman stats (if that), and the "psychological" chapters
in the poker books are just guess work. (E.g., is my opponent
working on the 2nd, or third, or mebbe 4th level of analysis,
but his play would be the same whether he is on the 2nd or the
4th, or even zeroth).
During the few times I've played in Vegas, I didn't find MM to have
a colossal ego; he's just an average mid-level pro, very quiet and
tight, who probably occasionally capitalizes on his auteur status
to steal pots.
Bill
Mason Malmuth
D Taylor wrote:
> I previously tried to post the following reply to Mason Malmuth on his
> 2+2 Forum, unsuccessfully.
>
We have not deleted any posts by Dick Taylor. We also do not appreciate the
implication that we delete posts that are critical of us. We don't. Anyone
is welcome to come to our web page at www.twoplustwo.com and check the forum
and our archives. You will see numerous posts critical of us. By the way,
for those of you who are not familiar with our web page there is some great
information on our forum as well as on our essay section.
> The issue: Mason denies having stated that the Sklansky Hold'em hand
> rankings have been changed.
This is part of what Mr. Taylor posted on our forum."Since Mason elsewhere
has written that the S&M starting hands published in
Holdem for Advanced Players have been estensively revised you no doubt will
have a leak in your game if you follow those."
I never wrote anything like this and it is not true.
> But, I quote from the Winter 1998 issue of
> The Intelligent Gambler, an article titled "Examining the Hand Rankings"
> by Mason Malmuth: "...these rankings have gone through several
> changes..."
The rankings originally appeared in HOLD "EM POKER by David Sklansky which
was published in 1976. In 1980 they were revised in his book ESSAYS ON POKER
to reflect structure changes as well as players getting better. (ESSAYS ON
POKER is now included in the book SKLANSKY ON POKER.) Since 1980 David has
made several minor changes. For instance when HPFAP came out in 1988 he
moved KTs from a Group 5 to a Group 4.
> Further, he states "I'm now going to reveal a secret. When I
> play I don't think in terms of the Sklansky Hand Groups. Neither does
> David Sklansky." He goes two examples, however, which are interesting
> only be reference to those starting hand groups.
The purpose of the article was to show that the hand rankings actually can
vary depending on the situation. There are numerous examples of this in
HPFAP.
> The implication, of
> course, is that anyone who takes the S&M hand rankings/groups too
> seriously is likely to have a leak in their game .. according to Mason
> Malmuth. We have, of course, re-examined these hand rankings in our
> study, available at
> http://www.cs.cmu.edu/~mummert/poker
> Our conclusion is that Sklansky merely needs to edit his list of
> starting hold'em hands in order to have the hands ranked in terms of
> their mathematical power across the full range of table playing
> conditions and numbers of players.
>
I want to state again that the Taylor model does a very poor job of
imitating real poker. I do not recommend any of the advice that he gives.
Mason Malmuth
D Taylor
2+2 Forum, unsuccessfully.
The issue: Mason denies having stated that the Sklansky Hold'em hand
rankings have been changed. But, I quote from the Winter 1998 issue of
The Intelligent Gambler, an article titled "Examining the Hand Rankings"
by Mason Malmuth: "...these rankings have gone through several
changes..." Further, he states "I'm now going to reveal a secret. When I
play I don't think in terms of the Sklansky Hand Groups. Neither does
David Sklansky." He goes two examples, however, which are interesting
only be reference to those starting hand groups. The implication, of
course, is that anyone who takes the S&M hand rankings/groups too
seriously is likely to have a leak in their game .. according to Mason
Malmuth. We have, of course, re-examined these hand rankings in our
study, available at
http://www.cs.cmu.edu/~mummert/poker
Our conclusion is that Sklansky merely needs to edit his list of
starting hold'em hands in order to have the hands ranked in terms of
their mathematical power across the full range of table playing
conditions and numbers of players.
Dick Taylor
HitTheFlop
D Taylor <ev...@home.com> writes:
>I previously tried to post the following reply to Mason Malmuth on his
>2+2 Forum, unsuccessfully.
>
>The issue: Mason denies having stated that the Sklansky Hold'em hand
>rankings have been changed.
Few people I know take criticism worse than Malmuth. He holds
his opinions as beyond debate and his great work suffers from
the damage done by his colossal ego.
Many folks believe that the original hand ranking found in HFAP
were generated using seat of the pants intuition and perhaps a
pinch of eye of newt. These hand rankings are not far off in the
main but I'm certain they suffer in certain areas.
Dick Taylor has done some of the finest modern work on starting
OnIQQker
>
>"I previously tried to post the following reply to Mason Malmuth on his
>2+2 Forum, unsuccessfully."
>
>group. It is extremely difficult for anyone to be totally objective, and
>therefore it is extremely difficult for any moderator to be fair all the
>time, and therefore it is almost inevitable that valuable material will be
>excluded by any moderator, at least occasionally. I have read Dick Taylor's
>work, and I think I understand it, but I am not qualified to comment on its
>correctness. Certainly it is plausible enough that it should be considered
>by qualified people, rather than excluded because it contradicts the gospel.
>
>(Oh, my. I'm starting to sound like the Grantor...)
>
You're right. It's easy for a moderator to fall into the trap of only allowing
material to appear if he agrees with it. I was a member of a list which was run
by a young lady who was almost nazi-like in her approach. SHe would rail about
postings that were too short, about postings that were off-topic, and generally
made everyone on the list miserable. People who criticized her were summarily
unsubbed.
I expect the TexasHoldem list to be a lot different. Since I am new to poker, I
have no axes to grind; no pre-conceived notions about who is right or more
right. I expect people on the list to get frustrated with each other from time
to time, but that's just part of the fun.
More than anything else, I just want to know that I'm talking and listening to
real poker players.... not cranks, salesmen and scammers. I also want my own
comments and experiences to be shared with people I feel comfortable with. The
mailing list has arhives, but the archives are available only to members. I've
never been real comfortable with the idea that everything I post to a NG is
freely available to anyone who cares to to a search on my name.
Les
Like holdem? Subscribe to the Texas Hold 'em mailing list by sending a blank
e-mail to: TexasHolde...@onelist.com
Paul Phillips
Mason Malmuth <masonm...@twoplustwo.com> wrote:
>Let me make an offer that I made in private to Mr. HitThe Flop a while
>back. He is welcome to come to Las Vegas and sit down with me and discuss
>any topic that he wishes. He can then report back to this group.
Is there some reason understandable by mortals why someone should have
to go to Vegas to ask you a question? Have the other great geniuses of
our time taken this approach?
"Dr. Bohr, your model of the atom is hogwash!"
"Otto, I'll tell you the same thing I told you last time: if you're
such a damn genius, why don't you come to Copenhagen and convince me
in my lab. I'll even buy lunch."
At least Hardy brought Ramanujan to Cambridge himself rather than
smugly challenging him to find his own way there.
--
Paul Phillips | been around the world and found that only stupid
Mordant Surfer | people are breeding, the cretins cloning and feeding,
<pa...@go2net.com> | and i don't even own a tv. -- harvey danger
+1 206 447 1595 |--------* http://www.go2net.com/people/paulp/ *--------
D Taylor
> We have not deleted any posts by Dick Taylor.
Six hours after I sent it, the message still had not appeared
in the 2+2 Forum. I suppose it depends on the your meaning
of "delete." ...
> We also do not appreciate the implication that we delete posts
> that are critical of us. We don't.
Oh yeah? Anyone is welcome to use dejanews to find previous rgp
complaints about having their posts to 2+2 killed, filtered and/or
deleted. There have been several.
> Anyone is welcome to come to our web page .. You will see numerous
> posts critical of us.
Hmmmm. While I don't really want to waste much time on this, I
did a scan of current messages and didn't find a single one.
<snip history>
> The purpose of the article was to show that the hand rankings
> actually can vary depending on the situation.
The hand rankings are a basis for making pre-flop playing
decisions. If properly chosen, using a probability model that
factors in the range of possible player styles before the flop
and varying the number of players in the game, the rankings
DO NOT actually change "depending on the situation."
Whatever method Sklansky used to arrive at his starting hand
rankings, it is remarkably accurate. I recommend them to Mason.
> I want to state again that the Taylor model does a very poor job of
> imitating real poker. I do not recommend any of the advice that he
> gives.
Thank you once again for your summary dismissal. I'll put it in
the same file where I tossed your current essay on the use of
personal computers ...
Dick Taylor
Sean Duffy
>Mason Malmuth wrote:
>> We have not deleted any posts by Dick Taylor.
>
> Six hours after I sent it, the message still had not appeared
> in the 2+2 Forum.
That's not a long turn over time for any forum, be it Web based or
Usenet.
Jaeger T. Cat
Mason Malmuth <masonm...@twoplustwo.com> wrote:
>Let me make an offer that I made in private to Mr. HitThe Flop a while
>back. He is welcome to come to Las Vegas and sit down with me and discuss
>any topic that he wishes. He can then report back to this group.
>
>
What's the matter, Mason? Can't debate in an _open_ forum, with
peer review?
Wuss.
Kim Scheinberg
>In article <369C280A...@twoplustwo.com>,
>Mason Malmuth <masonm...@twoplustwo.com> wrote:
>>Let me make an offer that I made in private to Mr. HitThe Flop a while
>>back. He is welcome to come to Las Vegas and sit down with me and discuss
>>any topic that he wishes. He can then report back to this group.
>
>to go to Vegas to ask you a question? Have the other great geniuses of
>our time taken this approach?
>At least Hardy brought Ramanujan to Cambridge himself rather than
>smugly challenging him to find his own way there.
Well, Doyle *did* refuse to sign SuperSystem, insisting Bill Gates come
sit in the game to get the autograph...
-k. who, at 27 weeks, feels more like the Mountain than Mohammed
11seat
>
> Let me make an offer that I made in private to Mr. HitThe Flop a while
> back. He is welcome to come to Las Vegas and sit down with me and discuss
> any topic that he wishes. He can then report back to this group.
>
______________________________________________________
Bob (Snookums) W. queried:
Is this offer open to someone who already admires and respects your
work, but may have a few questions? I would personally feel privileged,
but obligated to be objective.
I don't know if I would be qualified to dispute anything, but I could
report.
______________________________________________________
"Sometimes it's a game of cards played with money and
sometimes it's a game of money played with cards."
Jim Geary
whatever reason everyone jumps on the bandwagon. Let's apply some
critical thinking to what was actually written...
On Wed, 13 Jan 1999, D Taylor wrote:
> Mason Malmuth wrote:
> > We have not deleted any posts by Dick Taylor.
>
> Six hours after I sent it, the message still had not appeared
> in the 2+2 Forum. I suppose it depends on the your meaning
> of "delete." ...
What exactly does this mean, Dick? You "sent it"? One merely
posts. One does not send. Did you post a message thru the standard
interface and receive a message that says, "Your message has been posted"?
And after receiving this confirmation, it was subsequentely deleted
because the ideas went against MM's beliefs? I do not believe this.
Please tell us what you did EXACTLY. You are accusing 2+2 of deleting
your post but you are vague on the details of posting it ("sending..").
I (as well as anyone from rgp who visits 2+2) has seen numerous
examples of inflammatory posts (eg. you guys would sell more books
if you weren't such assholes) that remain on their board. Abdul
has questioned the starting hand schema at great length. His posts
have not been deleted. Are we to believe that your starting hand
analysis alone was so controversial as to meet deletion? I do not
believe this either.
> > We also do not appreciate the implication that we delete posts
> > that are critical of us. We don't.
>
> Oh yeah? Anyone is welcome to use dejanews to find previous rgp
> complaints about having their posts to 2+2 killed, filtered and/or
> deleted. There have been several.
I know of one valid example that was validly deleted (the vanity
press thing). If the rest of the complaints in the DJN archive show as
much merit as I believe your complaint does, I would not take these as
statistically significant evidence.
> > Anyone is welcome to come to our web page .. You will see numerous
> > posts critical of us.
>
> Hmmmm. While I don't really want to waste much time on this, I
> did a scan of current messages and didn't find a single one.
This language is characteristic of evasiveness. I found a whole thread
regarding "if Sklansky is so damn smart etc..." Numerous other examples
are in the archives (which kick in after just one week due to the volume
of the board). Some of these are blatant ad hominems. Yet they remain.
But your belief in your own hand rankings is so controversial that it
merits deletion. Exercise for the readers: Apply Bayes' Theorem and
tell me which is more likely, that all these inflammatory posts were
retained yet Dick Taylor's hand-ranking post was deleted, or that there
was in fact no hand-ranking post.
> > The purpose of the article was to show that the hand rankings
> > actually can vary depending on the situation.
I'm not familiar with the article in question, but I think this
statement is one we can all agree is true. This is certainly
addressed in HPFAP.
> The hand rankings are a basis for making pre-flop playing
> decisions. If properly chosen, using a probability model that
> factors in the range of possible player styles before the flop
> and varying the number of players in the game, the rankings
> DO NOT actually change "depending on the situation."
Oops, I spoke too soon above. Any standardized test taker knows
that a good answer to have in your bag is, "you need more information
to answer this question." Poker seems like the ultimate environ
for making this statement. Yet there is disagreement. I would
enjoy a debate on the analysis of this situation. Please post it
to 2+2 for critical debate. If you are worried that it will be
deleted, send it to me and I will post. If they delete my post,
I will give you $100 no strings attached. The same offer applies
for the hand-ranking debate. I can think of numerous reasons to
substantiate my position that pre-flop play is a situationally dependent
problem, and would enjoy discussing this here or on the 2+2 forum
(which I prefer due to my non-thread newsreader interface).
Regards,
Jim Geary
jaygee at primenet dot com
http://www.primenet.com/~jaygee/
Maverick
> Let me make an offer that I made in private to Mr. HitThe Flop a while
> back. He is welcome to come to Las Vegas and sit down with me and discuss
> any topic that he wishes. He can then report back to this group.
Why would he want to do that, Mason? From what I've heard, you'd just
call'em a liar and probably accuse him of being in league with the Caro.
gary_...@my-dejanews.com
Mason Malmuth <masonm...@twoplustwo.com> wrote:
>
>
> D Taylor wrote:
>
> > I previously tried to post the following reply to Mason Malmuth on his
> > 2+2 Forum, unsuccessfully.
> >
>
> implication that we delete posts that are critical of us. We don't. Anyone
Yes you do Mason. You deleted one of mine a while back.
> is welcome to come to our web page at www.twoplustwo.com and check the forum
> and our archives. You will see numerous posts critical of us. By the way,
So, you don't delete them all.
> > The issue: Mason denies having stated that the Sklansky Hold'em hand
> > rankings have been changed.
>
> This is part of what Mr. Taylor posted on our forum."Since Mason elsewhere
> has written that the S&M starting hands published in
> Holdem for Advanced Players have been estensively revised you no doubt will
> have a leak in your game if you follow those."
>
> I never wrote anything like this and it is not true.
He didn't say you wrote this, Mason. He says that what you did write implies
this. And, he's correct.
>
> I want to state again that the Taylor model does a very poor job of
> imitating real poker. I do not recommend any of the advice that he gives.
>
One of the problems Mason is that what you think is real poker isn't played
much outside of Las Vegas. The game characteristics elsewhere tend to be
different -- what you think of as a typical game and what most people
experience as a typical game are just different.
It's really sad to watch you keep trying to defend the indefensible, Mason.
Gary Carson
>
-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
gary_...@my-dejanews.com
pa...@go2net.com (Paul Phillips) wrote:
> In article <369C280A...@twoplustwo.com>,
> Mason Malmuth <masonm...@twoplustwo.com> wrote:
> >Let me make an offer that I made in private to Mr. HitThe Flop a while
> >back. He is welcome to come to Las Vegas and sit down with me and discuss
> >any topic that he wishes. He can then report back to this group.
>
>
> "Otto, I'll tell you the same thing I told you last time: if you're
> such a damn genius, why don't you come to Copenhagen and convince me
> in my lab. I'll even buy lunch."
I don't think Mason is offering to buy anyone lunch.
Maverick
> enjoy a debate on the analysis of this situation. Please post it
> to 2+2 for critical debate. If you are worried that it will be
> deleted, send it to me and I will post. If they delete my post,
> I will give you $100 no strings attached. The same offer applies
hey Jim...does this $100 offer apply to me if I can get one of my posts
deleted?
D Taylor
You have my permission to copy and post on the 2+2 Forum, for
discussion, the hold'em study available at
http://www.cs.cmu.edu/~mummert/poker
Please use my $100 to buy your beautiful daughter a puppy.
Dick Taylor
Heldar
"I want to state again that the Taylor model does a very poor job of
imitating real poker. I do not recommend any of the advice that he gives."
Those of us who are not capable of making our own models must try to decide
whose models are correct, and then try to learn to follow them. Could you
explain why Mr. Taylor's model does not imitate real poker, and what type of
model you would suggest instead?
Barbara Yoon
> We have not deleted any posts by Dick Taylor. We also do not
> appreciate the implication that we delete posts that are critical of us.
>
> by David Sklansky...1976. ...1980...revised in...ESSAYS ON POKER
> to reflect structure changes as well as players getting better. Since
> 1980 David has made several minor changes. ...HPFAP...1988...
> ...he moved KTs from a Group 5 to a Group 4.
>
> I want to state again that the Taylor model does a very poor job of
> imitating real poker. I do not recommend any of the advice that he gives.
OK...to clarify here, please tell us how K-10-suited was ranked relative
to J-10-suited (that is, higher or lower) in 1976 and 1980 and 1988?!
HitTheFlop
"Heldar" <hel...@ix.netcom.com> writes:
>Mason Malmuth wrote:
>
>"I want to state again that the Taylor model does a very poor job of
>imitating real poker. I do not recommend any of the advice that he gives."
Heldar asks:
>Those of us who are not capable of making our own models must try to decide
>whose models are correct, and then try to learn to follow them. Could you
>explain why Mr. Taylor's model does not imitate real poker, and what type of
>model you would suggest instead?
This gets str8 to the nub of the problem. We won't get a worthwhile
answer to this even if I personally crawl to Vegas.
Alan Bostick
> I previously tried to post the following reply to Mason Malmuth on his
> 2+2 Forum, unsuccessfully.
[etc. snipped.]
If the maintainers of the 2+2 Forum did indeed delete Dick Taylor's post,
they did so in error.
Here is an excerpt from 2+2's Acceptable Use Policy:
"In this vein, we ask all contributors to make their posts:
1.Gambling-related,
2.Strategic in nature,
3.Possess a minimum of courtesy."
Dick's alleged post on hand selection was surely gambling-related, and
equally surely strategic in nature. As for point three, not having seen
his actual post I can only hazard a guess, but if it were at all like
what he posts to r.g.p., it would in fact possess an absolute minimum of
courtesy.
Alan "Definitely NOT speaking for Mason Malmuth" Bostick
--
Alan Bostick |
mailto:abos...@netcom.com | "But sex *is* politics!" -- Gore Vidal
news:alt.grelb |
http://www.alumni.caltech.edu/~abostick
Jim Geary
Due to the length, I was forced to break it into two parts.
The msgid's are 12634, and 12635.
A very interesting work. I look forward to the debate.
Heldar
tough game? I would think they go up in value in a loose, wild game,
because they have a better chance to become a monster hand, which is
sometimes necessary to win a "family pot". Or am I missing something?
David Monaghan
wrote:
>Dick's alleged post on hand selection was surely gambling-related, and
>equally surely strategic in nature. As for point three, not having seen
>his actual post I can only hazard a guess, but if it were at all like
>what he posts to r.g.p., it would in fact possess an absolute minimum of
>courtesy.
Alan, you have quite a way with words - but - haven't you criticised
*me* up in the past for gratuitous rudeness?
DaveM
Dsklansky
fold even top pair under pressure. Gary Carson correctly points out that this
mainly describes only Vegas games. However under this criteria it should be
obvious why suited connectors go up in value in many situations.
gre...@aol.com
I play at Bellagio 3 - 4 times a week. I use the Tayloy starting hands very
successfully. I have only been playing poker for a few months, but I am a
winning player, at the low limit levels. I have read most of the books -
Jones, most of the Malmuth and Sklansky books, etc. I would love to sit and
learn about the starting hands and why you feel the Taylor hands are wrong?
Please accept this invitation.
> Mason Malmuth <masonm...@twoplustwo.com> wrote:
> >Let me make an offer that I made in private to Mr. HitThe Flop a while
> >back. He is welcome to come to Las Vegas and sit down with me and discuss
> >any topic that he wishes. He can then report back to this group.
>
-----------== Posted via Deja News, The Discussion Network ==----------
HitTheFlop
"Heldar" <hel...@ix.netcom.com> writes:
>I'm kind of slow, I guess.
Compaired to those on Mt Olympus we
are all slow. Don't expect elaboration.
JP Massar
wrote:
My guess:
Because of the semibluffing possibilities they present.
A hand like 76s will hit more flops in some (albeit small) way than a
hand like AT.
gary_...@my-dejanews.com
dskl...@aol.com (Dsklansky) wrote:
> My rankings assume players will bluff in a tough manner and and will sometimes
> fold even top pair under pressure. Gary Carson correctly points out that this
> mainly describes only Vegas games. However under this criteria it should be
Actually David, I think it also works pretty well in a few other places --
the mid limit games on the Miss. Gulf Coast is one place where the games are
often very similar in condition to those you might find in Las Vegas. That
wasn't the case before Lousiana got poker, but now that few Texans are
playing in Miss, the mid-limit (10/20 and up) games along the gulf coast have
been pretty tough the few times I've visited. NO games are similar also.
Shreveport and Lake Charles, however aren't even close to what you might find
in LV.
Which is what I think is wrong with Dick Taylors hand rankings. It looked to
me like he tried to compute some kind of average rank across different game
conditions. This will likely give you rankings that aren't right for any
game -- they just are off by some consistent (least squares) amount.
Sklansky's groups are pretty accurate for the game they are aimed at. Not,
for other games.
Overall, meaningful ranks are just not attainable. But, ranks for specific
conditions are attainable. So, I don't think Taylor's approach can possibly
give you correct rankings. Although I do find it interesting.
Gary Carson
Dsklansky
in tough games.
D Taylor
> My rankings assume players will bluff in a tough manner and and will
> sometimes fold even top pair under pressure. Gary Carson correctly
> points out that this mainly describes only Vegas games. However under
> this criteria it should be obvious why suited connectors go up in
> value in many situations.
Gary Carson wrote: <snip Gulf Coast anecdotes>
> ...Which is what I think is wrong with Dick Taylors hand rankings.
> It looked to me like he tried to compute some kind of average rank
> across different game conditions. This will likely give you
> rankings that aren't right for any game -- they just are off by
> some consistent (least squares) amount. Sklansky's groups are pretty
> accurate for the game they are aimed at. Not, for other games. ...
> So, I don't think Taylor's approach can possibly give you correct
> rankings. Although I do find it interesting.
The myth of suited connectors has, as I pointed out in our study, an
element of truth in it. It is that at tight tables (whether tough, or
not, such as T1-T2) a hand like 87s has measurably more value than other
hands that might normally outrank it, such as T9s. If play at the tight
table is so tough that the pots become worthwhile, then there will no
doubt be isolated occasions when 87s will seem worthwhile. But having
noted this element of truth, I must also note that the myth consists in
the belief that the entire set of small and medium suited connectors
therefore become playable. No savvy player will stay long at a very
tight table. Instead, the wise thing is to simply assume that small and
medium suited connectors are just small and medium ranking hands. Even
uder their very best overall table conditions they should only be played
by gamblers.
Dick Taylor
Tad Perry
And the flops it doesn't hit you can pretend.
tvp
Asha34
disrespect to MM intended). Thank you.
Ashley
Asha34
me some solace. After all, if people who know so much more than I about poker
disagree about such fundamental things then maybe my intuitive approach, and
uncertainty about the perfect poker play, isn't so far off the mark.
Thanks guys.
D Taylor
> The reasoning is similar to why in lowball draw weak pat hands go
> down in value in tough games.
We showed in our study that this reasoning is correct in hold'em, where
merely playing tough can have positive EV at loose tables ... unless the
other players are chimpanzees or otherwise are not self-aware.
Dick Taylor
Irish Mike
stud games at all limits.
Irish Mike
--
Chuck Weinstock
>Mason Malmuth wrote:
>> We have not deleted any posts by Dick Taylor.
>
> in the 2+2 Forum. I suppose it depends on the your meaning
> of "delete." ...
>
Dick,
Anything posted to the forum should show up essentially instantly. If
you reload the forum you should see your article. If you do not see
your article then something went wrong technically and I'd like to
know about it so that I can track down the problem.
Chuck
D Taylor
If my earlier post wasn't censored, then it is possible that I
neglected to have scrolled down to the transmit button. I wasn't used to
posting to 2+2 and may have exited the edit session prematurely. This
happened to my current reply to Mason, which I re-posted. It might be
more intuitve of a GUI to have the message options all precede the
message text window.
Dick Taylor
Andrew Prock
I was inspired to read the Taylor report on poker after seeing
the controversy that the 2+2 problem had caused. I think most
poker players who are serious about playing should read this
report.
The report is important for several key reasons:
1) It measures hands in an analytic and reproducible way
2) The games are played with some kind of strategy
3) Variance in play and opponents is measured
4) The work is (sort of) public and thus reviewable
In general, more poker research should be done this way.
Given a description of the methods, assumptions, and results
it is much easier to evaluate the creditability of the research
than if the results are simply stated as "truths" or "mantras"
based on experience and common-sense. If one thing is true in
poker, neither experience or common-sense forms the foundation of
the best play, only probability does. (One can make arguments to this,
but I feel that any information gathered must be used to adjust the
foundational plan, not vice-versa.)
This being said, the reports results have dubious relevance to
the actual game of poker, but they do provide a foundation from
which to work. However, it does provide a solid framework for
understanding their strategy, which S&M does not provide.
What is interesting is that the results are close to those of S&M,
and only vary greatly in the poor range of playable opening hands,
giving Ax and Kxs more weight and medium-small suited connectors
little or no weight.
Both Taylor and S&M generate a chart of all the possible opening
hold-em hands. This chart is then used as a basis for forming a
full strategy. Taylor does little more than generate a new chart,
and challenge some of the fine points of the S&M strategy. In this
respect the research falls far short of the S&M work. It does not
propose a new strategy, but rather a variant of that recommended by
S&M. Another point is that these strategies should only be used
as a basis for playing poker. By no means should they be played
at face value for every hand.
The natural question is whether the Taylor chart is better or worse
than the S&M chart. One may also ask whether some of the other
changes to the strategy are also.
Clearly, under the simulation conditions given, the Taylor table
is the best strategy. However, despite the general quality of
the simulation, it is deficient in many ways. These deficiencies
include:
1) Foe's all use an identical "old" strategy (denoted OS)
2) Foe's all play with the same degree of looseness
3) Foe's don't bluff at all.
4) Position play is not well defined/distinguished
5) Simplistic post-flop strategy
6) The Hero always plays to the end
Deficiencies 1 & 2
The "old" strategy (OS) is not defined in the text, but one can
safely assume that it is close to the S&M strategy. The problem
with every opponent using the same strategy at the same looseness
is that this is a very artificial table. One would only expect
to find a table like this in a game dominated by locals and
good tourists who have read, and identically understood the
poker literature. If such a table was found, it probably would
not be the table to play at. The only condition where this might
be a table that should be played would be at a tournament.
What is more common is to have a spectrum of talent and tightness.
Taylor's research would have been much more convincing if three or
four common types of players were represented. Aggregate skill and
looseness should also have been used instead of absolute looseness.
The study would have benefited from having mean looseness ratings
with a looseness variance.
The problem, of course, with adding multiple strategies, as well
as a non-zero looseness variance, is that the number of variables
that needs to be tracked increases greatly. This in turn complicates
the analysis.
Deficiencies 3, 4, & 5
When it comes to the actual strategy used by the opponents at the
table, it is surprising that bluffing and position play were not
included in some form. Bluffing, for good or ill, is one of the
most common tactics used in poker. Bluffing essentially amounts
to representing your hand as better than it actually is. This
could easily have been implemented in the simulation. Its
omission greatly detracts from the validity of the results.
Position play is also not well integrated into the strategy. There
is some accounting for it, but simple ploys like stealing the
blind, and defending the blind are not addressed. Admittedly, the
solution they have come up with seems like it could reasonably
handle bets and raises before the flop. I have no proof of this,
but then no one has a proof that their strategy works well either.
The post flop strategy is one of the biggest problems with the
simulation. The basic problem is that the opponents strategy
assumes that they are playing a game with random players. There
is no description of how each opponent deals with raises, re-raises
and possible bluffs. They simply calculate the odds, and do the
"right" thing. In particular it seems that the opponent assumes
that every one else left in the pot has a random hand, instead of
one worth staying in the pot with.
Some of the problems stated above are inherit to poker simulations.
The lack of more complex post-flop strategy and bluffing really
does make the quality of the simulation questionable.
Deficiency 6
The final problem is that the "hero" always stays in till the
end playing a showdown game. The author tries to account for this,
but it is unclear what the method is, and whether the method is
mathematically sound.
Given their results, Taylor does a good job of contrasting
it with the S&M strategy. Taylor comes up with a new opening
pair table which is well suited to playing in the simulation.
Some strategy alterations are also made from S&M. The
comparison with Jones' work is as incomplete as Jones strategy
is. Jones' work is on the other hand meant for beginners.
Taylor also seems to forget the promised comparison with
the Super/System strategies.
Some questions are left after the results are presented:
Why didn't Taylor present the percentages?
Why did Taylor use exactly the same number of cards in each group?
And
Why didn't Taylor simulate the new strategy?
All of these questions are important in understanding poker
and poker simulations. It could be that if the simulations were
rerun with the new strategy in place of OS, that the strategy of
S&M would prove the best. (This is not likely but it does illustrate
the point.)
With a list of the percentages, it would be possible to recast
the groups into more "correct" sizes. There doesn't seem to
be any real evidence that the group sizes must be (5,5,6,8,18,9,17,16).
It might make more sense for them to be (2,5,9,14,18,18,18).
In conclusion, the results presented by Taylor are interesting, and
should probably be investigated further. In particular, the evidence
that Ax, and Kxs should be played more often than the small suited
connectors is interesting and bears some definite investigation, both
from a theoretical probabilistic standpoint, and a simulation standpoint.
Again it is interesting to point out that the absolute rankings of
each hand are so close to S&M for the high value hands that one has
to seriously wonder, if this simulation is flawed in some ways, why
would the S&M ranking be any more sound?
In the end, poker is as much of a game of reading the opponents as
it is of hand selection. Armed with only one of these weapons, any
poker player is sure to meet with defeat.
- Andrew
--
"Intuitively obvious to the most casual of observers."
Andrew Prock
pr...@cs.wisc.edu
<a href="http://www.cs.wisc.edu/~prock/">home</a>
D Taylor
> I was inspired to read the Taylor report on poker after seeing
> the controversy that the 2+2 problem had caused. I think most
> poker players who are serious about playing should read this
I never cease being amazed at how an open, analytic approach
to evaluating the starting hands in hold'em could have
generated such heat in the two years since I first posted it
on the internet. I was glad to read your careful,
insightful review of it as a byproduct of that nonetheless.
I have just a few comments.
Identical between hands except for position and table looseness.
> 2) Foe's all play with the same degree of looseness
This is true for foes at the same table. But looseness is varied
systematically from table to table, so there is inter-table
looseness variance.
> 3) Foe's don't bluff at all.
Since the target hand will always go to the river, the only
potential impact of foes bluffing will be to succeed/fail in
bluffing other foes. This would show up as randomizxed noise
in our measures of hero hands.
> 4) Position play is not well defined/distinguished
Only pre-flop position play, as to whether or not to fold,
is defined.
> 5) Simplistic post-flop strategy
Yes. Nothing fancy. Foes just continue with playable hands.
> 6) The Hero always plays to the end
This is, of course, by design. It is the only way to directly
measure the limiting conditions of a hand bluffing successfully
(or not) as table looseness varies
>
> Deficiencies 1 & 2
>
> The "old" strategy (OS) is not defined in the text, but one can
> safely assume that it is close to the S&M strategy. The problem
> with every opponent using the same strategy at the same looseness
> is that this is a very artificial table. One would only expect
> to find a table like this in a game dominated by locals and
> good tourists who have read, and identically understood the
> poker literature. If such a table was found, it probably would
> not be the table to play at. The only condition where this might
> be a table that should be played would be at a tournament.
>
> What is more common is to have a spectrum of talent and tightness.
> Taylor's research would have been much more convincing if three or
> four common types of players were represented. Aggregate skill and
> looseness should also have been used instead of absolute looseness.
> The study would have benefited from having mean looseness ratings
> with a looseness variance.
Interesting idea. It is, of course, hard to (a) imagine exactly
what aggregate could be regarded as being representative, or
(b) whether the interactions between foes generated thereby
could have better qualified the hero hand measurements.
>
> The problem, of course, with adding multiple strategies, as well
> as a non-zero looseness variance, is that the number of variables
> that needs to be tracked increases greatly. This in turn complicates
> the analysis.
>
> Deficiencies 3, 4, & 5
>
> When it comes to the actual strategy used by the opponents at the
> table, it is surprising that bluffing and position play were not
> included in some form. Bluffing, for good or ill, is one of the
> most common tactics used in poker. Bluffing essentially amounts
> to representing your hand as better than it actually is. This
> could easily have been implemented in the simulation. Its
> omission greatly detracts from the validity of the results.
Foes would not ever succeed in bluffing the hero, of course,
and the interactions between foes would, we think, have led
to overestimating the value of bluffing by hero hands. We
think the bluffing potential for each starting hand can be
estimated as the difference between its n-player showdown
p(win) and its actual p(win) for a given table when all
players have folded by the river.
>
> Position play is also not well integrated into the strategy. There
> is some accounting for it, but simple ploys like stealing the
> blind, and defending the blind are not addressed. Admittedly, the
> solution they have come up with seems like it could reasonably
> handle bets and raises before the flop. I have no proof of this,
> but then no one has a proof that their strategy works well either.
>
> The post flop strategy is one of the biggest problems with the
> simulation. The basic problem is that the opponents strategy
> assumes that they are playing a game with random players. There
> is no description of how each opponent deals with raises, re-raises
> and possible bluffs. They simply calculate the odds, and do the
> "right" thing.
Actually, post flop the foes are not so smart. The play on
with a pair or better, or 1 card draws to straights or flushes.
> In particular it seems that the opponent assumes
> that every one else left in the pot has a random hand, instead of
> one worth staying in the pot with.
Not quite. But you are correct that the foes are not smart
or fancy either.
>
> Some of the problems stated above are inherit to poker simulations.
> The lack of more complex post-flop strategy and bluffing really
> does make the quality of the simulation questionable.
>
> Deficiency 6
>
> The final problem is that the "hero" always stays in till the
> end playing a showdown game. The author tries to account for this,
> but it is unclear what the method is, and whether the method is
> mathematically sound.
Sorry I haven't made that clear. There is only one table (5)
which is a "showdown" game. It can be thought of as a control
condition in which a given hand will win its absolute brute
force number of rounds. At all other (12) tables the hero is
the only one who plays to the river.
>
> Given their results, Taylor does a good job of contrasting
> it with the S&M strategy. Taylor comes up with a new opening
> pair table which is well suited to playing in the simulation.
> Some strategy alterations are also made from S&M. The
> comparison with Jones' work is as incomplete as Jones strategy
> is. Jones' work is on the other hand meant for beginners.
> Taylor also seems to forget the promised comparison with
> the Super/System strategies.
>
> Some questions are left after the results are presented:
>
> Why didn't Taylor present the percentages?
Why didn't Taylor also put the Figure numbers in the report,
or finish writing it? Same answers.
>
> Why did Taylor use exactly the same number of cards in each group?
>
> And
>
> Why didn't Taylor simulate the new strategy?
>
> All of these questions are important in understanding poker
> and poker simulations. It could be that if the simulations were
> rerun with the new strategy in place of OS, that the strategy of
> S&M would prove the best. (This is not likely but it does illustrate
> the point.)
>
> With a list of the percentages, it would be possible to recast
> the groups into more "correct" sizes. There doesn't seem to
> be any real evidence that the group sizes must be (5,5,6,8,18,9,17,16).
> It might make more sense for them to be (2,5,9,14,18,18,18).
Yes. Or others.
>
> In conclusion, the results presented by Taylor are interesting, and
> should probably be investigated further. In particular, the evidence
> that Ax, and Kxs should be played more often than the small suited
> connectors is interesting and bears some definite investigation, both
> from a theoretical probabilistic standpoint, and a simulation standpoint.
> Again it is interesting to point out that the absolute rankings of
> each hand are so close to S&M for the high value hands that one has
> to seriously wonder, if this simulation is flawed in some ways, why
> would the S&M ranking be any more sound?
>
> In the end, poker is as much of a game of reading the opponents as
> it is of hand selection. Armed with only one of these weapons, any
> poker player is sure to meet with defeat.
>
> - Andrew
>
> --
> "Intuitively obvious to the most casual of observers."
> Andrew Prock
> pr...@cs.wisc.edu
> <a href="http://www.cs.wisc.edu/~prock/">home</a>
Dick Taylor
D Taylor
His belief I owe 2+2 an apology for claiming they censored
my post(s). ...
Mason,
Since I believe I and others have had posts to 2+2 censored,
for whatever reasons, but in the last week I have also caught
myself exiting from the 2+2 edit session prematurely once,
I do apologize if I incorrectly concluded that one post was
censored when it was not. The one that never showed up,
however, may not have had your modicum of politeness.
Dick Taylor
holl...@ecn.ab.ca
Having complimented my antecedent, I must now point out that it seems that
he, like most everybody else, has missed the point entirely. I will
concede that this is partly Taylor's fault for his conclusions and strategy
guide derived therefrom. What Taylor tried to do was to separate tactics
(which includes aggression, bluffing, many positional considerations, etc)
from the pure strength inherent in the various starting hands. His
results and method seem very good from that point of view. Unfortunately,
he then tries matchup playability or quality with that inherent strength.
I think this is what leads people to attack his model/results. Andrew and
others speak of the deficiencies of the model (lack of bluffing, ignoring
position, averaging, and playing till the end, etc), but seem to forget or
ignore the fact that Taylor didn't want to find out how to play hands,
he wanted to try to find fundamental, underlying properties of the hands.
Anything further is tactical, and therefore a separate issue.
To determine the quality or playability of a hand, one must combine raw
strength with other, tactical factors, like the fact that real life
opponents won't behave exactly like the model foes. It may even be that
hands like 87s need to have a different concept of fundamental strength
associated with them (ie can't be measured the same way as one measures
the strength of AJo). One would have to qualitatively blend the hands
measured with different units to try to arrive at an overall hand ranking.
Or not. I suspect that is what theorists like S&M do, to some extent.
They "measure" the strength of the hands with sticks which are different
for various hand types, layer on some tactical considerations for each,
and then try to blend the whole mass together into a final, comprehensive
ranking/grouping. Unfortunately (in Taylor's opinion, I think) this
process was more art than science, and Taylor was trying to approach
things more scientifically. Good for him. Get people thinking.
By the way, there is nothing wrong with a simplistic model. It is the
best place to start. If it gives results anywhere close to reality, then
it was a good model, and a starting point for the next one. Since there
is a good deal of agreement with S&M results (our definition of reality
for the moment) at the upper end, then the model chosen by Taylor is a
good one in that sense. The fact that there is a great divergence lower
down could be the result of a first order model, or it could be we are
using the wrong definition of reality. Unfortunately, it seems that both
sides have entrenched themselves so that they are unable to impartially
and rationally examine the situation. In fact, it is probably the case
that both are factors, and the solution lies elsewhere. I don't know.
But I'm thinking about it.
Eric
Heldar
every post is a separate web page, and loading them or moving back and forth
is extremely cumbersome. Am I doing something wrong?
Charles Milton Ling
to sit in his game to get it? Great...
Charley
Kim Scheinberg <ik...@panix.com> wrote in message
news:77iafc$lfo$1...@news.panix.com...
>pa...@go2net.com (Paul Phillips) writes
>>In article <369C280A...@twoplustwo.com>,
>>Mason Malmuth <masonm...@twoplustwo.com> wrote:
>>>Let me make an offer that I made in private to Mr. HitThe Flop a while
>>>back. He is welcome to come to Las Vegas and sit down with me and discuss
>>>any topic that he wishes. He can then report back to this group.
>>
>>to go to Vegas to ask you a question? Have the other great geniuses of
>>our time taken this approach?
>[...]
>>At least Hardy brought Ramanujan to Cambridge himself rather than
>>smugly challenging him to find his own way there.
>
>Well, Doyle *did* refuse to sign SuperSystem, insisting Bill Gates come
>sit in the game to get the autograph...
>
>-k. who, at 27 weeks, feels more like the Mountain than Mohammed
gary_...@my-dejanews.com
D Taylor <ev...@home.com> wrote:
> Mason Malmuth wrote (by e-mail),
> His belief I owe 2+2 an apology for claiming they censored
> my post(s). ...
Gee Mason, do you owe me an apology for claiming you've never censored a post?
Alan Bostick
> If my earlier post wasn't censored, then it is possible that I
> neglected to have scrolled down to the transmit button. I wasn't used to
> posting to 2+2 and may have exited the edit session prematurely. This
> happened to my current reply to Mason, which I re-posted. It might be
> more intuitve of a GUI to have the message options all precede the
> message text window.
Never attribute to malice what is adequately explained by stupidity.
Alan Bostick
> I never cease being amazed at how an open, analytic approach
> to evaluating the starting hands in hold'em could have
> generated such heat in the two years since I first posted it
> on the internet.
Nothing amazing about it. With all the fuss you make about it, it's only
inevitable that sooner or later someone fusses back.
Accusations of censorship, for example, generate a lot more heat than
saying, "Oops, I forgot to hit the 'Post Message' button!"