Shuffling less is often good enough.
Salem
"You may have heard that it takes about seven shuffles to mix up a deck
of cards to near randomness. Turns out, though, that most of the time,
perfect randomness is more than you need. In blackjack, for example, you
don't care about suits. The same mathematician who developed the
original result now says that for many games, four shuffles is enough.
And the result isn't only important for card sharks. It helps reveal the
math underlying Markov Chain Monte Carlo simulations, telling applied
mathematicians when they can stop their simulations."
would take a little adapting to vtes i guess.
--
salem
(replace 'hotmail' with 'gmail' to email)
"In *my* Assamite deck, this would pwn you in teh FAEC, so shut up."
"Thats only cos u've never sene mi Gionavvi PUNCHnMUCNH u asshat."
- James Coupe
Chris Berger
> http://science.slashdot.org/article.pl?sid=08/11/14/1438252
>
> "You may have heard that it takes about seven shuffles to mix up a deck
> of cards to near randomness. Turns out, though, that most of the time,
> perfect randomness is more than you need. In blackjack, for example, you
> don't care about suits. The same mathematician who developed the
> original result now says that for many games, four shuffles is enough.
> And the result isn't only important for card sharks. It helps reveal the
> math underlying Markov Chain Monte Carlo simulations, telling applied
> mathematicians when they can stop their simulations."
>
> would take a little adapting to vtes i guess.
>
I wonder how "good" the shuffles need to be. Because if I start with
a well-ordered deck, seven shuffles isn't even enough. I suppose it
depends on what you consider a shuffle. Seven riffle shuffles is only
going to move the top card down about 10 cards or less most of the
time, unless you mix in some overhand shuffles in there. 4 riffle
shuffles with an overhand shuffle in between is likely to be enough to
take a deck that is somewhat random already and randomize it, assuming
that the shuffles are neither "perfectly" executed (if each shuffle
puts exactly one card from each pile in between each two cards from
the other pile, it's not random, now is it?), nor too haphazardly
executed (20 from one pile, 20 from another, rest of the first pile
rest of the second pile... that's not going to randomize very much
either). But with a perfectly ordered deck, I feel like at least 10
shuffles are required before I no longer see obvious patterns from the
original ordering. Granted, that's not scientific, it's just a
"feeling", but I doubt 4 shuffles, no matter how well executed, can
really randomize an ordered deck, unless it's a table shuffle that
just spreads out all the cards and mixes them around.
Blooded Sand
I have been experimenting with something, but am unsure as to its
legality.
The idea is to take a deck, define the optimum layout of cards, then
number all cards from 1-90. This is done by inserting a paper slip
with the numbers on it into the face side of the sleeve the card is
in. I then shuffle, doing 3 riffles and 3 overhands. I find this gives
an incredibly good spread, allowing optimum card flow. The question
remains whether the original "Stacking" is legal. It could be argued
that it does not matter what order you start from, so whether it is
the same 1-90 order or randomly ordered, as long as there is
sufficient shuffling afterwards.
I would actually like a response from LSJ on this, as it is a highly
efficient system which I do not use in tourneys simply to not knowing
the legal status of it.
LSJ
> I have been experimenting with something, but am unsure as to its
> legality.
>
> The idea is to take a deck, define the optimum layout of cards, then
> number all cards from 1-90. This is done by inserting a paper slip
> with the numbers on it into the face side of the sleeve the card is
> in. I then shuffle, doing 3 riffles and 3 overhands. I find this gives
> an incredibly good spread, allowing optimum card flow. The question
> remains whether the original "Stacking" is legal. It could be argued
> that it does not matter what order you start from, so whether it is
> the same 1-90 order or randomly ordered, as long as there is
> sufficient shuffling afterwards.
>
> I would actually like a response from LSJ on this, as it is a highly
> efficient system which I do not use in tourneys simply to not knowing
> the legal status of it.
If the original order matters, then the subsequent shuffling is not sufficient.
Blooded Sand
The original orders matters only so far as it is the most even spread
of cards. I f you are saying that it is illegal because of the
original order, then for all sets of order of cards, there is no
suffficient subsequent shuffling. If the order is Random,
alphabetical, grouped into type or whatever form, it still has some
kind of initial order.
In exactly the same manner, if I then, before every game, spread my
deck out, evened the cards out manually, then shuffled, this would be
illegal. Which is a bit tricky as quite a few people already do this.
This is also the same as saying that using the 6/8 card shuffle, where
everything is divided into separate piles is illegal, as in this
layout the order is defined as being important, even though it is a
randomised order.
(Arguing with a 39 C fever is maybe not the wisest course in the world
though)
LSJ
> On Nov 15, 12:13 pm, LSJ <vtes...@white-wolf.com> wrote:
>> If the original order matters, then the subsequent shuffling is not sufficient.
>
> The original orders matters only so far as it is the most even spread
> of cards. I f you are saying that it is illegal because of the
> original order, then for all sets of order of cards, there is no
> suffficient subsequent shuffling. If the order is Random,
> alphabetical, grouped into type or whatever form, it still has some
> kind of initial order.
If the original order matters in any sort of intentional or predictable way to
the ultimate ordering of the deck after shuffling, then that shuffling is not
sufficient.
Blooded Sand
But by the same token, ANY initial order, whether preordained or
random, matters to the eventual order of the deck. If the shuffling
itself is not enough to guarantee a sufficient level of randomness,
then all shuffling becomes insufficient. It does not matter from what
state you start in if the idea of shuffling is to create a
randomisation of the set. However if the argument is that shuffling
does not introduce a high enough level of randomness, then any and all
initial ordering is suspect, no matter the cause of that specific
order.
By the same token, if the shuffling IS an agreed method of introducing
a sufficient level of randomness, then the inital order makes no
difference.
LSJ
> On Nov 15, 12:53 pm, LSJ <vtes...@white-wolf.com> wrote:
>> Blooded Sand wrote:
>>> On Nov 15, 12:13 pm, LSJ <vtes...@white-wolf.com> wrote:
>>>> If the original order matters, then the subsequent shuffling is not sufficient.
>>> The original orders matters only so far as it is the most even spread
>>> of cards. I f you are saying that it is illegal because of the
>>> original order, then for all sets of order of cards, there is no
>>> suffficient subsequent shuffling. If the order is Random,
>>> alphabetical, grouped into type or whatever form, it still has some
>>> kind of initial order.
>> If the original order matters in any sort of intentional or predictable way to
>> the ultimate ordering of the deck after shuffling, then that shuffling is not
>> sufficient.
>
> But by the same token, ANY initial order, whether preordained or
> random, matters to the eventual order of the deck.
True, but moot. We're discussing different methods of randomization, not
different methods of selecting a starting order. See "intentional or predictable".
> If the shuffling
> itself is not enough to guarantee a sufficient level of randomness,
> then all shuffling becomes insufficient.
Incorrect.
If the shuffling method in question (say, 4 riffles) is not sufficient, that in
no way implies that all shuffling methods (say, 90 card spread mix picked up in
the order indicated by 90 sequential tosses of a d100) are insufficient.
> It does not matter from what
> state you start in if the idea of shuffling is to create a
> randomisation of the set.
... if the randomization method is "sufficient", correct.
If the randomization method is insufficient, then, by definition, the starting
state matters.
> However if the argument is that shuffling
> does not introduce a high enough level of randomness, then any and all
> initial ordering is suspect, no matter the cause of that specific
> order.
OK.
> By the same token, if the shuffling IS an agreed method of introducing
> a sufficient level of randomness, then the inital order makes no
> difference.
That's what I said, yes.
James Coupe
s.com>, Blooded Sand <sand...@gmail.com> writes:
>The original orders matters only so far as it is the most even spread
>of cards.
If it's giving you a noticeably better flow of cards after the shuffle,
however, then the original order mattered. Similarly, Magic players who
try to stack a land every third or so card are trying to stack their
decks.
The theoretical perfect shuffle is unlikely to be possible. However, if
you're getting better results when you start from one order than if you
start from another, you're:
a) not shuffling enough
b) cheating (possibly unintentionally).
Some players have suggested when shuffling using a dice during a pile
shuffle. e.g. pile the deck into 1d6+3 piles, a few overhand shuffles,
some riffling etc., then do it again for 1d8+2 piles, and so on.
--
James Coupe
PGP Key: 0x5D623D5D YOU ARE IN ERROR.
EBD690ECD7A1FB457CA2 NO-ONE IS SCREAMING.
13D7E668C3695D623D5D THANK YOU FOR YOUR COOPERATION.
James Coupe
>random, matters to the eventual order of the deck. If the shuffling
>itself is not enough to guarantee a sufficient level of randomness,
>then all shuffling becomes insufficient.
No, that's not true.
You start with a deck that's got mostly action modifiers at the top and
mostly actions at the bottom. You riffle shuffle it once, and get a
fairly predictable flow of actions and action modifiers. Clearly, this
is not sufficient.
With the same deck, you riffle shuffle it, cut it somewhere at random,
riffle shuffle it again, pile shuffle it into 1d6+3 piles, a few
overhand shuffles, cut it again, riffle shuffle it again, ask another
player to cut it for you, pile shuffle it into 1d8+2 piles, overhand
shuffle, riffle shuffle again, get someone else to cut the deck in two,
riffle shuffle the two halves, pile shuffle each half into 1d4+3 piles,
stack each half back together, and then riffle shuffle the two halves
back into one. That's much more random. The order is far less
predictable. This has introduced several more random elements into the
process - dice, other people, some overhand shuffles.
The first single riffle shuffle is clearly insufficient. The second
arrangement is clearly much more random, and so may well be sufficient.
It doesn't attempt to produce a better flow based on the original
sorting.
A sufficiently randomized deck should, once in a blue moon, come up with
the deck in alphabetical order.
In tournaments past, there have been issues where a judge did a deck
check. They handed the deck back to the player in deck-list order.
Various players then wanted to sort the deck into a "more random" order
before shuffling, because starting from deck-list order would screw
them. Problem: if your shuffling is good enough, the original order
doesn't matter, the outcome will be exactly as unpredictable regardless
of where you start from.
Johannes Walch
> In tournaments past, there have been issues where a judge did a deck
> check. They handed the deck back to the player in deck-list order.
> Various players then wanted to sort the deck into a "more random" order
> before shuffling, because starting from deck-list order would screw
> them. Problem: if your shuffling is good enough, the original order
> doesn't matter, the outcome will be exactly as unpredictable regardless
> of where you start from.
Agreed. What I would like to see is a tournament rule that gives a minimum
requirement for "good enough" shuffling. Most people are not math
geniuses (including me) so that would be really helpful.
--
If playing against Cock all you need to
remember is: Don´t get caught by Cock.
Blooded Sand
So my original point was that I riffle shuffle 3 times, mixed with 3
overhand shuffles. Is this not sufficient enough to make it random?
because at a certain point, more shuffling will actually result in a
less random result. Strange, but true. All i want to do is start from
a sufficiently homogenized position, and then shuffle. Why is this any
different from taking your deck after play and making sure there are
no clumps?
James Coupe
s.com>, Blooded Sand <sand...@gmail.com> writes:
>So my original point was that I riffle shuffle 3 times, mixed with 3
>overhand shuffles. Is this not sufficient enough to make it random?
Apparently not, based on your own words:
"I find this gives an incredibly good spread, allowing optimum card
flow."
Since you are getting "incredibly good" results, your shuffling is
clearly insufficient.
> All i want to do is start from
>a sufficiently homogenized position, and then shuffle.
The question is why. If your shuffling is good enough, it doesn't
matter if you start from a homogenized position - it's just a waste of
time ordering the deck to start with.
If your shuffling isn't good enough, which it clearly isn't because your
stacking is having an effect on your result, then clearly this is the
wrong thing to be doing.
>Why is this any
>different from taking your deck after play and making sure there are
>no clumps?
People unclumping their decks are saying "My shuffling isn't good
enough." A good, thorough shuffle overrides unclumping. If it doesn't,
you're not shuffling enough.
If your shuffling is good enough, you pick up your deck whatever state
it's in - ordered, unclumped, clumped, ordered by artist surname - and
the result is equally unpredictable.
Clearly, there are some practical concerns - few shuffles are perfect.
However, you have already told us that the stacking you do is resulting
in a better outcome. Therefore:
a) your shuffling isn't good enough
b) you are cheating.
LSJ
> On Nov 15, 1:26 pm, LSJ <vtes...@white-wolf.com> wrote:
>> Blooded Sand wrote:
>>> By the same token, if the shuffling IS an agreed method of introducing
>>> a sufficient level of randomness, then the inital order makes no
>>> difference.
>> That's what I said, yes.
>
> So my original point was that I riffle shuffle 3 times, mixed with 3
> overhand shuffles. Is this not sufficient enough to make it random?
Probably not. Certainly not if you can afterwards perceive anything about the
initial sorting or grouping or homogenization of the deck.
> because at a certain point, more shuffling will actually result in a
> less random result. Strange, but true.
This false perception comes from a common misunderstanding of what "random" is.
Random doesn't mean that the various categories of cards in the deck will be
uniformly distributed through the deck. It means that each possible ordering of
the deck, including the ones in which each group of cards in the various
categories are clumped together, is equally likely.
> All i want to do is start from
> a sufficiently homogenized position, and then shuffle. Why is this any
> different from taking your deck after play and making sure there are
> no clumps?
It is no different. Both suggest improper shuffling.
Proper shuffling (sufficient randomization) means that the starting set up has
no predictable bearing on the final distribution of cards in the deck. Which
means that any effort spent adjusting the initial order or clumping or
homogeneity of the deck before shuffling is wasted.
Chris Berger
>
> If the original order matters, then the subsequent shuffling is not sufficient.
>
But the original order could determine the amount of shuffling
required to be sufficient, can it not? If the deck was already in
random order, then it will still be randomized after 1 shuffle. If
the deck is mostly random - say I played a game and got ousted on turn
6, then took the cards that I had played and put them back in the deck
- then only a few shuffles will re-randomize it. If I take those
cards I played during the game and put them in pseudo-random spots in
the deck instead of just stacking them on top, I believe that it would
take still fewer shuffles to sufficiently randomize.
So, I guess the question becomes this: Taking a deck that is grouped
with all like cards next to each other - if you pile-shuffle it, does
the entropy of the deck go up? I think yes. On the other hand,
nobody really likes "mana-weaving." It just wastes our time because
it takes so much longer than riffle shuffling.
LSJ
> On Nov 15, 5:13 am, LSJ <vtes...@white-wolf.com> wrote:
>> If the original order matters, then the subsequent shuffling is not sufficient.
>>
>
> But the original order could determine the amount of shuffling
> required to be sufficient, can it not?
No.
> If the deck was already in
> random order, then it will still be randomized after 1 shuffle.
Uh. By convention, the thing called "original deck" is not in random order.
Obviously, if the deck is randomized, then it only takes "no action" to
randomize it.
> If
> the deck is mostly random - say I played a game and got ousted on turn
> 6, then took the cards that I had played and put them back in the deck
> - then only a few shuffles will re-randomize it.
Probably not. But I suppose if you drew only your initial seven cards on those
six turns and handed the still-randomized-and-unseen remaining deck and the
seven cards to a judge to randomly insert into the deck, then it would take
fewer shuffles to randomize that mostly-random deck than it would a known-state
deck.
But all of this is just smoke. It doesn't help make the point clear and it may
actually distract from the point.
> If I take those
> cards I played during the game and put them in pseudo-random spots in
> the deck instead of just stacking them on top, I believe that it would
> take still fewer shuffles to sufficiently randomize.
Nah, since you know the positions of those insertions as well as the distance
between those various insertions.
> So, I guess the question becomes this: Taking a deck that is grouped
> with all like cards next to each other - if you pile-shuffle it, does
> the entropy of the deck go up? I think yes.
Yes. Indeed, taking a deck in any condition (other than pure random) and
shuffling it properly increases the entropy. (Pure random already being maximum
entropy, so shuffling it only maintains the entropy level rather than increasing
it).
> On the other hand,
> nobody really likes "mana-weaving."
Perhaps, but I don't know what that is.
James Coupe
<vte...@white-wolf.com> writes:
>Chris Berger wrote:
>> On the other hand,
>> nobody really likes "mana-weaving."
>
>Perhaps, but I don't know what that is.
Inserting land at every third card in a Magic deck. (Or any other
insertion strategy, relevant to a particular deck.)
There were some periods of time at which the DCI said this was legal,
then they said it wasn't, then there was fun with players who
intentionally pile-shuffled decks into 3 piles (to unweave the land)
before shuffling, and so on.
Chris Berger
> Chris Berger wrote:
>
> > If
> > the deck is mostly random - say I played a game and got ousted on turn
> > 6, then took the cards that I had played and put them back in the deck
> > - then only a few shuffles will re-randomize it.
>
> Probably not. But I suppose if you drew only your initial seven cards on those
> six turns and handed the still-randomized-and-unseen remaining deck and the
> seven cards to a judge to randomly insert into the deck, then it would take
> fewer shuffles to randomize that mostly-random deck than it would a known-state
> deck.
>
> But all of this is just smoke. It doesn't help make the point clear and it may
> actually distract from the point.
>
> > If I take those
> > cards I played during the game and put them in pseudo-random spots in
> > the deck instead of just stacking them on top, I believe that it would
> > take still fewer shuffles to sufficiently randomize.
>
> Nah, since you know the positions of those insertions as well as the distance
> between those various insertions.
>
But you don't know the exact position of the insertions any better
than if you overhand shuffled. Theoretically, when you overhand
shuffle, you "know" that the top, say, 15 cards are now the bottom 15,
the next 27 are now the next 27 from the bottom, etc... If you don't
count where you insert the cards into the deck, and assuming you
insert them face down without looking at them, it's very close to
random.
> > On the other hand,
> > nobody really likes "mana-weaving."
>
> Perhaps, but I don't know what that is.
>
I'm nearly certain that you do.
I suppose that pile-shuffling and mana-weaving aren't exactly the same
thing. Maybe I'm using the wrong terminology, but I take pile
shuffling to be laying the deck out into a pre-determined number of
piles (James's version using a die is better - I like that), in order,
one card per pile - I don't think it's technically a shuffle because
it isn't random. Mana weaving would be putting, say, a Master card
every 3rd card in your deck, putting one bleed and one stealth card
next to each other... arranging the deck in some sort of crazy way
before shuffling. I've always kinda considered pile-shuffling to be
mana weaving because, while you're not looking at it, you're still
spreading out the master cards in an even way, etc... I lump them
together moreso because of the amount of time it takes. I can't
believe that there's any faster way to randomize your deck than riffle
shuffling it, with overhand shuffles mixed in, until it's "ready".
Sitting at the table mana-weaving or pile-shuffling should be
considered a delay of game if it causes a tournament director to not
be able to start a round on time (or to cause a game to start after
the round has already started).
And I agree with what I think was your rule (I could be wrong?), that
stacking your deck is illegal because if it accomplishes anything, it
is cheating; and allowing it leads to an easy opportunity to cheat by
not sufficiently randomizing after stacking.
Kevin M.
> Salem <kella...@hotmail.com> wrote:
>> http://science.slashdot.org/article.pl?sid=08/11/14/1438252
>>
>> "You may have heard that it takes about seven shuffles to mix up a
>> deck of cards to near randomness. Turns out, though, that most of
>> the time, perfect randomness is more than you need. In blackjack,
>> for example, you don't care about suits. The same mathematician who
>> developed the original result now says that for many games, four
>> shuffles is enough. And the result isn't only important for card
>> sharks. It helps reveal the math underlying Markov Chain Monte Carlo
>> simulations, telling applied mathematicians when they can stop their
>> simulations."
>>
>> would take a little adapting to vtes i guess.
>
> I wonder how "good" the shuffles need to be.
The definitive article:
http://projecteuclid.org/DPubS/Repository/1.0/Disseminate?view=body&id=pdf_1&handle=euclid.aoap/1177005705
Kevin M., Prince of Las Vegas
"Know your enemy, and know yourself; in one-thousand battles
you shall never be in peril." -- Sun Tzu, *The Art of War*
"Contentment... Complacency... Catastrophe!" -- Joseph Chevalier
Please visit VTESville daily! http://vtesville.myminicity.com/
Las Vegas NAQ 2009 http://members.cox.net/vtesinlv/
ira...@gmail.com
> James Coupe schrieb:
> > them. Problem: if your shuffling is good enough, the original order
> > doesn't matter, the outcome will be exactly as unpredictable regardless
> > of where you start from.
>
> Agreed. What I would like to see is a tournament rule that gives a minimum
> requirement for "good enough" shuffling. Most people are not math
> geniuses (including me) so that would be really helpful.
For a 90-card deck, 10 reasonably good riffle shuffles should be
sufficient.
You can read about what a riffle shuffle is here:
http://en.wikipedia.org/wiki/Shuffling
Here is a good article going into the math of one way to tell if
you're shuffling well enough. It uses Magic: The Gathering as the
example, but the same principles apply:
http://www.starcitygames.com/php/news/article/4004.html
Using that math for VTES, if you have a 90 card deck that has 17
Master cards in it, you should expect to draw two master cards in a
row ~3.1 times in a sufficiently randomized deck. If you have 30
combat cards in a 90 card deck, you should expect to draw two combat
cards in a row ~9.8 times in a sufficiently randomized deck, and 3
combat cards in a row ~3.1 times.
So, if you want to check to see if your method of shuffling is
sufficiently random, one way to check is to put all your master cards
and combat cards on top, shuffle the way you usually do, and then
check to see if you have a distribution close to what I list above.
Of course, you'll need to do the experiment several times and take the
averages, since randomness naturally has some variation.
Ira
Blooded Sand
> Chris Berger <ark...@ugcs.caltech.edu> wrote:
> > Salem <kella...@hotmail.com> wrote:
> >>http://science.slashdot.org/article.pl?sid=08/11/14/1438252
>
> >> "You may have heard that it takes about seven shuffles to mix up a
> >> deck of cards to near randomness. Turns out, though, that most of
> >> the time, perfect randomness is more than you need. In blackjack,
> >> for example, you don't care about suits. The same mathematician who
> >> developed the original result now says that for many games, four
> >> shuffles is enough. And the result isn't only important for card
> >> sharks. It helps reveal the math underlying Markov Chain Monte Carlo
> >> simulations, telling applied mathematicians when they can stop their
> >> simulations."
>
> >> would take a little adapting to vtes i guess.
>
> > I wonder how "good" the shuffles need to be.
>
>
> Kevin M., Prince of Las Vegas
> "Know your enemy, and know yourself; in one-thousand battles
> you shall never be in peril." -- Sun Tzu, *The Art of War*
> "Contentment... Complacency... Catastrophe!" -- Joseph Chevalier
> Please visit VTESville daily!http://vtesville.myminicity.com/
>
> - Show quoted text -
Hehehehehe. Nice answer. So for a deck of cards where n=90, what is
the required number of riffle shuffles?
Chris Berger
> Chris Berger <ark...@ugcs.caltech.edu> wrote:
> > Salem <kella...@hotmail.com> wrote:
> >>http://science.slashdot.org/article.pl?sid=08/11/14/1438252
>
> >> "You may have heard that it takes about seven shuffles to mix up a
> >> deck of cards to near randomness. Turns out, though, that most of
> >> the time, perfect randomness is more than you need. In blackjack,
> >> for example, you don't care about suits. The same mathematician who
> >> developed the original result now says that for many games, four
> >> shuffles is enough. And the result isn't only important for card
> >> sharks. It helps reveal the math underlying Markov Chain Monte Carlo
> >> simulations, telling applied mathematicians when they can stop their
> >> simulations."
>
> >> would take a little adapting to vtes i guess.
>
> > I wonder how "good" the shuffles need to be.
>
>
Very interesting. Some of the math is beyond me, but I get the
general gist. I don't think they mentioned anything about cutting or
overhand-shuffling between riffles, which to me seems to be a given.
Otherwise, the top card of the deck is going to, on average, move down
about n/2 positions after shuffling n times, leaving the probability
that the original top few cards are still in the top few after
shuffling unnaturally high. If they mentioned it, it was in a section
that I glossed over due to notation I didn't grok.
Also, they don't specifically state a requirement for how "good" the
shuffles have to be. They just state a formula for how people *do*
shuffle (approximately, verified by experimentation). If I were able
to shuffle "perfectly", i.e. split into two equal piles and then
alternate one card from each pile without fail, that would screw
things up quite a bit. I'm sure that doesn't really happen in
practice though. As long as people shuffle fairly close to the
probabilities they give, then I presume the conclusions hold up.
witness1
> Very interesting. Some of the math is beyond me, but I get the
> general gist. I don't think they mentioned anything about cutting or
> overhand-shuffling between riffles, which to me seems to be a given.
> Otherwise, the top card of the deck is going to, on average, move down
> about n/2 positions after shuffling n times, leaving the probability
> that the original top few cards are still in the top few after
> shuffling unnaturally high. If they mentioned it, it was in a section
> that I glossed over due to notation I didn't grok.
I think the top card goes down closer to ~2^n positions, right? I
mean, assume it goes either 0 or 1 position down on the first shuffle.
Then (if it's still the top card) it goes down either 0 or 1 again,
but if it's now in position 1 it's likely to go down 2-3 more
positions (ending at position 3-4). And so on (ending at 0-1, 3-4,
8-9, or 11-12 on the third shuffle, which clearly doesn't average to
position 1.5).
-witness1
Orpheus
> sufficient.
>
> You can read about what a riffle shuffle is here:
> http://en.wikipedia.org/wiki/Shuffling
As a poker dealer, I do a lot of rifle shuffling, usually with 100% plastic
cards.
Let me tell you I'll never do it with collectible cards, not even in plastic
sleeves.
I want my cards to stay in good shape and this is definitely a way to hurt
them. A lot. And a good deck costs a lot more than a pack of bicycle cards,
doesn't it (as a matter of facts a deck of good non-plastic cards costs
about the price of a booster, so you can afford to change it once in a poker
night) ?!
--------
Orpheus
Chris Berger
Yeah, I'm totally wrong on that n/2 comment. So you're probably right
that it's closer to 2^n, which would mean after log2 k (k being number
of cards in the deck) shuffles it could be anywhere, right? That
changes things and more or less removes my one point of contention. I
still plan to overhand shuffle in between riffles, because it still
feels to me like the top of the deck isn't randomized enough
otherwise, but I can see that it's more likely just a feeling than
anything having to do with the mathematics.
witness1
I'd say that's pretty reasonable. Most people don't have really really
good shuffling, and my own breaks down at the top few cards
occasionally.
But if you did have really really good riffle shuffling (which is
generally what mathematicians concern themselves with), you probably
wouldn't need to do that.
Likewise, pile shuffling is a pretty bad mechanism for randomization
on its own, but it can separate cards that were stuck together and
allow your normal randomisation techniques to work better.
-witness1
Kevin M.
> ira...@gmail.com wrote:
> For a 90-card deck, 10 reasonably good riffle
> shuffles should be sufficient.
>>
>> You can read about what a riffle shuffle is here:
>> http://en.wikipedia.org/wiki/Shuffling
>
> As a poker dealer, I do a lot of rifle shuffling,
> usually with 100% plastic cards.
>
> Let me tell you I'll never do it with collectible cards,
> not even in plastic sleeves.
Especially since playing cards are single-cut printed cards, whereas CCG
cards are mostly dual-layered (i.e. a two-card 'sandwich').
> I want my cards to stay in good shape and this is definitely a way to
> hurt them. A lot. And a good deck costs a lot more than a pack of
> bicycle cards, doesn't it (as a matter of facts a deck of good
> non-plastic cards costs about the price of a booster, so you can
> afford to change it once in a poker night) ?!
I would disagree that riffle shuffling CCG cards hurts them "a lot", as I
have been riffle shuffling MtG and VTES cards for literally decades and
they are just beat around the corners, and very far from unplayable.
Perhaps my riffle shuffle is more gentle than normal.
Kevin M., Prince of Las Vegas
"Know your enemy, and know yourself; in one-thousand battles
you shall never be in peril." -- Sun Tzu, *The Art of War*
"Contentment... Complacency... Catastrophe!" -- Joseph Chevalier
Orpheus
> Orpheus <orphe...@NOSPAMfree.fr> wrote:
>> ira...@gmail.com wrote:
>> For a 90-card deck, 10 reasonably good riffle
>> shuffles should be sufficient.
>>>
>>> You can read about what a riffle shuffle is here:
>>> http://en.wikipedia.org/wiki/Shuffling
>>
>> As a poker dealer, I do a lot of rifle shuffling,
>> usually with 100% plastic cards.
>>
>> Let me tell you I'll never do it with collectible cards,
>> not even in plastic sleeves.
>
> Especially since playing cards are single-cut printed cards, whereas
> CCG cards are mostly dual-layered (i.e. a two-card 'sandwich').
>
>> I want my cards to stay in good shape and this is definitely a way to
>> hurt them. A lot. And a good deck costs a lot more than a pack of
>> bicycle cards, doesn't it (as a matter of facts a deck of good
>> non-plastic cards costs about the price of a booster, so you can
>> afford to change it once in a poker night) ?!
>
> I would disagree that riffle shuffling CCG cards hurts them "a lot",
> as I have been riffle shuffling MtG and VTES cards for literally
> decades and they are just beat around the corners, and very far from
> unplayable. Perhaps my riffle shuffle is more gentle than normal.
>
> Kevin M., Prince of Las Vegas
Perhaps, Kevin. But even the "bent around the corners" part doesn't bode
well with me...
Anyway, when you do have plastic sleeves, the lateral shuffle (don't know
the real english name) is efficient enough. Or you can just make several
rows and columns, rinse and repeat, and it works fine too.
Orpheus
--
"Cynicism isn't maturity. Callousness isn't strength. Pretending you
don't care so you don't have to try isn't "winning" What you do with
your life matters."
Luther Manning.
Kevin M.
> Kevin M. wrote:
>> Orpheus <orphe...@NOSPAMfree.fr> wrote:
>>> ira...@gmail.com wrote:
>>> For a 90-card deck, 10 reasonably good riffle
>>> shuffles should be sufficient.
>>>>
>>>> You can read about what a riffle shuffle is here:
>>>> http://en.wikipedia.org/wiki/Shuffling
>>>
>>> As a poker dealer, I do a lot of rifle shuffling,
>>> usually with 100% plastic cards.
>>>
>>> Let me tell you I'll never do it with collectible cards,
>>> not even in plastic sleeves.
>>
>> Especially since playing cards are single-cut printed cards, whereas
>> CCG cards are mostly dual-layered (i.e. a two-card 'sandwich').
>>
>>> I want my cards to stay in good shape and this is definitely a way
>>> to hurt them. A lot. And a good deck costs a lot more than a pack of
>>> bicycle cards, doesn't it (as a matter of facts a deck of good
>>> non-plastic cards costs about the price of a booster, so you can
>>> afford to change it once in a poker night) ?!
>>
>> I would disagree that riffle shuffling CCG cards hurts them "a lot",
>> as I have been riffle shuffling MtG and VTES cards for literally
>> decades and they are just beat around the corners, and very far from
>> unplayable. Perhaps my riffle shuffle is more gentle than normal.
>
> bode well with me...
In 15+ years of playing CCGs, I have never encountered "bent around the
corners" cards.
> Anyway, when you do have plastic sleeves, the lateral shuffle (don't
> know the real english name) is efficient enough. Or you can just make
> several rows and columns, rinse and repeat, and it works fine too.
I have found anything other than brand-new sleeves near-impossible to
shuffle and very user-unfriendly.
I'd humbly suggest that if you believe your VTES cards will be devalued by
shuffling them then you might be in need of some help. ;)
Kevin M., Prince of Las Vegas
jwjbw...@gmail.com
> So my original point was that I riffle shuffle 3 times, mixed
> with 3 overhand shuffles. Is this not sufficient enough to
> make it random? because at a certain point, more shuffling
> will actually result in a less random result. Strange, but true.
> All i want to do is start from a sufficiently homogenized
> position, and then shuffle.
Players generally prefer even distribution of cards. However, as LSJ
pointed out, a random result need not be the same as an even result.
It is actually surprising the extent to which clumping tends to occur
in random results. I have a vague recollection of players complaining
about the results obtained by the computerized shuffler on JOL. The
difference was, of course, that the JOL shuffler was truly random --
unlike their own usual efforts.
An illustration of this is the famous birthday problem. Put 23 people
together, and it is more than 50% likely that 2 or more will share the
same birthday. Put 57 people together, and it is more than 99%
likely, and there will probably be multiple pairings. An "even
distribution", would result in everyone having birthdays 6 or 7 days
apart, but such a result is astronomically unlikley.
> Why is this any different from
> taking your deck after play and making sure there are
> no clumps?
Deliberate unclumping is cheating, unless you shuffle sufficiently
afterwards to ensure normal chances of re-clumping.
I am sure a certain amount of unconscious "cheating" often occurs.
For instance, if a player is satisifed with the even distribution of
his cards in the first round of a tournament, he may decide that his
deck is already well "randomized" and give it only the most cursory of
shuffles going into the second round. If, however, he runs into
clumping in his first round, he is more likely to do a vigorous
shuffle before the second round.
jwjbw...@gmail.com
> This false perception comes from a common
> misunderstanding of what "random" is. Random doesn't
> mean that the various categories of cards in the deck will
> be uniformly distributed through the deck. It means that
> each possible ordering of the deck, including the ones in
> which each group of cards in the various
> categories are clumped together, is equally likely.
I suspect such misperceptions are less common among poker players.
When they are dealt a hand of 4 aces, they are less likely to curse a
bad shuffle, and more likely to sing the praises of Lady Luck.
Orpheus
>>> as I have been riffle shuffling MtG and VTES cards for literally
>>> decades and they are just beat around the corners, and very far from
>>> unplayable. Perhaps my riffle shuffle is more gentle than normal.
>>
>> Perhaps, Kevin. But even the "bent around the corners" part doesn't
>> bode well with me...
>
> In 15+ years of playing CCGs, I have never encountered "bent around
> the corners" cards.
"Beat" is the word you used. Thank you to point out that english is you
first language and not mine. That was useful.
>> Anyway, when you do have plastic sleeves, the lateral shuffle (don't
>> know the real english name) is efficient enough. Or you can just make
>> several rows and columns, rinse and repeat, and it works fine too.
>
> I have found anything other than brand-new sleeves near-impossible to
> shuffle and very user-unfriendly.
Depends on the sleeves, and how you shuffle them. Some are bad, yes. It's
also a business so you can buy new ones from time to time. Better than
ruining my cards anyday.
> I'd humbly suggest that if you believe your VTES cards will be
> devalued by shuffling them then you might be in need of some help. ;)
Such humility forces one's admiration.
So you don't care in which condition the cards you get are, and it has no
influence over the price you pay for them ? Good for you.
Cards can be ruined by too much bad shuffling, that's a fact. No help
needed, thanks.
Anyway I don't care that much about the "value", I just like my cards mint
or close enough. And preferably flat (riffle shuffles tend to bend the
cards).
--
Frederick Scott
news:gfsltu$ia1$1...@aioe.org...
> I have found anything other than brand-new sleeves near-impossible to
> shuffle and very user-unfriendly.
Cut cards a few times, pile shuffle, cut cards a few times, pile shuffle,
cut cards a few time. This creates total randomness, as verified by the fact
that I can easily get totally screwed by card clumping, insanely bad
distributions, and other forms of deranged ordering. Whether I started from
a perfectly sorted deck or a carefully weaved one, doesn't matter. It
works.
The main downside (assuming you're not trying to cheat, of course) is that
it takes a bit of time. Opponents and judges tend to get impatient. If
I have time between games, I try to do at least one of the pile shuffles
ahead of time.
Fred
jwjbw...@gmail.com
> On Nov 15, 11:32 am, LSJ <vtes...@white-wolf.com> wrote:
> > > If I take those
> > > cards I played during the game and put them in pseudo-random spots in
> > > the deck instead of just stacking them on top, I believe that it would
> > > take still fewer shuffles to sufficiently randomize.
>
> > Nah, since you know the positions of those insertions as well as the distance
> > between those various insertions.
>
> But you don't know the exact position of the insertions any better
> than if you overhand shuffled. Theoretically, when you overhand
> shuffle, you "know" that the top, say, 15 cards are now the bottom 15,
> the next 27 are now the next 27 from the bottom, etc... If you don't
> count where you insert the cards into the deck, and assuming you
> insert them face down without looking at them, it's very close to
> random.
Suppose the 7 cards you know about include a Fame and a Bum's Rush.
These are cards that work well together, and ideally you would like to
see them together. After inserting, you need to shuffle them enough
so there is the normal chance that this will occur. If, after
insterting, you KNOW that the Fame and the Bum's Rush are NOT side by
side (or that they ARE side by side) then the result is not random.
Still, I'm not sure its entirely wrong to say that the less you know
about the contents of a deck, the less shuffling you need do to
completely lose track of what you know. For instance, if you have
seen only one card, the only thing you need do is place it back in the
deck, and then cut the deck enough times so that you have no clue as
to where that card is.
witness1
> Still, I'm not sure its entirely wrong to say that the less you know
> about the contents of a deck, the less shuffling you need do to
> completely lose track of what you know. For instance, if you have
> seen only one card, the only thing you need do is place it back in the
> deck, and then cut the deck enough times so that you have no clue as
> to where that card is.
Eliminating your knowledge of the order is not the same as randomizing
the order.
-witness1
jwjbw...@gmail.com
In the ultimate scheme of things, there is no such thing as random.
No matter how much I shuffle, God will always know the order the cards
are in, and will know precisely how I put them in that order.
But it suffices if I find a method to make sure that every mortal
concerned (not just me) has no basis for knowing or guessing the order
of the cards.
Eliminating all relevant knowledge is the key. If I shuffle a "deck"
of only 3 cards, it will be impossible to shuffle successfully unless
I allow myself to lose track of which card is which.
witness1
This is not actually correct. If I sort my deck alphabetically, and
stick it in a box, and pull it out a month later, the mere fact that
I've forgotten that I sorted the deck does not mean it's sufficiently
randomized.
Knowledge of the deck's order is unrelated to whether the deck is
randomized. The key is that no matter what order it started in, the
randomization method could result in any of the n! possible orderings
with (approximately) equal probability.
-witness1
jwjbw...@gmail.com
In that case, your method has failed to achieve the stated goal. As
soon as you start drawing cards, you and everyone else at the table
will have a basis for guessing the order of the remaining cards. It
will be an observable fact that the cards drawn so far are in
alphabetical order.
> Knowledge of the deck's order is unrelated to whether the deck is
> randomized.
Lack of any basis for knowledge or guesswork is the sole basis for
"random".
> The key is that no matter what order it started in, the
> randomization method could result in any of the n! possible orderings
> with (approximately) equal probability.
Lack of knowledge is the basis for equal "probability". If the only
information I have about the seven horses, in a horserace, are their
names, then, from my point of view, each horse is equally likely to
win.