shuffling in general

[Prism]

yesterday, i borrowed a ponza deck for fridaynight magic. the deck ran 24 sources of mana, being 20 mountains and 4 chrome moxen. what happened was that i was constantly drawing 1 land hands, and when i drew a 2 land hand, i never drew into the third until it was too late.

i took it back to the person who lent me the deck, and said that the deck sucks because it's so inconsistent, and he shuffled deck and drew seven. it was a good draw. he did it again. and drew well. then i tried. and a 1 land hand again. we both did this about five times, with same results.

this made me realise, that i've been getting manascrewed pretty often, but neglected the possibility that i might be shuffling the wrong way.

my shuffling method is to combine any number of cut-and-drop or riffle a number of times, usually between 9 and 15. the other guy just pileshuffled in 8 piles. when i showed him my shuffling method, he shook his head and said that all the lands will go together; he didn't know why.

all that said, i have a few questions that have been boggling my mind for ages:

1) does cut-and-drop shuffling achieve the same kind of shuffling effect as riffling?

2) in the thread about shuffling in type one, there were some people that said pileshuffling does not actually shuffle the deck, nor should you do it with the number of piles that divides the deck, or is a prime; why is this?

3) in a dci sanctioned tournament, would a single 8-pile shuffle be enough to 'randomise' a deck?

4) is it possible to overshuffle?

5) finally, if you have some personal shuffling tips, they would be appreciated.

thanks in advance everyone!

[Machinus]

Effective distribution is achieved only through mixing shuffling methods. My preferred method is several riffles followed by a prime pile shuffle - repeat this two or three times before every match, and once before every game.

[Tempe]

Technically, the chances are greater to get mana clumps if you excessively shuffle. A tip is to make sure all of your mana is split up in the deck by looking through it, and then riffle shuffling it a few times. You won't know the order of the deck, but the mana will be sufficiently spread out.

A note here is that if your land isn't clumped, pile shuffling may make it worse. The solution is to riffle when your mana isn't clumped, and pile shuffle when it is clumped.

--Tempe

[dmessiah]

Well damn everyone who shuffles one way when one thing is true and another when something else is true.  That aspect of Magic is supposed to be RANDOM.  Before every matchup, I riffle shuffle, and between each of these I pull out a chunk of cards from the center and alternate them between the front and the back of the deck.  I repeat this maybe a dozen times.  I think this is the only way to get a nice, random shuffle - through "micro movements" (a riffle shuffle will usually place adjacent cards near one another) as well as "macro movements" (large chop shuffles move cards that sit either at the top or bottom of your deck when you riffle to random spots in your deck).  If I get mana scewed, I shrug and take a mulligan.

     Also, as a point of order, if you take a deck and assign it a random distribution, you will get land clumps.  Statistically speaking, after a large number of shuffles, your lands will NOT become more evenly distributed.  If you have Magic Workstation, feel free to use the easiest example of this by shuffling and viewing your library a few times.

     Land clumps are part of the game.  Get used to it - it's what Brainstorm/Fetch is for.  Only pussies cheat and don't shuffle well.

[Lord_Drazinus]

The best way of shuffling I've found is riffling before every game 5 or 6 times, and pile shuffle maybe 2 or 3 times before a match.

[-Draz-]

[Azhrei]

With a poker deck, it takes about 7 riffle shuffles to randomize the deck adequately.

I find it's generally more important to shuffle your opponent's deck. A word to the wise: never, ever, EVER trust anyone who is pile shuffling. That's the easiest way to mana weave a deck that there is. If someone pile shuffles, pile shuffle them back into three piles-- if they mana-wove, they end up with major clumping.

Really, you absolutely must assume that your opponent is cheating. People mulligan against me ALL THE TIME.

[monstre]

With a poker deck, it takes about 7 riffle shuffles to randomize the deck adequately.

Looks like we have a winner  
It turns out that you have to riffle shuffle a poker deck seven times for a perfect shuffle. I don't know the details, but lots of maths were involved and this is a definitive answer. Any less than seven and the deck isn't completely random. Any more than that is pointless but not harmful, as there is no such thing as overshuffling. Since a typical magic deck is larger than a poker deck, maybe an 8th shuffle is required, but I have no idea where the threshold is in terms of deck size.

Every other method (pile shuffling, mana weaving, etc.) is really an attempt to decrease randomness in order to avoid the occasional "wild hand". Many players do this to some degree, so maybe they don't realize that mana clumps are to be expected every now and then. You might want to think about this before complaining that shuffling in apprentice is "broken" (a pet peeve of mine, sorry).

[Prism]

first of all, thanks to everyone that put in the effort to respond. they are greatly appreciated.

yes, manascrew/flood is part of the game. i accept that, and am usually cool about it.

however, as i said in my initial post, i was getting manascrewed so much that led me to investigate any possible answers to it. for your information, i played 7 rounds between yesterday and today, and mulliganed 6 times in total and then still lost some games due to manascrew, because i was not willing to go down to 5 sometimes (mountain/mox+red spellbomb hands). so it's not like i'm being stubborn with my hands that i should have thrown away.

also, i think i do excessively shuffle- between rounds, i'd just sit down and cut and drop shuffle my deck ad infinitum.

now i'm trying out different shuffling methods, it seems everybody has their own personal style about it... so i hope to find my own. ;D

thanks everyone, and if you have anything to add, feel free!

[Zharradan]

You cannot just rely on a single shuffling method... while riffling multiple times is good (and is something you should do), you should not rely on only riffling. Riffling by itself will leave cards that were near the top of the deck still near the top of the deck, for example.

[Dr. Sylvan]

You cannot just rely on a single shuffling method... while riffling multiple times is good (and is something you should do), you should not rely on only riffling. Riffling by itself will leave cards that were near the top of the deck still near the top of the deck, for example.

Only if you don't cut it at all. Especially with the large number of shuffles during a game, I am never too worried about the "perfect riffling doesn't move cards from one end to the other" problem.

[dmessiah]

You cannot just rely on a single shuffling method... while riffling multiple times is good (and is something you should do), you should not rely on only riffling. Riffling by itself will leave cards that were near the top of the deck still near the top of the deck, for example.

Only if you don't cut it at all. Especially with the large number of shuffles during a game, I am never too worried about the "perfect riffling doesn't move cards from one end to the other" problem.

You still have the problem of "my Lotus is a few cards away from my Yawgmoth's Will - somewhere in there" if you just cut.  I don't think a single cut is sufficient if the person has ONLY riffle shuffled.  You just move that cards that didn't move much relative to one another to somewhere ELSE in the deck.

[Dr. Sylvan]

You still have the problem of "my Lotus is a few cards away from my Yawgmoth's Will - somewhere in there" if you just cut. I don't think a single cut is sufficient if the person has ONLY riffle shuffled. You just move that cards that didn't move much relative to one another to somewhere ELSE in the deck.

The point of a riffle shuffle is that it does fix this problem my moving each card a minimum of several places from its original position, and moving most of the deck fifteen-plus places with each shuffle. If you really think you're not altering the position of the few uppermost and lowest cards, then make sure several of the cards from the middle on each shuffle end up on top of whatever was there before.

Riffle shuffles are brokenly good. I only pile shuffle when my deck is literally clumped almost perfectly into mana and nonmana, and in this case I will riffle multiple times after a four-pile shuffle.

[dmessiah]

You still have the problem of "my Lotus is a few cards away from my Yawgmoth's Will - somewhere in there" if you just cut. I don't think a single cut is sufficient if the person has ONLY riffle shuffled. You just move that cards that didn't move much relative to one another to somewhere ELSE in the deck.

The point of a riffle shuffle is that it does fix this problem my moving each card a minimum of several places from its original position, and moving most of the deck fifteen-plus places with each shuffle. If you really think you're not altering the position of the few uppermost and lowest cards, then make sure several of the cards from the middle on each shuffle end up on top of whatever was there before.

Riffle shuffles are brokenly good. I only pile shuffle when my deck is literally clumped almost perfectly into mana and nonmana, and in this case I will riffle multiple times after a four-pile shuffle.

     While this is true, I'd say it's more true for the middle cards than it is for the top and bottom ones.  I think a few chop shuffles between the riffles fixes this nicely.  The first from top and the third from top, say, won't move far from one another (if at all) in 4 riffle shuffles unless you are horrendously meticulous about it.

[Matt]

The cut "shuffle" is completely useless except to tie up the loose ends that riffle shuffling leaves (leaving cards that began near the top near the top, and same for bottom). If you want to see how well a particular shuffle works, try it face-up*. It literally takes about thirty cut shuffles to do the job of one riffle shuffle. It's the most worthless randomization method there is.

Also, the point of a pile shuffle is to ensure that any two cards that started out next to each other end up with at least one card between them. That's a surprisingly efficient mixing algorithm, and it's best when the number of piles is coprime with the number of cards in your deck - and the lowest such number is seven (technically 1 is coprime with 60, but that's a degenerate case). Eleven or thirteen piles would also work, but seven is easier to actually do, so I reccommend that number. I would never go without at least one pile shuffle for this reason.

If I want to not mulligan, I do this:

1. Riffle shuffle at least five times, more like eight or ten.
2. Pile shuffle once.
3. Repeat step one.

*It helps to do it with a standard poker deck that you've put in order, because it gives a definite order to the entire deck (because multiples of, say, FoW rapidly become indistinguishable). This lets you easily see just how "out of order" any given shuffling method leaves you, and you should quickly notice that riffle shuffling is the single best randomizer - but a combination of two disjoint methods like the one I outline above works even better.

[theorigamist]

A pile shuffle method I've found to be good is six piles; then take the leftmost pile and pile shuffle it among the remaining five; then take the leftmost pile and pile shuffle it onto the remaining four, etc. until you've only got two piles left.  Those two piles are each thirty cards, and that makes for a good point to start riffling.

Also, about the poker shuffling thing, you have to keep in mind that that's with a 52 card deck, which makes a ton of difference to the math involved.  For example, somebody said seven is good, and any more than that will not hurt the randomization.  But eight perfect shuffles with a deck of playing cards will return the deck to it's original position.  That's not true of 60 cards, but the fact that seven is the right number of riffle shuffles also might not be true of 60 cards.

[Azhrei]

Okay, a couple points I feel I have to make as someone who one day realized that after casting his 3000th Gaea's Blessing that if he knew where the Ancestral Recall was when he started shuffling, he knew where it was when he finished. I can track 4-5 cards nowadays if I try-- I can do it in decks someone else is shuffling too-- I once cut a friend to his Black Lotus 5 times in a row to demonstrate. I'm really, really out of practice but for a while I'd have friends put Academy, Timetwister, and 3 Moxen in a deck at random places and then I'd shuffle them all to the opening hand maybe 4 in 5 times. That was back in about 1999 I think.

1) You *absolutely* must assume that anyone pile shuffling is cheating. It's the easiest way to weave, and a 3 pile shuffle will punish them for this.

2) Seven riffle shuffles are the ONLY thing you can trust someone doing. If done with uneven sides, this will mix the cards up so that the top cards don't remain at the top, etc..

3) I recommend giving people a 3 pile shuffle regardless, because if they had a relatively even distribution of spells and mana it's a good way to force a mulligan. If their deck is 50% mana, make it a two pile shuffle. People mulligan against me ALL THE TIME.

4) Even if you don't cheat, you need to learn how people do it so you can watch for it. For example, way back in 1997 or so I was playing some kitchen table games with a good friend of mine, and he broke down and asked me how I always had the Abyss against him when he was cutting my deck. I had a good laugh and told him that he always cut my deck at the same spot. I didn't feel bad because the game before I'd found out that every time I cast Blessing and looked through my graveyard, when I wasn't paying attention he'd draw extra cards.  

After that, I started watching him shuffle and cut him to mana clumps.

5) Basically, the best advice I can give to you is that there are about a dozen or more good, legal tricks that an give you an advantage during the shuffling stage, especially ones that deal with screwing someone else's weave. There are also a lot of other things you can do that are impossible to NOT do once you're able to-- for example, if you CAN count cards playing blackjack, it's impossible to not count cards because your brain does it and you can't ignore that knowledge. You just have it.

The safest way to handle this is to always shuffle your opponent's deck, and always have them shuffle yours. The only problem with that is that many, many Magic players are revoltingly filthy and get grease on the sleeves.

No joke, a friend of mine was playing at a PTQ in High Point, NC, and someone shuffles his deck that was so dirty his sleeves actually STANK for the rest of the day, and he had to buy new ones.

[RoadTrippin]

The cut "shuffle" is completely useless except to tie up the loose ends that riffle shuffling leaves (leaving cards that began near the top near the top, and same for bottom). If you want to see how well a particular shuffle works, try it face-up*. It literally takes about thirty cut shuffles to do the job of one riffle shuffle. It's the most worthless randomization method there is.

The way I look at it, and I'm sure most others do as well, is that the cut obviously doesn't serve in place of a shuffle, but rather to avoid letting an opponent stack their deck. This works, because they don't know where you're going to cut. It is absolutely not meant to replace shuffling.

[Matt]

also, i think i do excessively shuffle- between rounds, i'd just sit down and cut and drop shuffle my deck ad infinitum.

Roadtrippin', there are people who do this and think it works. That part of my post was for them, not you who know how bad it is.

[Rainula]

Shuffling is an imperfect art.  The goal in shuffling is to completely randomize your deck.  Unfortunately, absolute randomization is not a feasible goal.  Another issue is that in magic, you are not rewarded by a good randomization, you are rewarded by having an even mix of lands, threats, and counters in your opening hand.  This problem in goals causes players to dislike landscrew and landfllood enough that they start to question their shuffling techniques.  Fact; if you are consistently getting 2-4 land hands almost every game (read over 95 %), then you are cheating (given an infinite set of data).  Additionally, if you are getting 1-land hands almost every game, then you are also cheating.  Basically, if you are shuffling to evenly disribute rather than randomize and you are doing that task well, then you are cheating.  I just thought someone needed to say something to that affect.  The easiest way to avoid issues like the two mentioned above is to shuffle a lot, do not pile shuffle as it is an ordered method and therefore does not provide any randomization unless you randomly place cards into different piles, and try not to shuffle in the same way all the time.

[yodoblec]

Additionally, if you are getting 1-land hands almost every game, then you are also cheating.

Why would you want only 1 land? Other than a combo deck that only needs 1 land is the only time you want to do that.

[Rainula]

Cheating does not have to be to your advantage.  Getting a one land hand is, unless you are playing two-land belcher, going to be a disadvantage.  My point was not whether or not it is good for your game or not, my point was that in getting a one land hand almost all the time means that you are not randomizing your deck, which is cheating.

[Machinus]

Azhrei:

Are you referring to remembering where cards are during riffle shuffles or pile shuffles...or do you mean you track cards through the entire shuffling process? This seems like it would be really difficult to do if you mixed shuffling methods. How do you do this?

[Jacob Orlove]

The safest way to handle this is to always shuffle your opponent's deck, and always have them shuffle yours. The only problem with that is that many, many Magic players are revoltingly filthy and get grease on the sleeves.

No joke, a friend of mine was playing at a PTQ in High Point, NC, and someone shuffles his deck that was so dirty his sleeves actually STANK for the rest of the day, and he had to buy new ones.

You can legally request a judge shuffle your deck in lieu of having your opponent do it. That seems like a case where such a request would be warranted.

[Godot]

You can legally request a judge shuffle your deck in lieu of having your opponent do it. That seems like a case where such a request would be warranted.

If my opponent requests that a judge shuffle his deck in lieu of me, can I request the manner in which the judge shuffles the deck--i.e. x-pile shuffle, riffle shuffle, etc.-- or is it purely up to the judges discretion how he shuffles the deck?

[Jacob Orlove]

Here's everything the universal tournament rules has to say on shuffling:

21.   Shuffling
Shuffling must be done so that the faces of the cards cannot be seen. Regardless of the method used to shuffle, players’ decks must be sufficiently randomized. Each time players shuffle their decks, they must present their decks to their opponents for additional shuffling and/or cutting. At a judge’s discretion, players may request to have a judge shuffle their cards rather than pass that duty to their opponents. By presenting their decks to their opponents, players are stating that their decks are correct, legal, and sufficiently randomized.

After decks are presented and accepted, any player who does not feel his or her opponent has made a reasonable effort to sufficiently randomize his or her deck must notify a judge. The head judge has final authority to determine whether a deck has been sufficiently randomized. The head judge also has the authority to determine if a player has used reasonable effort to randomize his or her deck. If the head judge feels that either the deck has not been sufficiently randomized or that a player has not made a reasonable effort to randomize his or her deck, the player will be subject to the appropriate provisions of the DCI Penalty Guidelines.

To aid in randomization, at REL 3 and higher events players must always shuffle their opponents’ decks at the beginning of games. The head judge can mandate the shuffling of opponents’ decks at lower RELs (1 and 2) as long as he or she announces this at the beginning of the tournament. If a shuffling effect takes place, players may shuffle and must cut their opponents’ decks after the shuffling effect is completed.

Once players have had the opportunity to shuffle and/or cut their opponents’ decks, the cards are returned to their original owners. If the opponent has shuffled the player’s deck, that player may make one final cut.

[Nukaramento]

A Land-filled or Land-light hand can't be prevented.  The real question is what kind of shuffling we must do when we mulligan.

[defector]

We are a community of the obsessive-compulsive(many with poor hygeine), this thread should be ported to a sociology classroom for further study.  The idea of randomization through an ideal method is a little ridiculous, the idea of stacking a deck through a concentrated effort is illegal.  Put your cards into piles if you want, when you get tired of that riffle shuffle them.  I pile shuffle them every time, in an obsessive-compulsive 6 to seven pile method simply because I don't want to damage my cards through a bad riffle shuffle.
defector

[johnstown713]

I found that I oile shuffle into 6 piles followed by a couple(usually 4) riffles.  I hardly have mana problems but never get the same cards in my hand.  This is how the pile goes:  
1   2   3
4   5   6

First I take pile number 1 and then take pile 5 and put it over it.  Then put pile 3 over pile 5.  Then put pile 6 over pile 3 and then 2 over 6.  Finally put 4 over 2 for maximum distibution.  I have had judges say it was perfectly legal as long as I did normal suffles after the pile shuffling.  I do have to mulligan sometimes but usually get a good distribution of cards.  And for any of you that are going to cheat my shuffling must be pathetic if im trying to cheat because I haven't won or top 8ed a tournament yet.

Johnstown713

[M]

Shuffling is an imperfect art.  The goal in shuffling is to completely randomize your deck.  Unfortunately, absolute randomization is not a feasible goal.  Another issue is that in magic, you are not rewarded by a good randomization, you are rewarded by having an even mix of lands, threats, and counters in your opening hand.  This problem in goals causes players to dislike landscrew and landfllood enough that they start to question their shuffling techniques.  Fact; if you are consistently getting 2-4 land hands almost every game (read over 95 %), then you are cheating (given an infinite set of data).  Additionally, if you are getting 1-land hands almost every game, then you are also cheating.  Basically, if you are shuffling to evenly disribute rather than randomize and you are doing that task well, then you are cheating.  I just thought someone needed to say something to that affect.  The easiest way to avoid issues like the two mentioned above is to shuffle a lot, do not pile shuffle as it is an ordered method and therefore does not provide any randomization unless you randomly place cards into different piles, and try not to shuffle in the same way all the time.

Thanks Rainula, for being correct in this jungle of confusion.
92% of the other posters should read it, and learn the very important difference between 'evenly distributing' and randomizing.

[Prism]

i have been deliberately trying to avoid a semantic/philosophic discussion, because there is a contradiction at the heart of the issue.

consider this:
player x uses any combination of shuffling method to sufficiently randomise his/her deck. he draws 7, it only has 1 land in it. he repeats the process only to draw a 1 land hand again. he does this 20 times, each time drawing a 1 land hand.

now, because randomisation can give any kind of hand, it is possible to say 'oh well, i randomised the deck enough each time, so i'm just having a bad day'.

however, that's not what i wish to discuss, if it's not evident from my first two posts. yes, manascrew sucks. but no, we still have to 'sufficiently randomise' our decks as the dci has set out. since dci does not describe what exactly sufficient randomisation is, i want to find out about maximising my chances to draw a 'good' hand, within dci policy. the issue of even distribution vs randomisation is rather irrelevant here. in fact, i feel that i should randomise as little as possible, as long as dci is fine with it.

before anyone calls me a cheater and points a finger at me, this is tied with sirlin's 'playing to win' philosophy: this is a game. nothing is cheap. nothing is honorable, just as nothing is dishonorable, as long as it's allowed in the game. just as manascrew is part of the game, so should maximising the chance to avoid manascrew be. (and i'm not talking about manaweaving/stacking of any kind here)

that said:

Unfortunately, absolute randomization is not a feasible goal.

what do you mean by absolute randomisation? what is a deck that has been absolutely randomised? one that has been shuffled 10 times? 50 times? 200 times? infinite? (i'm guessing infinite, which leads to the next question...)

Fact; if you are consistently getting 2-4 land hands almost every game (read over 95 %), then you are cheating (given an infinite set of data).

i'm not sure how you can draw out a percentage from an infinite set of data. i'd assume that in an infinite set of data, anything is possible.

edit: i feel that this issue needs more attention, because i rarely see it being discussed, if ever. i think shuffling well is vital in becoming a good player, and am really interested in what others have to say about it. so, please share your thoughts!