Sucking at Legacy: Rakdos in 1.5

morgan_coke · 07-03-2006, 10:05 PM

This thread is for the discussion of my latest article, Sucking at Legacy: Rakdos in 1.5. We would be grateful if you would let us know what you think, but please keep your comments on topic.

JayC · 07-04-2006, 12:28 AM

Dang... that was a pretty rough comment about standstill man

.

Zyrakris · 07-04-2006, 02:23 AM

You know what I've learned? Just don't test on MWS. It's foolish to do so, because the shuffler is crap to the point that it makes accurate testing impossible. Yeah, drawing 3 to 4 Lightning Helix doesn't happen that often. Neither does drawing 7 to 9 swamps in a row, with no mountains or other lands in between. If you doubt me, shuffle a couple times, then view your library and look at that steaming pile of BS. Check out all the hot chunky action in there.

Now that's not even starting on how TERRIBLE the players are on there... I just don't even bother anymore.

Akral · 07-04-2006, 03:08 AM

Oh, come on, Zyrakris. Do you understand anything in programming?
MWS shuffler is a simple yet effective randomization. A program by itself doesn't care if a card is a land, so it puts cards in totally random order, like you should have cards in RL. In reality, after shuffling, your deck should have land piles.
Moreover, you don't have to shuffle multiple times in mws, it doesn't change anything. Anything at all.

stanalquin · 07-04-2006, 03:59 AM

Thanks for the article, i would like to say that you could improve your actual writing in a lot of ways:
1) Contractions: it's not like you have to make like Data and never use them, but you have to be aware of two things: over-using them and occasions where the same contraction has two different meanings. In paragraph two you use "I'd" for "I had" and "I would": this is lazy writing.
2) Repetition: using a particular word repeatedly throughout an article is not that bad per se, but when that word is "suck" or "figure" (figure only really features in paragraph 2, but THREE times!) or any word for which there is a non-colloquial word, it again makes a writer look a bit lazy (or too cheap to afford a thesaurus!).
3) Sentence construction: "But" is not a good word with which to start a sentence: it isn't wrong as such, but you need to be a very skilful writer to use it correctly in this way. "They hurt, a lot." is not a sentence except in the sense that it starts with a capital and ends with a full-stop.
I suppose it comes down to attention to detail: there are a few other minor examples of this: using the / symbol when you can use the word 'or', use of apostrophes ("Dark Ritual had screwed up the decks' mana base"- singular deck so apostrophe before the 's'), inaccurate use of adjectives and adverbs "First, take a look at creatures." should, contextually, read "FirstLY....". Iron these problems out and your general writing will benefit as a result.

Chimaera · 07-04-2006, 04:03 AM

Quote:

Originally Posted by Akral

Oh, come on, Zyrakris. Do you understand anything in programming?
MWS shuffler is a simple yet effective randomization. A program by itself doesn't care if a card is a land, so it puts cards in totally random order, like you should have cards in RL. In reality, after shuffling, your deck should have land piles.
Moreover, you don't have to shuffle multiple times in mws, it doesn't change anything. Anything at all.

Like how in 250 i looked at the top 3 with a slyvan library, put a restricted card on top, then shuffled with a fetchland, then drew my next card next turn, the same restricted card at the top? ya right, thats random. Its happened more then once.

To the article though, I actually liked seeing the use of the word "sucks" alot, somewhat humerous, but maybe its cause i was tired and it was different for once.

Zyrakris · 07-04-2006, 06:10 AM

Quote:

Originally Posted by Akral

Oh, come on, Zyrakris. Do you understand anything in programming?
MWS shuffler is a simple yet effective randomization. A program by itself doesn't care if a card is a land, so it puts cards in totally random order, like you should have cards in RL. In reality, after shuffling, your deck should have land piles.
Moreover, you don't have to shuffle multiple times in mws, it doesn't change anything. Anything at all.

The MWS Shuffler has more than land clumps, which I can except, but it has card clumps. I'll give you an example.

Example decklist for T2 MBA:

Deck
Creatures - 4 Dark Confidant (RAV) 4 Ravenous Rats (9th) 4 Festering Goblin (9th) 4 Hypnotic Specter (9th) 4 Hand of Cruelty (SOK) 4 Plagued Rusalka (GP) 2 Ink-Eyes, Servant of Oni (BOK)	Spells- 4 Umezawa's Jitte (BOK) 4 Cruel Edict (9th) 4 Last Gasp (RAV) Land- 22 Swamp (1) (9th)

A bad deck, but that's not the point. I am now going to test this deck in "solo mode" and I'm going to shuffle three times (not that it matters) and reveal the decklist. For this first test, I won't be stating what set the card is from, and though you may question the relevance now, I'll get to that later.

Spoiler:

Clearly there is a huge chunk in the swamps in the first 20, one not easy to shrug off, at that. Festering Goblins are fairly close, though not suspiciusly so, as are all 3 Ravenous Rats, seeing as they find themselves within the first 20, each 5-7 cards apart. It gets more interesting though.

Spoiler:

This part is just golden, and if you can't see the horrid shuffler at work here, then you MUST be blind. 3 Last Gasps within this section, 2 of them together, 1 other only 4 cards away. All 4 Cruel Edicts show up in this one section, however, the first two only 4 cards apart, and the second two being only 2 cards apart. The 2 Ink-Eyes in the deck are only seperated by 1 cards, while 2 Confidants share a similar fate, also being seperated only by one card. Jittes are, as you see, next to eachother, while the only 2 swamps in this section are seperated by two cards.

Yeah, totally randomized here. Come on, it's like you skeletoned the deck, seperating lands and creatures and so forth, and then you jammed all the piles together and riffled a bit. It really doesn't seperate anything.

Spoiler:

Oh, look, the rest of my land! You can see the Hands are in a relative proximity of 4 cards, while 3 Specters reside in this section together, 2 of them walking hand-in-hand. Aside from that, just lots of sporadically clumped swamps.

I hope I won't have to go on any further, but in case I do, one fair solution to this problem (thus confirming that there is a problem to begin with) is that you have to take all cards that had been printed in another time (so Swamps, Hyppies, WoGs, whatever) and add in the cards from the other set. So you'll have Hyppie from Alpha, one from Beta, one from Revised, another from 9th. This REALLY helps stop the chunking, so long as you continue to do this with your lands (especially important) and other cards. This is, of course, more difficult according to the format you choose (hardest with Block, easiest with Vintage), and there will be cards that had only been printed once (Dark Confidant), so you can't use the trick with them.

Along with the fact that not many other players know this trick (and thus you'll be forced to face an opponent who is overconfident about his deck's ability, all because he can topdeck 3 Goblin Piledrivers), this is just one of the unfortunate things about MWS that has led me to stop testing with it except in cases of absolute necessity.

Now, I'm going to do the same experiment, and I'm going to use the trick:

Example decklist 2 for T2 MBA:
Creatures-
4 Hand of Cruelty (SOK)
4 Dark Confidant (RAV)
4 Plagued Rusalka (GP)
2 Ink-Eyes, Servant of Oni (BOK)
2 Festering Goblin (ON)
2 Festering Goblin (9th)
1 Hypnotic Specter (A)
1 Hypnotic Specter (B)
1 Hypnotic Specter (R)
1 Hypnotic Specter (9th)
1 Ravenous Rats (UD)
1 Ravenous Rats (IN)
1 Ravenous Rats (8th)
1 Ravenous Rats (9th)

Spells-
4 Last Gasp (RAV)
4 Umezawa's Jitte (BOK)
2 Cruel Edict (P2)
2 Cruel Edict (9th)

Land-
22 Swamp (Varied, trust me, I just don't want to list it)

Spoiler:

Nice land distribution here, and as you can see, some cards that have not or could not have been changed have still stuck together, like the Cruel Edicts. I won't bother commenting any further on the cards that the trick couldn't be done with.

Spoiler:

Not much to comment on here, the Last Gasps have chunked, and there's a clump of 3 swamps, which isn't entirely out of the norm, because it's not a clump of 7 following a clump of 5.

Spoiler:

Again, Jittes stick together, as do two of the inseperable Cruel Edicts, unfortunately. Also, land chunk is much, much better this time around. Anyway, I'm tired, and I hope that my arguments and evidence weren't a total waste.

bateleur · 07-04-2006, 08:15 AM

Writing software random number generators (or more correctly: pseudo-random number generators) is extremely easy. To seriously allege that a piece of software like MWS might get this wrong just seems ridiculous. It has undergone multiple revisions. Randomization is a critical factor for the operation of the software. Why on earth would the package's author get it wrong ?

I've used MWS quite a bit and never experienced this odd "chunking" Zyrakis describes. Also, the workaround he describes makes zero sense. If the shuffling algorithm is duff it can't suddenly become good just because the sixty (abstract) objects it's randomising are now all different.

People complain about the shuffler on MtGO too, which as far as I can tell has one key feature in common with MWS in this respect: it shuffles thoroughly, unlike some human players.

El Jugador · 07-04-2006, 12:06 PM

Quote:

Originally Posted by bateleur

Writing software random number generators (or more correctly: pseudo-random number generators) is extremely easy. To seriously allege that a piece of software like MWS might get this wrong just seems ridiculous. It has undergone multiple revisions. Randomization is a critical factor for the operation of the software. Why on earth would the package's author get it wrong ?

While it is easy to write a random number generator, it isn't trivial to write a correct permutation randomizer, so that every permutation has a probability 1/n!. We know MTGO's is right because Elf has said they used Knuth's algorithm from the Art of Computer Programming. We don't know what MWS uses and it could be some sort of incorrect, hacky mess.

ToshiUmezawa · 07-04-2006, 04:10 PM

However, testing on MWS has one big flaw still: everyone is testing. So you get about 40% teir one decks, 10% bad decks that used to be good and people won't admit suck (*cough*Landstill*cough*) and 50% rogue decks. Also, people don't want to seem "unorginal" so nobody plays Ichorid in Extended, unless they just like crushing people.

JayC · 07-04-2006, 07:01 PM

Is there any real evidence to support such claims? I mean... you're actually saying it doesn't have a correct code to produce pseudo-random shuffles and draws? Prove it.

Until you do, you simply remind me of little kids when they start to get so upset losing to the computer of a video game and actual cry "it's not fair! its not faiiiirrrr!"

Lanky · 07-04-2006, 09:33 PM

Ehhh, I have seen the same phenomenon. I can't really explain it, since random-ness is one of the easiest programmable functions. Unless you screw it up by getting too fancy..

bateleur · 07-05-2006, 01:45 AM

Quote:

Originally Posted by El Jugador

While it is easy to write a random number generator, it isn't trivial to write a correct permutation randomizer

Is there something wrong with putting a random card on top, a random card second, a random card third and so on until we run out of cards ?

Sounds trivial to me...

knightfall · 07-05-2006, 08:38 AM

Quote:

Originally Posted by bateleur

Is there something wrong with putting a random card on top, a random card second, a random card third and so on until we run out of cards ?

Sounds trivial to me...

I am no expert on RNG's so I will only speculate. First, they are not truly random. They require a seed (some people use time) to generate a number. The problem with that is if you were to stack a deck by selecting card after card, you are using the same seed over and over and hence will get poor results. Like I said, I only know the basics.

I do find, however, that people complain way too much about random shuffling. Some people do not understand that random distribution and even distribution are completely different.

As much as I would like to say that people blame shufflers because of insufficient shuffling in real life, I think it is something else. I think it is a psychological thing. Some people just never point blame at themselves. How many times has anybody ever heard how awesome the shuffler is because it won a game for them? Yet there are endless complaints about people getting screwed by them.

Top Deck · 07-05-2006, 10:35 AM

Quote:

Originally Posted by Zyrakris

You know what I've learned? Just don't test on MWS. It's foolish to do so, because the shuffler is crap to the point that it makes accurate testing impossible. Yeah, drawing 3 to 4 Lightning Helix doesn't happen that often. Neither does drawing 7 to 9 swamps in a row, with no mountains or other lands in between. If you doubt me, shuffle a couple times, then view your library and look at that steaming pile of BS. Check out all the hot chunky action in there.

Now that's not even starting on how TERRIBLE the players are on there... I just don't even bother anymore.

the irony is that yes the shuffler on both MWS and apprentice are accurate. it is accurate if the way we shuffle was totally random; however, if you color code your deck with different sleeves and then you shuffle you will see that when we shuffle it isnt random at all. most of the time we get the same things over and over simply because we shuffle a very fixed way.

07-03-2006, 10:05 PM	#1
morgan_coke Wizard Mentor Join Date: Mar 2005 Posts: 718	Sucking at Legacy: Rakdos in 1.5 This thread is for the discussion of my latest article, Sucking at Legacy: Rakdos in 1.5. We would be grateful if you would let us know what you think, but please keep your comments on topic.

07-04-2006, 12:28 AM	#2
JayC Banned (Ripping) Join Date: Feb 2006 Location: California, USA Posts: 3,759 Blog Entries: 2	Dang... that was a pretty rough comment about standstill man . __________________ This member has been found guilty of ripping a number of users on MTGSalvation. As a result he has been banned on this site and all his positive feedback has been removed. The MTGSalvation Staff wants to strongly encourage you NOT to trade with this member ever again in whatever situation might come up. Spread the word.

07-04-2006, 02:23 AM	#3
Zyrakris A Portrait of a Poison Join Date: Jul 2005 Location: With my eyes at the keyholes and ears to the walls. Posts: 2,984 Blog Entries: 4	You know what I've learned? Just don't test on MWS. It's foolish to do so, because the shuffler is crap to the point that it makes accurate testing impossible. Yeah, drawing 3 to 4 Lightning Helix doesn't happen that often. Neither does drawing 7 to 9 swamps in a row, with no mountains or other lands in between. If you doubt me, shuffle a couple times, then view your library and look at that steaming pile of BS. Check out all the hot chunky action in there. Now that's not even starting on how TERRIBLE the players are on there... I just don't even bother anymore. __________________ Getting the last word does not mean that you win the argument.

07-04-2006, 03:08 AM	#4
Akral Experienced Mage Join Date: Sep 2005 Location: Vilnius Posts: 52	Oh, come on, Zyrakris. Do you understand anything in programming? MWS shuffler is a simple yet effective randomization. A program by itself doesn't care if a card is a land, so it puts cards in totally random order, like you should have cards in RL. In reality, after shuffling, your deck should have land piles. Moreover, you don't have to shuffle multiple times in mws, it doesn't change anything. Anything at all. __________________ Judge level 1. Chloe - the true Noir...

07-04-2006, 03:59 AM	#5
stanalquin Banned Join Date: Sep 2005 Location: In the city which is twinned with Your Darkest Thoughts (no, really, it is!) Posts: 649	Thanks for the article, i would like to say that you could improve your actual writing in a lot of ways: 1) Contractions: it's not like you have to make like Data and never use them, but you have to be aware of two things: over-using them and occasions where the same contraction has two different meanings. In paragraph two you use "I'd" for "I had" and "I would": this is lazy writing. 2) Repetition: using a particular word repeatedly throughout an article is not that bad per se, but when that word is "suck" or "figure" (figure only really features in paragraph 2, but THREE times!) or any word for which there is a non-colloquial word, it again makes a writer look a bit lazy (or too cheap to afford a thesaurus!). 3) Sentence construction: "But" is not a good word with which to start a sentence: it isn't wrong as such, but you need to be a very skilful writer to use it correctly in this way. "They hurt, a lot." is not a sentence except in the sense that it starts with a capital and ends with a full-stop. I suppose it comes down to attention to detail: there are a few other minor examples of this: using the / symbol when you can use the word 'or', use of apostrophes ("Dark Ritual had screwed up the decks' mana base"- singular deck so apostrophe before the 's'), inaccurate use of adjectives and adverbs "First, take a look at creatures." should, contextually, read "FirstLY....". Iron these problems out and your general writing will benefit as a result. __________________ Q: How many Magic players does it take to screw in a lightbulb? A: 2. One to screw it in, and one to tell him he could have done it three turns quicker. Last edited by stanalquin : 07-04-2006 at 04:07 AM.


Home Articles Zendikar Spoiler (249/249) Radar Forums Blogs Wiki Writing/Contests Chat About

07-04-2006, 08:15 AM	#8
bateleur Random Scrub MTGS Writer Join Date: Jan 2005 Location: London Posts: 7,746	Writing software random number generators (or more correctly: pseudo-random number generators) is extremely easy. To seriously allege that a piece of software like MWS might get this wrong just seems ridiculous. It has undergone multiple revisions. Randomization is a critical factor for the operation of the software. Why on earth would the package's author get it wrong ? I've used MWS quite a bit and never experienced this odd "chunking" Zyrakis describes. Also, the workaround he describes makes zero sense. If the shuffling algorithm is duff it can't suddenly become good just because the sixty (abstract) objects it's randomising are now all different. People complain about the shuffler on MtGO too, which as far as I can tell has one key feature in common with MWS in this respect: it shuffles thoroughly, unlike some human players. __________________ -- Forum Awards: Best Writer 2005, Best Limited Strategist 2005-2008 5CB PotM - June 2005, November 2005, February 2006, April 2008, May 2008, Feb 2009 MTGSalvation Articles: 1-19 Also - Guest appearance on MTGCast #86!

07-04-2006, 04:10 PM	#10
ToshiUmezawa Experienced Mage Join Date: Dec 2005 Posts: 55	However, testing on MWS has one big flaw still: everyone is testing. So you get about 40% teir one decks, 10% bad decks that used to be good and people won't admit suck (coughLandstillcough) and 50% rogue decks. Also, people don't want to seem "unorginal" so nobody plays Ichorid in Extended, unless they just like crushing people.

07-04-2006, 07:01 PM	#11
JayC Banned (Ripping) Join Date: Feb 2006 Location: California, USA Posts: 3,759 Blog Entries: 2	Is there any real evidence to support such claims? I mean... you're actually saying it doesn't have a correct code to produce pseudo-random shuffles and draws? Prove it. Until you do, you simply remind me of little kids when they start to get so upset losing to the computer of a video game and actual cry "it's not fair! its not faiiiirrrr!" __________________ This member has been found guilty of ripping a number of users on MTGSalvation. As a result he has been banned on this site and all his positive feedback has been removed. The MTGSalvation Staff wants to strongly encourage you NOT to trade with this member ever again in whatever situation might come up. Spread the word.

07-04-2006, 09:33 PM	#12
Lanky Ascended Mage MTGS Writer Join Date: Aug 2005 Location: Ottawa, a campus bench therein. Posts: 379	Ehhh, I have seen the same phenomenon. I can't really explain it, since random-ness is one of the easiest programmable functions. Unless you screw it up by getting too fancy.. __________________

Thread Tools	Search this Thread
Show Printable Version Email this Page	User Name: Search this Thread: Advanced Search