mulligan issue

[Palbert6393]

This is a thread i posted on the PTCGO forums, but I wanted it to be seen here to hear any feedback from you guys as well (since it might be seen by different people here).

I apologize if this is a duplicate report, but I've never seen anything similar to this reported.

I noticed while playing that whenever I didn't get a basic pokemon in my hand while playing, the game appeared to string multiple starting hands together. So I decided to run a little test.

I built a 60-card deck which had 60 different cards, and only one basic pokemon (Regigigas EX). I then proceed to test this deck against the computer. I started 4 games. The first game, I had to reshuffle 5 times to find Regigigas-EX. The second, 7. The third, 4. The fourth, 29. Now by itself this information doesn't appear to be much. The important part is that for ALL 45 times I had to reshuffle my hand into my deck, I got one of the same cards in my next hand. I repeat ALL 45 times. 100% of the time.

Using my knowledge in Java programming, I wrote a program using the built in random function. This program is designed to get 7 different numbers between 1-60 (representing the cards in the starting hand.) It would then draw a new hand, compare it to the old hand and replace the old hand with the new hand. The program runs through 1 million shuffles, and then reports:

1) the percentage of the time that a card was in consecutive hands

2) the most times in a row that this happened

I ran this program ten times. The percentage ranged between 60.7% and 60.9%. While this was higher than I expected, it is still far away from the 100% that I got on PTCGO.

The longest streak of consecutive hands where any one card carried over from one to the next was 26. In 10000000 tries. On PTCGO I got a streak of 45 in 45 tries.

Why is this significant? As stated in the beginning of this point, this may be leading to a higher chance of having repeated starting hands without basics.

I know that the above info might not be enough to prove that something is wrong, but I do believe that this is enough info to warrent the possible issue being looked into.

I also don't know if this possible issue is only related to mulligans, or to any time the deck is shuffled, and I'm not really certain how to consistently check it in other situations.

I really appreciate anyone taking the time to read all this, and would be more than happy to post (if anyone is interested):

1) the java program i wrote

and/or

2) the list of all the cards in the hands from the tests on PTCGO

 

[ticka222]

ptcgo doesn't run on java.it runs on flash.about mulligans-random is random.it was explained on ptcgo a few times already how it works.can't remember in which forums though

 

[Palbert6393]

ptcgo doesn't run on java.it runs on flash.about mulligans-random is random.it was explained on ptcgo a few times already how it works.can't remember in which forums though

i know ptcgo doesn't run on java. My sample program that I created does, which i used to compare percentages and such. The point of using java random was so I could use a random function that I was sure worked.

 

[trainerjeff]

you have to also take into consideration that it is probability and you only have a 1 in 60 chance of pulling one card, thats not even a 2% chance per card. if you have 4 of the same card, which im assuming since you had multiples to get the same card in your hand you have essentially a 1 in 15 shot to pull the card so you can see why the difference can be made which is quite a difference in pulling a certain card.

i'm fairly positive the programmers at ptcgo are well equipped to handle something as simple as a random hand generator,and it still falls well within the probability of this occurring. how many durants do you see that don't at least have a few Mulligans (not talking real life play because deck manipulation takes away from this).

 

[Palbert6393]

you have to also take into consideration that it is probability and you only have a 1 in 60 chance of pulling one card, thats not even a 2% chance per card. if you have 4 of the same card, which im assuming since you had multiples to get the same card in your hand you have essentially a 1 in 15 shot to pull the card so you can see why the difference can be made which is quite a difference in pulling a certain card.

i'm fairly positive the programmers at ptcgo are well equipped to handle something as simple as a random hand generator,and it still falls well within the probability of this occurring. how many durants do you see that don't at least have a few Mulligans (not talking real life play because deck manipulation takes away from this).

I can't tell what your 1st paragraph is trying to say... anyway, to be clear the deck I created to test this on PTCGO had 60 unique cards, no duplicates.
Also, I understand where you are trying to go with the probability thing, but my sample program shuffled and redrew 10,000,000 times, and I the longest streak I got with that was barely 1/2 as long as 45 times on PTCGO. Doesn't that seem like it should at least be looked at?

 

[chrataxe]

I, for one, think what you've done is awesome, even if everyone else is being...uppity about it. So, let me start by saying, this has been a very well known issue since its release. In the beginning, the last card to be shuffled in on a mulligan was ALWAYS the first to be drawn when you drew your next hand...hence the reason I'm saying everyone is being uppity with all of their "I think they know how to program randomness" when its already been documented (WELL documented) they don't. Even now, its clear that drawing one of the same cards is inevitable on a mulligan and anytime you mulligan, you're almost always guaranteed to string a few of them together. I can count the number of times on one had where I've seen ONE mulligan and probably count the number of times I seen TWO mulligans on two hands.

There is DEFINITELY something wrong with their programming.

---------- Post added 03/05/2012 at 11:18 PM ----------

ptcgo doesn't run on java.it runs on flash.about mulligans-random is random.it was explained on ptcgo a few times already how it works.can't remember in which forums though

Read his post, then reply. Being a programmer, I'm sure he's more aware of the differences of flash and java than you and I. And, if you pay any attention at all, you'd know its been documented MANY times that there is a huge problem with hand randomization on mulligans.

you have to also take into consideration that it is probability and you only have a 1 in 60 chance of pulling one card, thats not even a 2% chance per card. if you have 4 of the same card, which im assuming since you had multiples to get the same card in your hand you have essentially a 1 in 15 shot to pull the card so you can see why the difference can be made which is quite a difference in pulling a certain card.

i'm fairly positive the programmers at ptcgo are well equipped to handle something as simple as a random hand generator,and it still falls well within the probability of this occurring. how many durants do you see that don't at least have a few Mulligans (not talking real life play because deck manipulation takes away from this).

As he's stated, he made a 60 card deck on PTCGO that had 60 DIFFERENT cards and only one of which was a basic, meaning, there was NO multiples. He then wrote a simple program to run the randomization on that numbered 1-60, meaning there was NO multiples. He was comparing apples to apples, you would know that if you read his post, that was the point of him posting.

 

[ChadtheDragonLord]

I've played somewhere in the neighborhood of 900 to 1000 expert games on PTCGO and I also believe there is a higher percentage chance of "clumping" with the online game, far more so than in real life, where I've also played hundreds of games. I very very very rarely get three of the same card (other than energy) in a hand in real life, but on PTCGO I have seen three Junipers, three Junk Arms, three of the same pokemon etc quite often. It all has to do with the randomizer.

Now, the other difference of course is the human factor. In real life, many players will look through their decks while shuffling and separate any clumps they find so the cards are more spread out. I am not sure if that is legal or not, maybe that is considered stacking your deck but everyone does it to some extent. So it is a manner of "sorted shuffling". That helps to avoid clumping. But on PTCGO it happens more often than it should I think. I haven't done any testing of this to gain proof, but I have gotten the sense of that just from playing so many games.