quote:
Originally posted by DarkMaster_MX:
I'm just saying the frequency in which it happens online vs when I play my local FNMs or tournies is very noticeable.
A poor RNG is just one conclusion you can draw from this observation. The other is that your experience with paper magic is actually skewed.
Most complaints I've heard about computer generated randomization is due to the fact that it doesn't behave the way that people expect it to, either compared to their expectations or their real life experiences.
Misguided expectations may be the result of poor statistical understanding. If you flip a coin 4 times, you can expect to get 2 heads on average, but getting only one head or getting 3 heads isn't a sign of rigged coins. There are 16 possible sequences of 4 coin flips, but only 6 (HHTT HTHT HTTH THHT THTH TTHH) result in exactly two heads. Within those 6, note the number with clusters of heads or tails. Clusters or clumps are another thing people often don't expect with randomization. On the other hand, there are 8 possible outcomes with either 3 tails or 3 heads. 1/8th of the time, you get the extreme cases where you get NO heads or NO tails. If we extend this out for 7 coin flips, you can see that you expect on average to get 3.5 coin heads, but obviously you can't get exactly 3.5 heads. There are 128 different sequences of 7 coin flips, but only 70 of them have exactly 3 heads or 4 heads. 42/128, or nearly 1/3 will have exactly 2 heads or 2 tails. 14/128 will have exactly 1 head or 1 tail and 2/128 have no heads or no tails.
If you play a large number of games, you can expect to run into a few of these highly improbably, but still statistically possible outcomes.
And it's human nature to ignore the 70 outcomes when the average happens and downplay the 42 outcomes when the slightly less than average happens and focus on the 16 outcomes where it's really frustrating.
The math for drawing your opening hand works out slightly different since the odds of drawing a land is probably closer to 40% and a non-land spell closer to 60%, but the same principle is still there.
Drawing from real life experiences can skew your view of the online world, as you may be shuffling in a manner that actually results in far fewer bad outcomes than should happen by chance than when you experience the more "pure" randomization, it won't mirror your real life experience. If your lucky quarter is 50 years old and one side has been rubbed so much that it is unrecognizable, your strategy of "Tails Never Fails" may pay off, resulting in tails 55+% of the time. When a friend flips his new coin, it may not be that his coin is flawed, but that yours has been rigged all along. I've seen people only do pile shuffles, or mana weave, or randomly distribute their lands at the end of a game by shoving them into the side of the deck, equal distance apart and these methods may result in mana screw/flood occurring less often than it should.
I know many people who refuse to use the spin-down dice at the beginning of a match because the numbers are very clearly ordered and it could be possible to get a non-random distribution of rolls if a person picks up the die after one person rolls, or if one person always orients the die to one side, say with the 20/logo on top.
Any game with random elements also leads to some very improbable, but still possible events. Any one who has played an online computer game with something akin to critical damage can probably tell of epic losses and wins when one person got their 1/20 ability two times in a row for massive damage and a win. That goes all the way back to D&D days when someone could roll two 20's in a row or two 1's in a row. That single event has a 1/20 * 1/20 chance of occurring, but if you play ten campaigns in one weekend or play a campaign every weekend for a year, you're bound to see some strange happenings.
Magic Online Trading League Bulletin Board
UBBFriend: Email This Page to Someone! 
posted September 13, 2010 11:52 AM
!).