Land drops in MTGA
Online Magic forum
Posted on July 16, 2020, 8:55 a.m. by spacecoyote1313
We all know land drops are key to any game of magic. However, over the last few months it seems that on MTGA the land count never seems to matter.
Generally you want to play with 22-24 land right? Time after time I will be playing a deck rocking 24 lands and I can't, for the life of me, get past 3 lands and I lose miserably. Then I will play an aggro deck running 20-22 lands and I will get flooded. None of this makes any sense and it is infuriating.
I understand, some games these things happen. But MTGA compared to paper magic, the land inconsistency is ridiculous. I have been playing paper magic for 8 years now and have never encountered as many land issues as I have on MTGA.
Am I crazy and quarantine getting to me or have other people also been experiencing their own issues with lands?
Seems random to me when I play. Are you playing a hundred games a week or just a few a day? Any clue how many games you played to get to this conclusion? Are you playing best of 1 or best of 3?
Also the previous poster is correct. In best of 1 games it secretly draws two hands every time and gives you the one that has the more even mana distribution.
July 16, 2020 11:45 a.m.
jonno.scott says... #4
Pervavita and Rzepkanut are right, in BO1 Arena draws two hands and gives you the one that best reflects your deck's land count - not necessarily an optimal split of spells and lands. For example if you run 15 lands then Arena will choose the hand with two lands if possible since you're running 25% lands and 0.25 * 7 = 1.75 lands. So knowing if you play BO1 or BO3 makes a difference.
July 16, 2020 12:55 p.m.
The MTGO coding is very bad about this type of thing because it's as close to true randomness in card order as is possible. What this means is that small sample sizes (anything less than 1000 matches) is going to be wildly inconsistent compared to paper magic. In paper, players often ensure the spells and lands are mixed nicely before shuffling and the land counts of opening hands and early draws is significantly more consistent in small sample sizes while just as consistent in large sample sizes.
You're not crazy, it's an issue about MTGO that's been discussed to death. I even sent in some of my own data to wizards on the topic using a 60 card deck with 24 lands and checking cycles of 100 games opening hand land counts. Statistically, you should get 1 land or 6 or more lands less than 1% of the time in such a scenario. Unfortunately in MTGO, of the 10 cycles of 100 games I submitted as data 6 of those cycles had this occurence happen at a 3-5% rate and 4 had it happen less than 1%. Over the 1000 total games the rate was fairly close to correct (though still 50% greater than the expected statistical average), but the 100 game sets were horrendously far from the mark. I submitted another 9000 games of opening hand data just to annoy them and see if things smoothed out after a truly significant sample size. They did not, even after 10,000 games of data there was still a 50% greater than expected rate. Wizards stated this was "working as intended" despite clear evidence to the contrary.
It would not surprise me at all if MTGA is similarly poorly coded in this respect.
July 16, 2020 1:30 p.m. Edited.
Flooremoji says... #6
jaymc1130: Is your argument players aren't shuffling enough for a true random state and often mix their spells and lands for a more favorable distribution, so when they change to MTGO the randomness throws them off?
The second part confused me.
July 17, 2020 3:42 a.m.
You could look at it like that I suppose. True randomness is just as impossible for a player to achieve in real life as it is for a computer program, so it's not really something that should be the goal for a person or a program. What players want from a game of MTG is a fair chance to compete in an enjoyable match. No one likes games where one player basically doesn't get to play purely due to chance and misfortune, there's no joy for the winner or the loser and this is an unsatisfactory result for all parties involved.
The better goal for either a program or a player is sufficient randomness that ensures knowledge of the card ordering is impossible but the overwhelming majority of opening hands will at least be playable even if they aren't ideal. This way the games played have a high likelihood of providing positive entertainment value for all parties.
July 17, 2020 8:30 a.m.
Flooremoji says... #8
Thanks for clairifing! I heard you needed to shuffle like... 40somthin times to achieve a true random state.
Pervavita says... #2
I can't quote it but I remember back about a year ago that there was an update where to reduce the mulligans Arena would draw two hands in secret and give you the hand that was the most optimal based off your land count for an opening hand between the two.
I know that doesn't address exactly what you are talking about but it is related and probably having some impact on what your experiencing.
July 16, 2020 11:08 a.m.