21 Land deck = only Land draws, 20 Land deck = never get Land

[Brave Pixel]

what's the deal with the coding of this game

[DontMisunderstand]

Publicado originalmente por Kurt Angle's Neck:

Publicado originalmente por DontMisunderstand:

I would argue that the existence of a hand smoothing algorithm is in itself a way that the shuffler is "poorly coded", as the rules of the actual game demand proper randomization, which the hand smoother would not qualify as. But that's an entirely different topic.

The hand-smoothing function shouldn't affect the shuffler at all. All it does is generate 3 different random shuffles of your deck then looks at the opening draw of each and gives you whichever one has the closest land/non-land ratio to the ratio of your deck. It makes your opening hand not truly random, sure, but that doesn't mean that the shuffler itself isn't sufficiently random.

3? I thought it was 2. And even 2 has an extremely harsh effect on optimal deckbuilding.

[Nec]

There is a lot of people here who would put a man in prison just because someone said they committed a crime. Scary

[Ripshot Blue]

Publicado originalmente por Kurt Angle's Neck:

Publicado originalmente por DontMisunderstand:

I would argue that the existence of a hand smoothing algorithm is in itself a way that the shuffler is "poorly coded", as the rules of the actual game demand proper randomization, which the hand smoother would not qualify as. But that's an entirely different topic.

The hand-smoothing function shouldn't affect the shuffler at all. All it does is generate 3 different random shuffles of your deck then looks at the opening draw of each and gives you whichever one has the closest land/non-land ratio to the ratio of your deck. It makes your opening hand not truly random, sure, but that doesn't mean that the shuffler itself isn't sufficiently random.

I can safely say I can count the amount of times I drew 0 land hands on just 1 hand. I wont say its rigged but I will say I would rather just true random like paper magic then any systems they put to force feed cards.

[Kurt Angle's Neck]

Publicado originalmente por DontMisunderstand:

3? I thought it was 2. And even 2 has an extremely harsh effect on optimal deckbuilding.

Pretty sure there was a dev somewhere that said it started as 2 then they went to 3. I could be imagining that though, I don't wanna go searching for it to back myself up.
Either way, doesn't matter if it's 2 or 3, it shouldn't affect your deckbuilding at all. All three copies of your deck are getting shuffled randomly just as they would without the smoothing operation. It's basically like getting two free mulligans, then getting the option to go back and pick one of the first two draws over the third (except it's all done automatically, behind the scenes).

[Kurt Angle's Neck]

Publicado originalmente por Ripshot Blue:

I can safely say I can count the amount of times I drew 0 land hands on just 1 hand. I wont say its rigged but I will say I would rather just true random like paper magic then any systems they put to force feed cards.

In order for that to happen, all of the generated shuffles would need to contain no lands. If the smoothing operation generates 3 separate shuffles (which to my understanding it does), and we assume you're using a standard 60 card deck with 24 lands, you would have a 0.0000013% chance of that happening. That's actually pretty incredible if you have to count how many times that's happened to you, I don't think I've ever had it happen.

Also just to throw out a comparison, if there was no hand smoothing operation there would be a 0.24% chance of drawing zero lands in your opening hand, that's a pretty significant difference.

[Ripshot Blue]

Publicado originalmente por Kurt Angle's Neck:

Publicado originalmente por Ripshot Blue:

I can safely say I can count the amount of times I drew 0 land hands on just 1 hand. I wont say its rigged but I will say I would rather just true random like paper magic then any systems they put to force feed cards.

In order for that to happen, all of the generated shuffles would need to contain no lands. If the smoothing operation generates 3 separate shuffles (which to my understanding it does), and we assume you're using a standard 60 card deck with 24 lands, you would have a 0.0000013% chance of that happening. That's actually pretty incredible if you have to count how many times that's happened to you, I don't think I've ever had it happen.

Also just to throw out a comparison, if there was no hand smoothing operation there would be a 0.24% chance of drawing zero lands in your opening hand, that's a pretty significant difference.

Personally I play brawl so the land draw odds are even worse but even so having no system would still be better to me as thats how paper magic is. You will draw hands with no lands, you will draw hands with all lands, while the odds are very low, that seems to never happen on arena unlike paper.

[DontMisunderstand]

Publicado originalmente por Kurt Angle's Neck:

Publicado originalmente por DontMisunderstand:

3? I thought it was 2. And even 2 has an extremely harsh effect on optimal deckbuilding.

Pretty sure there was a dev somewhere that said it started as 2 then they went to 3. I could be imagining that though, I don't wanna go searching for it to back myself up.
Either way, doesn't matter if it's 2 or 3, it shouldn't affect your deckbuilding at all. All three copies of your deck are getting shuffled randomly just as they would without the smoothing operation. It's basically like getting two free mulligans, then getting the option to go back and pick one of the first two draws over the third (except it's all done automatically, behind the scenes).

Deckbuilding starts first and foremost with probabilities of land draw in the starting hand. Of course the hand smoother has a major impact on that. With the general rule of thumb 24 lands in a 60 card deck, the baseline is a 30.87% chance of 3 lands. With the hand smoother taking 2 hands and choosing the more likely land:nonland ratio, that goes up to 52.21%. If it takes 3, that goes up to 67%. And of course, the math gets even more absurd if we're counting 2 and 4 land hands as acceptable as well... starting at 77.45% to get an acceptable start without the smoother, up to 94.02% with 2 hands, and 98.85% with 3. The absurd levels of consistency that come from the smoother inherently change the optimal way to build a manabase. Going from a 77.45% chance to get a usable starting hand with 24 lands all the way up to a 98.5% chance with 22 lands is absurd. That's 2 entire card slots opened up by the shuffler alone, while also making the deck entire orders of magnitude more consistent on top of that.

For the record, though... just going over the math makes it seem very much like the hand smoothing algorithm isn't a thing in the current iteration of the game. Or at least that the version currently implemented doesn't work the way we were told. During the course of 20 games today alone my 24 land 60 card deck had to mulligan to 4 on 4 separate occasions due to not having 3 or 4 lands in any of the hands. If the hand smoother does select 3 separate hands for the initial starter, the odds of that happening at all is 1.47%. Meaning the odds of it occurring once in 20 games would be 25.63%... meaning the odds of it happening 4 times would be a mere 0.43%. And today's experience wasn't even an outlier. It was the norm. So... going over the same math to calculate the odds of it happening purely randomly with no hand smoother, we'd arrive at at 71% chance of occurring once in 20 games, and thus a 25.41% chance of seeing it 4 times in a 20 game stretch. Sure, this is all anecdotal evidence, but the math, to me, indicates the best of 1 doesn't feature a hand smoother at all, given the frequency of events that should be extremely improbable with the hand smoother as described.

[TheStoryteller01]

Publicado originalmente por Brave Pixel:

21 Land deck = only Land draws, 20 Land deck = never get Land

Watch a RL tourney at your local gaming store and you'll see players complaing about the same thing.

Chance is a fickle mistress.

[Ripshot Blue]

Publicado originalmente por DontMisunderstand:

Publicado originalmente por Kurt Angle's Neck:

Pretty sure there was a dev somewhere that said it started as 2 then they went to 3. I could be imagining that though, I don't wanna go searching for it to back myself up.
Either way, doesn't matter if it's 2 or 3, it shouldn't affect your deckbuilding at all. All three copies of your deck are getting shuffled randomly just as they would without the smoothing operation. It's basically like getting two free mulligans, then getting the option to go back and pick one of the first two draws over the third (except it's all done automatically, behind the scenes).

Deckbuilding starts first and foremost with probabilities of land draw in the starting hand. Of course the hand smoother has a major impact on that. With the general rule of thumb 24 lands in a 60 card deck, the baseline is a 30.87% chance of 3 lands. With the hand smoother taking 2 hands and choosing the more likely land:nonland ratio, that goes up to 52.21%. If it takes 3, that goes up to 67%. And of course, the math gets even more absurd if we're counting 2 and 4 land hands as acceptable as well... starting at 77.45% to get an acceptable start without the smoother, up to 94.02% with 2 hands, and 98.85% with 3. The absurd levels of consistency that come from the smoother inherently change the optimal way to build a manabase. Going from a 77.45% chance to get a usable starting hand with 24 lands all the way up to a 98.5% chance with 22 lands is absurd. That's 2 entire card slots opened up by the shuffler alone, while also making the deck entire orders of magnitude more consistent on top of that.

For the record, though... just going over the math makes it seem very much like the hand smoothing algorithm isn't a thing in the current iteration of the game. Or at least that the version currently implemented doesn't work the way we were told. During the course of 20 games today alone my 24 land 60 card deck had to mulligan to 4 on 4 separate occasions due to not having 3 or 4 lands in any of the hands. If the hand smoother does select 3 separate hands for the initial starter, the odds of that happening at all is 1.47%. Meaning the odds of it occurring once in 20 games would be 25.63%... meaning the odds of it happening 4 times would be a mere 0.43%. And today's experience wasn't even an outlier. It was the norm. So... going over the same math to calculate the odds of it happening purely randomly with no hand smoother, we'd arrive at at 71% chance of occurring once in 20 games, and thus a 25.41% chance of seeing it 4 times in a 20 game stretch. Sure, this is all anecdotal evidence, but the math, to me, indicates the best of 1 doesn't feature a hand smoother at all, given the frequency of events that should be extremely improbable with the hand smoother as described.

Pretty sure theres some details they never said as the shuffler helps with 1 - 2 land hands but 3+ is just as rare as 0 land without having a 1/3 land ratio it seems

[Soji]

This is not a coding thing, this is a MTG thing. Lands are a bad mechanic and there's a reason most card games don't put your resource you need to be able to even play as something you also have to draw from your deck. Take it or leave it really, I take breaks from MTG all the time because of it.

[DontMisunderstand]

Publicado originalmente por Soji:

This is not a coding thing, this is a MTG thing. Lands are a bad mechanic and there's a reason most card games don't put your resource you need to be able to even play as something you also have to draw from your deck. Take it or leave it really, I take breaks from MTG all the time because of it.

... it's a good mechanic. There's a reason MtG makes up significantly more than half of the entire tcg market.

[Gryfn]

Publicado originalmente por Kurt Angle's Neck:

Publicado originalmente por DontMisunderstand:

I would argue that the existence of a hand smoothing algorithm is in itself a way that the shuffler is "poorly coded", as the rules of the actual game demand proper randomization, which the hand smoother would not qualify as. But that's an entirely different topic.

The hand-smoothing function shouldn't affect the shuffler at all. All it does is generate 3 different random shuffles of your deck then looks at the opening draw of each and gives you whichever one has the closest land/non-land ratio to the ratio of your deck. It makes your opening hand not truly random, sure, but that doesn't mean that the shuffler itself isn't sufficiently random.

It's funny how so many people confuse the bo1 hand smoothing algorithm with the shuffler when they are not the same thing. The bo1 hand smoothing algorithm is basically a free, automatic mulligan system for both players. And yes, it was updated to look at 3 hands and is also used when players manually mulligan.
"Arena uses “hand smoothing” in best-of-one games: the game shuffles the deck thrice and draws three hands and leans to choose the one with the land/spell ratio closest to the deck’s ratio (no color checks). Wizards tested other land consistency tech but confirmed to use only this one."

The bo1 hand smoothing algorithm is NOT the shuffler. Decks are still randomized.
"To shuffle decks in MTG Arena we use Fisher-Yates, pulling numbers from a Merseene Twister (MT199937), which is seeded with 256 cryptographically secure randomized bits. We use the same approach for coin tosses, only we’re looking for a 1 or a 2 rather than a whole deck of cards."
https://mtgazone.com/inner-workings-of-arena/

[Soji]

Publicado originalmente por DontMisunderstand:

Publicado originalmente por Soji:

This is not a coding thing, this is a MTG thing. Lands are a bad mechanic and there's a reason most card games don't put your resource you need to be able to even play as something you also have to draw from your deck. Take it or leave it really, I take breaks from MTG all the time because of it.

... it's a good mechanic. There's a reason MtG makes up significantly more than half of the entire tcg market.

nobody in the world would say the land mechanic is the reason MTG makes up a large portion of the market. Are you reading what you are typing right now?

[Ripshot Blue]

Publicado originalmente por Soji:

Publicado originalmente por DontMisunderstand:

... it's a good mechanic. There's a reason MtG makes up significantly more than half of the entire tcg market.

nobody in the world would say the land mechanic is the reason MTG makes up a large portion of the market. Are you reading what you are typing right now?

I wont say its a reason why its strong in the market but I will say its alot easier to learn then yugioh or pokemon & the land system has leverage to that.

[DontMisunderstand]

Publicado originalmente por Soji:

Publicado originalmente por DontMisunderstand:

... it's a good mechanic. There's a reason MtG makes up significantly more than half of the entire tcg market.

nobody in the world would say the land mechanic is the reason MTG makes up a large portion of the market. Are you reading what you are typing right now?

Somebody who had actually analyzed what works about the game would. It's the primary difference between Yugioh and MtG right now, and most of the other major competitors literally copied MtG's resource mechanics.

Hyper-consistent strategies are bad for ANY game. The land mechanic is an elegant solution to exactly that problem.