« first day (2997 days earlier)      last day (1911 days later) » 

5:42 PM
Huh, Angel of Grace: If you have less than 1 life and somehow haven't lost the game, damage dealt to you reduces your life total further below 0 (as normal).
 
Didn't think about this: Cindervines can be the target of its own activated ability. The ability won't resolve since it won't have a legal target, and no player will be dealt damage, but this does allow you to sacrifice Cindervines without another enchantment to target if you want to.
 
neat can save itself from exile effects
I am kind of confused why <0 is not also <1 seems kind of counter intuitive....
 
5:58 PM
what do you mean?
 
in RE to ruling on Angel of Grace
 
I think you have to think of it like "less than 1" is a range that you're either in or not. If damage would put you into the "less than 1" range, it puts you at 1 instead
 
ah, if you would enter the range of less than 1
if you are already in it you are not entering it
 
right
 
there was another card that this came up on recently.... but can't quite remeber
I think it had something to do with targets... or spellskite...
 
6:04 PM
Huh Incubation Druid can add colorless mana
 
That's how that kind of effect usually works
 
Well, it's usually phrased "any color" which excludes colorless
 
There are 3 of that effect that say "any type" and just one from 2002 that says "any color"
Generalizing to include lands other players own, it's about even
But any effect like that that was originally printed recently says "any type"
 
6:19 PM
"of any color": 204 cards. "of any type": 27 cards.
 
note that all the mana of any type cards also reference lands in some way
 
I was looking at "could produce"
Which finds every cards that generates mana based on what type lands could produce
There's a split between "any type" vs "any color" "that a land [someone controls] could produce", but the ones that say "any color" were all originally printed years ago
 
oooh, i see
 
 
2 hours later…
8:44 PM
Hmm, for the math people who care Arena Dev has spelled out a bit more on how the shuffler works
> Since people are asking, and it's no great secret. 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.
 
9:17 PM
Basically, Fisher-Yates is the standard shuffling algorithm, a Mersenne Twister is a very standard pseudo-random number generator, and I don't know what kind of real impact a cryptographically secure seed has.
But it's definitely positive if anything
 
9:32 PM
The only benefit of a crypto secure seed I can guess is that it will be harder for players to cheat? The Deck is only randomized whenever it is shuffled so theoretically if a player knew the seed and could "shuffle" their deck in an external program with the same generator and the same inputs they might be able to play with effectively a "face-up" library?
not a big deal for ladder, but online tournaments are becoming a bit more popular so there could be monetary incentive to cheating like that
That could be an interesting constructed "streamer event" though, players play with their library's revealed only to themselves
 
I'm sure the intent is something like that, but I don't know how much the source of the seed impacts whether it can be reconstructed from the generator output
I guess the fact that it's cryptographically secure just eliminates any possibility of some kind of long term pattern analysis to infer something about the seed
 
cost-to-benefit seems like it is a nobrainer to just use it
 
Arguably, depending on how important that aspect is, they should be using a cryptographically secure random number generator instead of a Mersenne Twister, but it probably doesn't matter considering that there are other layers of protection for examining the PRNG output in detail.
That may be slower, though
 
would be interesting though if someone does break it and finds a way to influence the shuffler. Turn evolving wilds into a vampiric Tutor
 
Even if someone does find a way to reconstruct the PRNG state early enough in a game to take advantage of it, that lets you figure out the result of the next shuffle, but that doesn't mean that you can decide what goes on top
In a lot of games, RNG manipulation is useful when the RNG is called frequently and for innocuous things, so you can take idle actions or wait for the next frame to get a different output
If the RNG is only used for shuffling decks and the opening coin flip, there isn't a good opportunity to manipulate it to a significant extent
 
9:51 PM
with evolving wilds type cards the player has the ability to choose when they randomize their deck, so it would depend on what other inputs the randomizer takes before spitting out a number (ie. Turn/current game time/how much turn time the player has left/ global time etc.)
not going to be too useful if everything is out of the players control (say like the specfic time the packet arrives at the server or something)
 
10:12 PM
@Malco Based on the original quote, there are two reasonable possibilities here, neither of which can be meaningfully manipulated
 
that is good to hear
 
One is that each time the deck is shuffled, a new random number generator is initialized using a new cryptographically secure random number. In that case, no manipulation is even theoretically possible because you never have any information about what will happen
The other is that once per game (or even once per server) a single PRNG is initialized using a cryptographically secure random number, and it is used for every shuffle. In that case, it may be possible to eventually infer the state of the PRNG, but that doesn't mean that you can meaningfully manipulate it, because it will only change when you actually shuffle
In either case, information like the turn and current game time wouldn't be inputs, because we already know what the input is: 256 cryptographically secure randomized bits
 
ahh ok, I don't deal with crypto and RNG on this level often (or at all really) so a lot of the specifics of how they exactly work are a bit beyond me
 
10:40 PM
@Malco So, the basic idea here with a PRNG such as a Mersenne Twister is that you input one number, and you get out a sequence of numbers, one at a time. And if done right, the numbers you get out are statistically uncorrelated, basically as good as actual random
When you talked about using information about the current game as an input, that's one method of constructing that one number. But they described a different method, using "cryptographically secure random bits", which basically means that that data comes from a (probably slow) true random source.
 

« first day (2997 days earlier)      last day (1911 days later) »