1. I like the flavour. It's like a fixed Kamigawa: asian flavour, but without the all-legendary and (hopefully) without every single card having a hard-to-remember name.
2. I like multi-color sets. It lends to interesting new card with new ability combos.
3. Specifically, I prefer the wedge color combo. I like RUG, BUG and BGW and RUW. Only RWB leaves me cold.
4. The spoiled card strikes a good balance.
5. The arts hint at interesting things, like the ninja-like acrobatics, the black wedge not being too overly ultra-black (i.e. demons / vampire / zombie), much like in Kamigawa where the rats angle gave black an interesting feel instead of the same rehash.
NFLed: Barring certain rares, Morph has never been 'overly powerful'. (That is, bomb with Morph is still a bomb.)
That is good. It just seems like a big advantage over a non-morph that in my view they would need to reduce the power for an equivalent card, so hopefully that will be the case as you indicate.
As just an obvious example, if they have a common blue 3/4 flyer for 5 mana which also has the Morph mechanic, that just seems way too good for a limited common because it can ambush a 3/3 flyer in addition to just being a 3/3 flyer for 5 which is okay. I am not worried or raising a flag here, it is just a random comment about what I hope will not occur.
I think that I like the mechanic (pending details) because it gives us more to think about.
It's a good discussion for those not familiar with Morph. You have to remember, un-Morphing a creature is usually quite expensive since it requires 3 + the Morph cost spread over multiple turns usually. So yea while you can ambush stuff, it's an expensive strategy. Also, in your example above, your 3/4 Flier would be vulnerable to Shock for a few turns in between -- major downside.
Morph is a very skill testing mechanic. Despite my complaints that it can turn into a coin flip guessing game (because it can) it's still a mechanic with a ton of design space and in-game challenges.
Also, because they're basically split card, both halves of a Morph creature are usually overcosted. A 3/4 Flier for 5 would be way above curve for a Morph creature.
3. Specifically, I prefer the wedge color combo. I like RUG, BUG and BGW and RUW. Only RWB leaves me cold.
I'm going to do a speculative prediction and say that RWB will be the strongest wedge in Limited. All the others are popular and strong (to the point where they have Legacy and Modern good-stuff decks); RWB is clearly the least-loved combination. In RTR block, I believe they intentionally made the least popular guilds the strongest (Selesnya and Rakdos in RTR, Boros in GP) while making the most popular ones the weakes (Izzet and Dimir) to balance things.
Also, because they're basically split card, both halves of a Morph creature are usually overcosted. A 3/4 Flier for 5 would be way above curve for a Morph creature.
Depends on the Morph cost - if it costs 5 to unmorph, it's not above the curve at all. Even for 4 it's very reasonable. Compare to Whip-Spine Drake. Being off-color is relevant of course (but we'll probably see that in Khans as well, as people said earlier) but creatures have gotten stronger since TS (and much strongers since Onslaught!).
You really just need to embrace the rage. I keep a small colony of hamsters next to my computer and every time I lose a match to mana screw I throw one against the wall.
It's a good discussion for those not familiar with Morph. You have to remember, un-Morphing a creature is usually quite expensive since it requires 3 + the Morph cost spread over multiple turns usually. So yea while you can ambush stuff, it's an expensive strategy. Also, in your example above, your 3/4 Flier would be vulnerable to Shock for a few turns in between -- major downside.
Morph is a very skill testing mechanic. Despite my complaints that it can turn into a coin flip guessing game (because it can) it's still a mechanic with a ton of design space and in-game challenges.
Also, because they're basically split card, both halves of a Morph creature are usually overcosted. A 3/4 Flier for 5 would be way above curve for a Morph creature.
That is good, and that makes sense. I think of that 3 mana cost 2/2 artifact creature in THS which you can sac and pay 3 to search for 2 lands, it can be good but it's slow. The ambush is the difference here but if the best half of a Morph creature is overcosted then that is good.
It's really important to remember that even if a creature has an attractive Morph cost, you're still sacrificing your board position by playing a 2/2 for 3 and then spending another turn's worth of mana (most likely) to turn it face up. You really need to ambush something, or get a 1-for-1 on the Morph trigger like Skirk Commando, to get value and turn your creature into a 2-for-1 (the creature that lives, plus the card's worth of value you received turning it face up). If you're Morphing just to trade 1-for-1 with a creature or spell, you're likely going to fall behind.
Morph has a lot of potential, but also a lot of hidden costs that are easily forgotten. That's why it's such a skill testing mechanic. Less talented players can shoot themselves in the foot trying to get "too cute" with Morph instead of just playing good creatures and spells.
Now that we've seen more of the set, I have to ask about mulligans. Is it safe to keep a 2-land hand in a 3-colour format with morph? Thanks! Reading through this thread has been very helpful for me.
Now that we've seen more of the set, I have to ask about mulligans. Is it safe to keep a 2-land hand in a 3-colour format with morph? Thanks! Reading through this thread has been very helpful for me.
Yes, but you're taking a risk. Time for some maths:
Assuming that you're running an 18 land deck, if you have 2 land in your opening hand you have 16 still in the deck, 33 cards total.
On the play you have 2 draws to get to 3 land. The probability of you whiffing is 17/33 * 16/32 ~= 0.25. That's basically a 1/4 chance of losing on the spot.
On the draw the probability of you whiffing is about 1/8.
So you definitely keep it on the draw, and I also would snap-keep it on the play if I also had a 2-drop. Maybe I would keep it on the play too. It depends on how fast/consistent I think my opponents deck is.
Riffing off of merl's excellent maths, let's look at the counteracting risk: how likely is a mulligan going to improve your chances?
The probability with an 18-land deck that you draw 3 or more lands from a 6-card hand (i.e., after a mulligan) is only 39% or so. (Also note that's any 3 lands, and if you want something more than 3 lands--e.g., at least two colors of mana production--it's lower.)
So if you're on the play, you have a 25% chance of missing your 3rd land drop, but a 61% chance of not improving your situation if you mulligan. In fact, you make it worse, as they probabilities of drawing that third land is ever-so-slightly lower.
I calculated the hypergeometric distribution for 18,X,40,6-X for X = 0, 1, and 2, added them up, and then subtracted that from 1.
Must find where I made the mistake...
Apparently HYPGEOMDIST is deprecated in Excel? I mean, that's the original formula I used, but when I use HYPGEOM.DIST, it gives the correct result.
Regardless!
Main point is that on the play and with two lands in hand, you have 3:1 odds of drawing a third land by the time you need it. On the draw with two lands, you have 7:1 odds of the same.
By contrast, you have more or less even money to get 3 lands if you mulligan.
It may be too early to say for sure, but from the spoilers I'm seeing a number of 2/* for 2 and 3/3 for 4 at common so my guess is that this set won't be defined by the 2/2 for 3 the way ONS block was even though there will be a higher number of them. So while missing your third land drop is still bad, if you have a bear or some other way to get back into the game it's not a death sentence the way it was in ONS.
Dolphan: Ah, right... I was somehow thinking (again, I plead very late night, etc.) that it wanted the number of non-sample cards, not the number of total cards viewed...
*sigh* Really, I do know my statistics... though my track record in this forum would not make you believe that... ><
Want to see some more maths before I'm convinced mulling 2 landers is correct, esp if you have 2 drops in hand. Sure, a 2-land 7-card hand has a pretty bad chance of hitting land drops 4 and 5 on time. But what are the odds of getting a better 6-card hand after mulling?
I did some more maths of my own to convince myself, using some basic Hypergeometric Distribution calculations and conditional probabilities.
First, some probabilities. Keeping a 2-land hand:
P(4 lands by turn 4, on play) = 48%
P(4 lands by turn 4, on draw) = 68%
P(5 lands by turn 5, on play) = 28%
P(5 lands by turn 5, on draw) = 47%
Keeping a 3-land hand:
P(4 lands by turn 4, on play) = 84%
P(4 lands by turn 4, on draw) = 92%
P(5 lands by turn 5, on play) = 60%
P(5 lands by turn 5, on draw) = 75%
Keeping a 4-land hand:
P(4 lands by turn 4, on play) = 1
P(4 lands by turn 4, on draw) = 1
P(5 lands by turn 5, on play) = 90%
P(5 lands by turn 5, on draw) = 94%
So you're about 50-50 on the mulligan actually making your hand worse. 29% shot of just getting another 2-lander only with fewer spells. 20% of getting a complete garbage unkeepable 6 (0-1 lands or 0-1 nonlands) that is much worse than your 2-lander 7. Even if you mull to a better 5, you are down 2 cards and probably won't get all your land drops anyway.
Start with a 7-card 2-lander and 18 lands in your 40 card deck. If you make a decision tree (mull vs don't mull) and then apply probabilities of hitting 4th and 5th land drops on time conditional on the number of lands in the hand you keep, assuming those 20% of garbage hands post-mull as failures, you get the following success rates for making land drops on time:
Don't mull to 6:
P(4 lands by turn 4, on play) = 48%
P(4 lands by turn 4, on draw) = 68%
P(5 lands by turn 5, on play) = 28%
P(5 lands by turn 5, on draw) = 47%
Mull to 6:
P(4 lands by turn 4, on play) = 60%
P(4 lands by turn 4, on draw) = 68%
P(5 lands by turn 5, on play) = 44%
P(5 lands by turn 5, on draw) = 56%
So once you factor in the failure rate (mulling to garbage) and the chance of just getting another 2-lander, you're nowhere near improving your chances by 50% by mulliganning! In fact, there is virtually no difference at all on the draw. On the play, you increase your chances of hitting your turn 5 land drop by a factor of 1.6. That's pretty respectable. So if your deck really needs to curve out its 5 drop immediately and you don't have ramp, I guess it's worth mulling away your 2-lander. Otherwise you're not really gaining much.
Keep in mind that with mulling you also add variance. A bird in the hand is worth 10 in the bush. You might already have a keepable 2-lander 7 with playable 2-drops and a chance of curving out. You're trading that for a gamble. Sure, in expectation your land drops might be slightly better. But there's also a 20% chance your 6 cards are terrible. And you're playing down 1 card.
So IMO you just keep your 7. And if you need to make your 5 drop that badly, just choose to draw instead of play.
Dolphan: Checking the first number for FTW's probabilities, I get the same if you use cumulative probabilities.
I.e., P(4 lands by turn 4, on play) is P(2 lands in top 3 cards) + P(3 lands in top 3 cards). The latter is about 10%, so that might explain the difference?
P(4 lands on turn 4, on play| X lands in opener)
= P(drawn at least 4 lands by turn 4 | X lands in opener)
= 1 - Sum_(k=0 to 3-X) P(k+X lands on turn 4 | X lands in opener)
We consider "at least" 4 instead of just the probability of exactly 4 because you are still making 4 lands on turn 4 if you have a 5th in hand (i.e. 2 in opener and drew a land each turn). The easiest way to get the "at least Y" probability in Excel is to go 1 - the cumulative distribution up to (Y - 1).
So I used, for example
P(4 lands on turn 4, on play| X lands in opener) = 1 - HYPGEOM.DIST(3-X successes, 3 draws, 18-X tot lands, 33 tot cards, TRUE)
P(5 lands on turn 5, on draw| X lands in opener) = 1 - HYPGEOM.DIST(4-X successes, 5 draws, 18-X tot lands, 33 tot cards, TRUE)
*34 cards if mulled to 6
Should be easily verifiable in Excel.
For the "mull to 6" success rates I used a little Bayesian logic, multiplying the chance of mulling to a certain keepable number of lands (2,3,4) by the probability of successfully making your land drops given that many cards in your 6-card opener and summing across the possibilities. I give a utility of 1 to making the required land drops, 0 otherwise. I gave the 0,1,5 and 6-land hands a utility of 0, since you're basically forced to mull to 5. This is a pessimistic assumption, since there is a non-zero chance of drawing a 3-land 5-carder with 2 very strong nonland cards that can get you back in the game and still curve out your land drops. You could work out the probabilities of making land drops after mulling to 5, etc. as well. But that would be an optimistic assumption, since you're not accounting for being down 2 cards. With fewer nonlands it's possible you just cant cast them, you have no creatures, they're only expensive drops, or they're just average creatures that don't make up for being down 2 cards. Sure, you can hypothetically mull to 2 and still curve out turn 3 Woolly Thoctar turn 4 Shield of the Oversoul [/game]. But it's way more likely your hand of 5 has fewer than 3 lands and that the spells you draw aren't strong enough to make up for -2 cards. We're just talking about the chance of curving out smoothly to turns 4 and 5, not of winning the game, so to me having to mull past 6 seems like a "failure". Probably worse than keeping your original 7.
Dolphan: Checking the first number for FTW's probabilities, I get the same if you use cumulative probabilities.
I.e., P(4 lands by turn 4, on play) is P(2 lands in top 3 cards) + P(3 lands in top 3 cards). The latter is about 10%, so that might explain the difference?
Yep, but you should use cumulative probabilities since you are making that 4th land drop in both cases.
FTW: Oh, agreed. I was just looking at that one number, is all.
If you want, and we can agree on land play priority, I could create a simulation model that will simulate this a lot... Dunno, feel very empirical right now...
Private Mod Note
():
Rollback Post to RevisionRollBack
To post a comment, please login or register a new account.
So that means I can tap mana in response?
1. I like the flavour. It's like a fixed Kamigawa: asian flavour, but without the all-legendary and (hopefully) without every single card having a hard-to-remember name.
2. I like multi-color sets. It lends to interesting new card with new ability combos.
3. Specifically, I prefer the wedge color combo. I like RUG, BUG and BGW and RUW. Only RWB leaves me cold.
4. The spoiled card strikes a good balance.
5. The arts hint at interesting things, like the ninja-like acrobatics, the black wedge not being too overly ultra-black (i.e. demons / vampire / zombie), much like in Kamigawa where the rats angle gave black an interesting feel instead of the same rehash.
So just personal preferences.
Except morph is better since you can unmorph things (and use their triggers) in response to a split second card.
Well, it's kinda like an instant that you can play at split second.
That is good. It just seems like a big advantage over a non-morph that in my view they would need to reduce the power for an equivalent card, so hopefully that will be the case as you indicate.
As just an obvious example, if they have a common blue 3/4 flyer for 5 mana which also has the Morph mechanic, that just seems way too good for a limited common because it can ambush a 3/3 flyer in addition to just being a 3/3 flyer for 5 which is okay. I am not worried or raising a flag here, it is just a random comment about what I hope will not occur.
I think that I like the mechanic (pending details) because it gives us more to think about.
Morph is a very skill testing mechanic. Despite my complaints that it can turn into a coin flip guessing game (because it can) it's still a mechanic with a ton of design space and in-game challenges.
Also, because they're basically split card, both halves of a Morph creature are usually overcosted. A 3/4 Flier for 5 would be way above curve for a Morph creature.
I'm going to do a speculative prediction and say that RWB will be the strongest wedge in Limited. All the others are popular and strong (to the point where they have Legacy and Modern good-stuff decks); RWB is clearly the least-loved combination. In RTR block, I believe they intentionally made the least popular guilds the strongest (Selesnya and Rakdos in RTR, Boros in GP) while making the most popular ones the weakes (Izzet and Dimir) to balance things.
Depends on the Morph cost - if it costs 5 to unmorph, it's not above the curve at all. Even for 4 it's very reasonable. Compare to Whip-Spine Drake. Being off-color is relevant of course (but we'll probably see that in Khans as well, as people said earlier) but creatures have gotten stronger since TS (and much strongers since Onslaught!).
That is good, and that makes sense. I think of that 3 mana cost 2/2 artifact creature in THS which you can sac and pay 3 to search for 2 lands, it can be good but it's slow. The ambush is the difference here but if the best half of a Morph creature is overcosted then that is good.
Morph has a lot of potential, but also a lot of hidden costs that are easily forgotten. That's why it's such a skill testing mechanic. Less talented players can shoot themselves in the foot trying to get "too cute" with Morph instead of just playing good creatures and spells.
Okay, thanks!
Yes, but you're taking a risk. Time for some maths:
Assuming that you're running an 18 land deck, if you have 2 land in your opening hand you have 16 still in the deck, 33 cards total.
On the play you have 2 draws to get to 3 land. The probability of you whiffing is 17/33 * 16/32 ~= 0.25. That's basically a 1/4 chance of losing on the spot.
On the draw the probability of you whiffing is about 1/8.
So you definitely keep it on the draw, and I also would snap-keep it on the play if I also had a 2-drop. Maybe I would keep it on the play too. It depends on how fast/consistent I think my opponents deck is.
The probability with an 18-land deck that you draw 3 or more lands from a 6-card hand (i.e., after a mulligan) is only 39% or so. (Also note that's any 3 lands, and if you want something more than 3 lands--e.g., at least two colors of mana production--it's lower.)
So if you're on the play, you have a 25% chance of missing your 3rd land drop, but a 61% chance of not improving your situation if you mulligan. In fact, you make it worse, as they probabilities of drawing that third land is ever-so-slightly lower.
Just a few numbers to keep in mind.
I calculated the hypergeometric distribution for 18,X,40,6-X for X = 0, 1, and 2, added them up, and then subtracted that from 1.
Must find where I made the mistake...
Apparently HYPGEOMDIST is deprecated in Excel? I mean, that's the original formula I used, but when I use HYPGEOM.DIST, it gives the correct result.
Regardless!
Main point is that on the play and with two lands in hand, you have 3:1 odds of drawing a third land by the time you need it. On the draw with two lands, you have 7:1 odds of the same.
By contrast, you have more or less even money to get 3 lands if you mulligan.
*sigh* Really, I do know my statistics... though my track record in this forum would not make you believe that... ><
I did some more maths of my own to convince myself, using some basic Hypergeometric Distribution calculations and conditional probabilities.
First, some probabilities.
Keeping a 2-land hand:
P(4 lands by turn 4, on play) = 48%
P(4 lands by turn 4, on draw) = 68%
P(5 lands by turn 5, on play) = 28%
P(5 lands by turn 5, on draw) = 47%
Keeping a 3-land hand:
P(4 lands by turn 4, on play) = 84%
P(4 lands by turn 4, on draw) = 92%
P(5 lands by turn 5, on play) = 60%
P(5 lands by turn 5, on draw) = 75%
Keeping a 4-land hand:
P(4 lands by turn 4, on play) = 1
P(4 lands by turn 4, on draw) = 1
P(5 lands by turn 5, on play) = 90%
P(5 lands by turn 5, on draw) = 94%
Mulling to 6:
P(2 lands) = 29%
P(3 lands) = 33%
P(4 lands) = 18%
P(0,1,5 or 6) = 20% (i.e. garbage!)
So you're about 50-50 on the mulligan actually making your hand worse. 29% shot of just getting another 2-lander only with fewer spells. 20% of getting a complete garbage unkeepable 6 (0-1 lands or 0-1 nonlands) that is much worse than your 2-lander 7. Even if you mull to a better 5, you are down 2 cards and probably won't get all your land drops anyway.
Start with a 7-card 2-lander and 18 lands in your 40 card deck. If you make a decision tree (mull vs don't mull) and then apply probabilities of hitting 4th and 5th land drops on time conditional on the number of lands in the hand you keep, assuming those 20% of garbage hands post-mull as failures, you get the following success rates for making land drops on time:
Don't mull to 6:
P(4 lands by turn 4, on play) = 48%
P(4 lands by turn 4, on draw) = 68%
P(5 lands by turn 5, on play) = 28%
P(5 lands by turn 5, on draw) = 47%
Mull to 6:
P(4 lands by turn 4, on play) = 60%
P(4 lands by turn 4, on draw) = 68%
P(5 lands by turn 5, on play) = 44%
P(5 lands by turn 5, on draw) = 56%
So once you factor in the failure rate (mulling to garbage) and the chance of just getting another 2-lander, you're nowhere near improving your chances by 50% by mulliganning! In fact, there is virtually no difference at all on the draw. On the play, you increase your chances of hitting your turn 5 land drop by a factor of 1.6. That's pretty respectable. So if your deck really needs to curve out its 5 drop immediately and you don't have ramp, I guess it's worth mulling away your 2-lander. Otherwise you're not really gaining much.
Keep in mind that with mulling you also add variance. A bird in the hand is worth 10 in the bush. You might already have a keepable 2-lander 7 with playable 2-drops and a chance of curving out. You're trading that for a gamble. Sure, in expectation your land drops might be slightly better. But there's also a 20% chance your 6 cards are terrible. And you're playing down 1 card.
So IMO you just keep your 7. And if you need to make your 5 drop that badly, just choose to draw instead of play.
I.e., P(4 lands by turn 4, on play) is P(2 lands in top 3 cards) + P(3 lands in top 3 cards). The latter is about 10%, so that might explain the difference?
P(4 lands on turn 4, on play| X lands in opener)
= P(drawn at least 4 lands by turn 4 | X lands in opener)
= 1 - Sum_(k=0 to 3-X) P(k+X lands on turn 4 | X lands in opener)
We consider "at least" 4 instead of just the probability of exactly 4 because you are still making 4 lands on turn 4 if you have a 5th in hand (i.e. 2 in opener and drew a land each turn). The easiest way to get the "at least Y" probability in Excel is to go 1 - the cumulative distribution up to (Y - 1).
So I used, for example
P(4 lands on turn 4, on play| X lands in opener) = 1 - HYPGEOM.DIST(3-X successes, 3 draws, 18-X tot lands, 33 tot cards, TRUE)
P(5 lands on turn 5, on draw| X lands in opener) = 1 - HYPGEOM.DIST(4-X successes, 5 draws, 18-X tot lands, 33 tot cards, TRUE)
*34 cards if mulled to 6
Should be easily verifiable in Excel.
For the "mull to 6" success rates I used a little Bayesian logic, multiplying the chance of mulling to a certain keepable number of lands (2,3,4) by the probability of successfully making your land drops given that many cards in your 6-card opener and summing across the possibilities. I give a utility of 1 to making the required land drops, 0 otherwise. I gave the 0,1,5 and 6-land hands a utility of 0, since you're basically forced to mull to 5. This is a pessimistic assumption, since there is a non-zero chance of drawing a 3-land 5-carder with 2 very strong nonland cards that can get you back in the game and still curve out your land drops. You could work out the probabilities of making land drops after mulling to 5, etc. as well. But that would be an optimistic assumption, since you're not accounting for being down 2 cards. With fewer nonlands it's possible you just cant cast them, you have no creatures, they're only expensive drops, or they're just average creatures that don't make up for being down 2 cards. Sure, you can hypothetically mull to 2 and still curve out turn 3 Woolly Thoctar turn 4 Shield of the Oversoul [/game]. But it's way more likely your hand of 5 has fewer than 3 lands and that the spells you draw aren't strong enough to make up for -2 cards. We're just talking about the chance of curving out smoothly to turns 4 and 5, not of winning the game, so to me having to mull past 6 seems like a "failure". Probably worse than keeping your original 7.
Yep, but you should use cumulative probabilities since you are making that 4th land drop in both cases.
If you want, and we can agree on land play priority, I could create a simulation model that will simulate this a lot... Dunno, feel very empirical right now...