Man, do i hate the beginning of a new deck. Unless you're lucky (or probably lsv) there are going to be some bad beats, especially in 100CS! Still stick in there, it'll come good eventually.
So you're playing a combo deck with no blue, so you have very few ways to protect your combo. So you need to look at other approaches to forcing your combo through. More discard will help, if they have no hand they can't stop you. The reusable nightmare void might help and Persecution in the sideboard. Recursion will mean that you can reuse your combo even if they stop it the first time round. how about adding a copy of volrath's stronghold and some of the green return to you hand spells? Better and quicker tutors would help get the combo out when you have one of the pieces. Try adding Tainted Pact for a hail mary pass. Finally better answers to other problems may help you - more removal like maelstrom pulse and putrefy. And what about adding spike weaver as protection from your own progenitus (at least for a few turns).
Hope some of this helps. I look forward to seeing the evolution of the deck!
Comments here are amazingly awesome. To Shard- please dont insult yourself ever again when saying that your degree is History and Bachelors. Ever! Please. I graduate with a bachelors this semester with an Asia Focused History degree with both 400 level Korean and Japanese and it isnt that great. Spike, Timmy, and Vorthos can get a Bachs degree in any social science they please without any difficulty. Obtaining anything higher is the challenge. Do not call yourself a Historian either please (unless you have credentials). It insults me and others who will graduate. On the other hand, if your had your Masters, I respect you and are free to make assumptions as not everyone aspires to obtain a higher degree.
Onto the article- I agree with you wholeheartedly. I try to always make bad plays to inform the opponent that I suck. I will sometimes (to my benefit) use my main phase inefficiently (ie. play a creature + attack) However, I will try to do this only once during my entire match. If they attack with mana open, I will most likely block. Id rather lose an OK creature to a trick instead of losing a winning creature to a trick.
I believe that by playing the opposite of common sense is a brilliant way of playing your opponent(whether its Magic, Games, Life, School Etc). It will force your opponent to think differently and lose his state of balance and not understand the situation at hand.
I was running a Jund-Rock deck, with several choices you have listed. I dropped the deck because I wasn't winning versus the prevalent Gobos. Subterranean Shambler is a great idea.
If you truly are going to replace the Armadillo Cloaks then I would suggest splashing black instead of white. Putrid Leech is a legitimate house and Blightning is great versus control.
The other suggestions I would have, is put Golgari Browscale in the board. Its interaction with Wild Mongerel is probably the most powerful sb tool I have versus aggro. Finally, (this is probably personal taste) I would try Yavimaya Elder over Civic Wayfinder. In my list, I was running mostly Green, with only black and red splashes. Being able to find both Red and Black with one card is phenomenal. Your own Shamblers would hit it, but it has a sac outlet to negate that cost.
Also, thanks again for the Subterranean Shambler idea. The other deck that I run is Blink. It has been very good versus every aggro deck other than Gobos. Blink on Shambler is a house.
The biggest issue with passing the Bog Tatters pack 2 is who you are passing them to. Its pretty clear from pack 1 that several of the players to your right are taking black because you're able to cut it so completely. That takes help. Then rather than reap the rewards of cutting a color in pack 2 you start taking blue cards.
Now on a card-for-card basis I can't fault any of those picks, but in the general sense of who's drafting what it was bound to bite you. They got both the better black in packs 1/3 and the black you passed pack 2.
I honestly think your best deck would have come from money drafting Ob Nix and then playing UG. Or paired him with G perhaps (harrow, etc). I just don't see how to get full value from Ob Nix in a limited deck that isn't half green.
@ Godot - You said: "I don't understand your P3P4 comment, though. If I'm taking AEther Figment specifically to fill the two-drop hole, then I am most definitely in a world of trouble.
Finally, I didn't say I *expected* this deck to make the finals."
To take the last part first. I didn't mean to imply that you had, so no offense intended - just that *I* didn't expect it.
I disagree about being in a world of trouble if you are taking Figment as a two drop. AEther figment, while no world beater, should not only be thought of as a 5 drops, though that IS safest for curve purposes since you would rather be heavy on two drops in this format than heavy on 5 drops. However it fulfills two important roles: 1) Potential finisher 2) It trades with all manner of 2/1 two drops. I am perfectly happy to use it a "removal" to stem the beats in this kind of deck which need to survive to the mid-game in a situation where you don't have to leave Umara Raptor back to block.
It is hard to say what would have happened if you had picked the scorpion. It probably didn't put anyone into black but it seems like it probably did solidify someone one in black two seats away based on what you saw in pack two.
Finally, two drops matter so much in this format that scrabbler (who I love) makes my final deck after your draft. With Ob in my pocket I probably cut the sky ruin to make room on the play (you should rule the skies anyway)... and an island on the draw.
41 cards sometimes makes sense if you have a high curve in a slow format, and you really need both 23 cards and 18 lands (or 22 and 19) but that is pretty rare.
A couple of things:
Djinn is just pronounced "Jinn"
As long as I am talking pronunciation I have a gripe left over from your last videos: the card is Forbidden OrCHARD (like an apple orchard!) *not* Forbidden OrCHID.
Good call, I failed to see that option, and I think it would have been correct. Not much was going to save me that game, but that would have given me a slightly better shot.
Good comments, and I too am looking to plug holes in pack three (see: P3P3 comment), but sometimes the cards just aren't there to plug them, particularly when you are fighting for black in Zendikar. None of the pick changes you suggest truly solve the deck's glaring problems of lack of removal or lowering the curve, except maybe the puma P2P5, which I can totally get behind.
P1P5 was a very interesting pick, though. I find there come points in pack one after starting black where you have to decide between a "fight for black" pick or setting yourself up for a solid second color, and that was the case for this pick. An unanswerable question is, if I take the scorpion there, does that deter someone from going black and free up better black for me later? Maybe. It felt like the better value was cutting the blue to ensure at least *one* color would flow freely.
I don't understand your P3P4 comment, though. If I'm taking AEther Figment specifically to fill the two-drop hole, then I am most definitely in a world of trouble.
Finally, I didn't say I *expected* this deck to make the finals. I simply said it had the potential to make the finals with a little luck, which, while perhaps the least-helpful sentence I've ever typed on Limited Magic, is still true. I wish the second game hadn't busted, it really demonstrated what the deck was capable of "with a little luck."
Tough opponent for a first round.
Evaluating your draft, I wondered what would have happened if you had forced black, by picking Hagra croc third pick, and then the Scorpion 5th. It probably would've been a gamble, but there's also the chance that one of your leftward drafters was considering what to take for a second color and any of those might have been a defining factor. I agree that both are bellow your pick, especially the croc, but you could have gambled it. Also second pack I would have picked the Rite, it's so much fun when you replicate an Umara Raptor like I did once, that in my mind it offsets the risks involved in having your target killed or bounced.
I was also wondering, since your first pick being a bomb and all, would you have considered droping it entirely if you hadn't got any other black card to go along with it? In other words, would you have splashed for it?
By the way, how's Oliver doing this days? Did he get any Magic related items from Santa?
dude...acaidic slime over looter? Phantom warrior over the looter? no way, especially when the slime is off color. Looter is the stone cold nuts in that format. Also, i would hav taken the ranger over the verdict but thats just me. Btw ice cage is great in GU.
and always, always, always play 40 in limited. 41 isnt even an option. unless your deck is literally all 1 spell, some cards will be better than others, so just play the better ones and draw them more.
In Game 1 when your opponent has a Giant Scorpion with a Spidersilk Net you choose to bounce the Scorpion and attack with the Eel to get 4 points of damage in, knowing that the Scorpion is going to come down again next turn. My inclination in this spot is to attack with the Eel, let him block with the Scorpion, and then Into The Roil the Spidersilk Net, letting the Eel trade with the Scorpion.
Killing off the Scorpion seems really important here, as the only ways you have to deal with it are Roil/Cancel and trading with a 6/6 or bigger Ob Nixilis. You can't "race" a netted Scorpion, you only have two damage sources that the Scoprion can't stop (Ob + Landfall and Aether Figment). Also, there is a fair chance that he doesn't have another creature he can cast since there was no turn 4 play, and even if he does have another creature it won't shut down your entire deck the way the Scorpion does ...
If it's relatively easy (I have to assume it is) to edit out the "dead air" while waiting on OK's from your opponent in the videos, I'd suggest doing that, but otherwise top notch as always, sir.
Bummer - sometimes things don't go your way and you get beaten by landwalkers, but that was hardly this decks only weakness, halfway through your draft I could tell you were in trouble. Here are the notes I took:
p1p5 - Fight for black and take the great scorpion - it is practically removal.
p1p8 - You always have to take a first ally sometime. Why not get one under your belt early, and essentially for free?
p2p5 - Puma and Roil both interest me more. Its a really tough call, but with two raptors in the fold I probably grab the puma.
p3p2 - You got conservative, the Rite was the correct four drop to pick. Don't forget that it doubles as an ally.
p3p4 - You are lacking early drops, I like Sky Ruin more than most, but here I pick up the curve filling two drop, especially since I would have grabbed the puma earlier.
p3p6 - Again, you are missing early drops more than anything. With two cancels in the pile the hatchling is an easy pick here. You may only start one, but having an extra to board in could be huge.
Ultimately, my concern with your few low drops was not your undoing (in this round) and even a single disfigure falling to you would have been a huge boon. However, I like to start drafts by building towards an "archetype" (however strictly or loosely you care to define such things), but by pick 3 or so in the second pack you usually have a good idea of what your actual deck's strengths and weaknesses are and you have to stop building towards a general concept and actually start plugging specific holes and building on existing strengths. I don't think this draft was successful at that, and did not expect to see this deck make the finals.
But, it only exists if both players have knowledge of the true prior. If you wanted to exend the anaylsis to the case where both players have some belief about P(Good), you would need to introduce a conjugate prior and each player would need knowledge of how the other's conjugate is parameterized; and then a semi-separating equilibrium will only exist if both players believe themselves and believe that the other believes that there are more good players than bad *and that the payoff matrix is well-behaved]. But you would need to know the conjugate parameterization in order to discover it. You could extend this ad-infinitum by having conjugates of the conjugates.
I'm not trying to argue about how the model is set up per se, but to point out the fact that I don't feel bayesian game theory is generally a good model for this type of game. When we apply mathematical models like this to a complex game we fall victim to our assumptions. I just feel that it important for readers to realize this.
Take the tatters over what? I can't get behind taking Bog Tatters over Welkin Tern, AEther Figment, or Umara Raptor in UB, which is what four of the five opportunities for a tatters were in this draft. The roil vs. tatters P3P5 was close, and I said as much, but it's results-oriented and counterproductive to look back at a draft like this and say, "I should have taken all those Bog Tatters, then I wouldn't have lost to them."
Or.. you could do what your opponent did and take the Tatters for your black deck, knowing so many opponents will be playing black and allowing them to attack unhindered. Also making his deck worse against your's as instead of his unblockable Tatters he has to attack with his.. 1/3 Scorpion or whatever he picks up to replace it.
Tracking down "exact" payoffs for certain outcomes isn't as important as you might think. Game theory functions as long as you have a complete and transitive preference ranking over outcomes. In a game of Magic, it's pretty obvious you have a preference ranking of win --> draw --> loss. From there, you can assign and numerical values, and the equilibria will remain the same. (The values of a certain mixed strategy will change, but we are more interested in knowing that the equilibrium is to mix than tracking down what that exact mixing ratio is.)
Here, I have assumed that cracking a Terramorphic Expanse immediately is a worse play, but it is more important to identify your opponent as a good player or a bad player. My conclusions follow directly from those premises. Pete's conclusions follow directly from the premises if you believe that the timing of the fetch is more important than knowing whether your opponent is good or bad. The nice thing about game theory is that it spells out all these assumptions and tells you exactly what is required for your conclusions to follow from them.
Props on the use of game theory, I so love that stuff i made it my major!
In my opinion the advice is way more general than just "when to crack your expanse", the point is: making the play that is not "optimal" in the common sense will sometimes throw your opponent off guard. Example: tapping out before attacking when you know it won't matter (opp tapped out) so that the opp instinctively will believe to some degree that 1. you are a bad player, 2. you dont have a trick in hand, 3. if you dont play something before attacks next turns you may have drawn a trick.
I assigned the 80-20 split because I did not want to make things super-complicated by having nature draw from p and 1-p. The semi-separating equilibrium exists as long as there are more good players than bad players, which I believe is an accurate assumption. I chose .8 and .2 because they fit those requirements.
I like the effort, but you have to be careful here. I do not believe that this is properly modeled as a Bayesian game. To put it simply, there is no prior from nature known to all players. You have arbitrarily assigned an 80-20 split between good and bad players in nature and assume that both players have the same belief in this prior. Additionally, you assume that the true payoff matrix is known to both players. The analysis simply breaks down without these assumptions.
Man, do i hate the beginning of a new deck. Unless you're lucky (or probably lsv) there are going to be some bad beats, especially in 100CS! Still stick in there, it'll come good eventually.
So you're playing a combo deck with no blue, so you have very few ways to protect your combo. So you need to look at other approaches to forcing your combo through. More discard will help, if they have no hand they can't stop you. The reusable nightmare void might help and Persecution in the sideboard. Recursion will mean that you can reuse your combo even if they stop it the first time round. how about adding a copy of volrath's stronghold and some of the green return to you hand spells? Better and quicker tutors would help get the combo out when you have one of the pieces. Try adding Tainted Pact for a hail mary pass. Finally better answers to other problems may help you - more removal like maelstrom pulse and putrefy. And what about adding spike weaver as protection from your own progenitus (at least for a few turns).
Hope some of this helps. I look forward to seeing the evolution of the deck!
Comments here are amazingly awesome. To Shard- please dont insult yourself ever again when saying that your degree is History and Bachelors. Ever! Please. I graduate with a bachelors this semester with an Asia Focused History degree with both 400 level Korean and Japanese and it isnt that great. Spike, Timmy, and Vorthos can get a Bachs degree in any social science they please without any difficulty. Obtaining anything higher is the challenge. Do not call yourself a Historian either please (unless you have credentials). It insults me and others who will graduate. On the other hand, if your had your Masters, I respect you and are free to make assumptions as not everyone aspires to obtain a higher degree.
Onto the article- I agree with you wholeheartedly. I try to always make bad plays to inform the opponent that I suck. I will sometimes (to my benefit) use my main phase inefficiently (ie. play a creature + attack) However, I will try to do this only once during my entire match. If they attack with mana open, I will most likely block. Id rather lose an OK creature to a trick instead of losing a winning creature to a trick.
I believe that by playing the opposite of common sense is a brilliant way of playing your opponent(whether its Magic, Games, Life, School Etc). It will force your opponent to think differently and lose his state of balance and not understand the situation at hand.
As usual, I got some good tips here.
I was running a Jund-Rock deck, with several choices you have listed. I dropped the deck because I wasn't winning versus the prevalent Gobos. Subterranean Shambler is a great idea.
If you truly are going to replace the Armadillo Cloaks then I would suggest splashing black instead of white. Putrid Leech is a legitimate house and Blightning is great versus control.
The other suggestions I would have, is put Golgari Browscale in the board. Its interaction with Wild Mongerel is probably the most powerful sb tool I have versus aggro. Finally, (this is probably personal taste) I would try Yavimaya Elder over Civic Wayfinder. In my list, I was running mostly Green, with only black and red splashes. Being able to find both Red and Black with one card is phenomenal. Your own Shamblers would hit it, but it has a sac outlet to negate that cost.
Also, thanks again for the Subterranean Shambler idea. The other deck that I run is Blink. It has been very good versus every aggro deck other than Gobos. Blink on Shambler is a house.
The biggest issue with passing the Bog Tatters pack 2 is who you are passing them to. Its pretty clear from pack 1 that several of the players to your right are taking black because you're able to cut it so completely. That takes help. Then rather than reap the rewards of cutting a color in pack 2 you start taking blue cards.
Now on a card-for-card basis I can't fault any of those picks, but in the general sense of who's drafting what it was bound to bite you. They got both the better black in packs 1/3 and the black you passed pack 2.
I honestly think your best deck would have come from money drafting Ob Nix and then playing UG. Or paired him with G perhaps (harrow, etc). I just don't see how to get full value from Ob Nix in a limited deck that isn't half green.
@ Godot - You said: "I don't understand your P3P4 comment, though. If I'm taking AEther Figment specifically to fill the two-drop hole, then I am most definitely in a world of trouble.
Finally, I didn't say I *expected* this deck to make the finals."
To take the last part first. I didn't mean to imply that you had, so no offense intended - just that *I* didn't expect it.
I disagree about being in a world of trouble if you are taking Figment as a two drop. AEther figment, while no world beater, should not only be thought of as a 5 drops, though that IS safest for curve purposes since you would rather be heavy on two drops in this format than heavy on 5 drops. However it fulfills two important roles: 1) Potential finisher 2) It trades with all manner of 2/1 two drops. I am perfectly happy to use it a "removal" to stem the beats in this kind of deck which need to survive to the mid-game in a situation where you don't have to leave Umara Raptor back to block.
It is hard to say what would have happened if you had picked the scorpion. It probably didn't put anyone into black but it seems like it probably did solidify someone one in black two seats away based on what you saw in pack two.
Finally, two drops matter so much in this format that scrabbler (who I love) makes my final deck after your draft. With Ob in my pocket I probably cut the sky ruin to make room on the play (you should rule the skies anyway)... and an island on the draw.
41 cards sometimes makes sense if you have a high curve in a slow format, and you really need both 23 cards and 18 lands (or 22 and 19) but that is pretty rare.
A couple of things:
Djinn is just pronounced "Jinn"
As long as I am talking pronunciation I have a gripe left over from your last videos: the card is Forbidden OrCHARD (like an apple orchard!) *not* Forbidden OrCHID.
Nassif had a bazaar looking sealed curve and went with 41-42 cards. He curved out at 8 with 3 eight drops and wanted to hit a land every turn.
Good call, I failed to see that option, and I think it would have been correct. Not much was going to save me that game, but that would have given me a slightly better shot.
Good comments, and I too am looking to plug holes in pack three (see: P3P3 comment), but sometimes the cards just aren't there to plug them, particularly when you are fighting for black in Zendikar. None of the pick changes you suggest truly solve the deck's glaring problems of lack of removal or lowering the curve, except maybe the puma P2P5, which I can totally get behind.
P1P5 was a very interesting pick, though. I find there come points in pack one after starting black where you have to decide between a "fight for black" pick or setting yourself up for a solid second color, and that was the case for this pick. An unanswerable question is, if I take the scorpion there, does that deter someone from going black and free up better black for me later? Maybe. It felt like the better value was cutting the blue to ensure at least *one* color would flow freely.
I don't understand your P3P4 comment, though. If I'm taking AEther Figment specifically to fill the two-drop hole, then I am most definitely in a world of trouble.
Finally, I didn't say I *expected* this deck to make the finals. I simply said it had the potential to make the finals with a little luck, which, while perhaps the least-helpful sentence I've ever typed on Limited Magic, is still true. I wish the second game hadn't busted, it really demonstrated what the deck was capable of "with a little luck."
Thanks for taking the time to comment.
Tough opponent for a first round.
Evaluating your draft, I wondered what would have happened if you had forced black, by picking Hagra croc third pick, and then the Scorpion 5th. It probably would've been a gamble, but there's also the chance that one of your leftward drafters was considering what to take for a second color and any of those might have been a defining factor. I agree that both are bellow your pick, especially the croc, but you could have gambled it. Also second pack I would have picked the Rite, it's so much fun when you replicate an Umara Raptor like I did once, that in my mind it offsets the risks involved in having your target killed or bounced.
I was also wondering, since your first pick being a bomb and all, would you have considered droping it entirely if you hadn't got any other black card to go along with it? In other words, would you have splashed for it?
By the way, how's Oliver doing this days? Did he get any Magic related items from Santa?
dude...acaidic slime over looter? Phantom warrior over the looter? no way, especially when the slime is off color. Looter is the stone cold nuts in that format. Also, i would hav taken the ranger over the verdict but thats just me. Btw ice cage is great in GU.
and always, always, always play 40 in limited. 41 isnt even an option. unless your deck is literally all 1 spell, some cards will be better than others, so just play the better ones and draw them more.
In Game 1 when your opponent has a Giant Scorpion with a Spidersilk Net you choose to bounce the Scorpion and attack with the Eel to get 4 points of damage in, knowing that the Scorpion is going to come down again next turn. My inclination in this spot is to attack with the Eel, let him block with the Scorpion, and then Into The Roil the Spidersilk Net, letting the Eel trade with the Scorpion.
Killing off the Scorpion seems really important here, as the only ways you have to deal with it are Roil/Cancel and trading with a 6/6 or bigger Ob Nixilis. You can't "race" a netted Scorpion, you only have two damage sources that the Scoprion can't stop (Ob + Landfall and Aether Figment). Also, there is a fair chance that he doesn't have another creature he can cast since there was no turn 4 play, and even if he does have another creature it won't shut down your entire deck the way the Scorpion does ...
heya :)
If it's relatively easy (I have to assume it is) to edit out the "dead air" while waiting on OK's from your opponent in the videos, I'd suggest doing that, but otherwise top notch as always, sir.
To put it more succinctly: a game does not reach nash equilibrium unless both players have true knowledge of nature.
Bummer - sometimes things don't go your way and you get beaten by landwalkers, but that was hardly this decks only weakness, halfway through your draft I could tell you were in trouble. Here are the notes I took:
p1p5 - Fight for black and take the great scorpion - it is practically removal.
p1p8 - You always have to take a first ally sometime. Why not get one under your belt early, and essentially for free?
p2p5 - Puma and Roil both interest me more. Its a really tough call, but with two raptors in the fold I probably grab the puma.
p3p2 - You got conservative, the Rite was the correct four drop to pick. Don't forget that it doubles as an ally.
p3p4 - You are lacking early drops, I like Sky Ruin more than most, but here I pick up the curve filling two drop, especially since I would have grabbed the puma earlier.
p3p6 - Again, you are missing early drops more than anything. With two cancels in the pile the hatchling is an easy pick here. You may only start one, but having an extra to board in could be huge.
Ultimately, my concern with your few low drops was not your undoing (in this round) and even a single disfigure falling to you would have been a huge boon. However, I like to start drafts by building towards an "archetype" (however strictly or loosely you care to define such things), but by pick 3 or so in the second pack you usually have a good idea of what your actual deck's strengths and weaknesses are and you have to stop building towards a general concept and actually start plugging specific holes and building on existing strengths. I don't think this draft was successful at that, and did not expect to see this deck make the finals.
But, it only exists if both players have knowledge of the true prior. If you wanted to exend the anaylsis to the case where both players have some belief about P(Good), you would need to introduce a conjugate prior and each player would need knowledge of how the other's conjugate is parameterized; and then a semi-separating equilibrium will only exist if both players believe themselves and believe that the other believes that there are more good players than bad *and that the payoff matrix is well-behaved]. But you would need to know the conjugate parameterization in order to discover it. You could extend this ad-infinitum by having conjugates of the conjugates.
I'm not trying to argue about how the model is set up per se, but to point out the fact that I don't feel bayesian game theory is generally a good model for this type of game. When we apply mathematical models like this to a complex game we fall victim to our assumptions. I just feel that it important for readers to realize this.
Well spotted, it is indeed a faerie.
Of the 5 he saw in the draft, they were up against good cards each time:
1: Umara Raptor
2: Umara Raptor
3: Aether Figment / Into the Roil
4: Welkin Tern
5: Into the Roil
I think those are all the correct picks over the Tatters.
Take the tatters over what? I can't get behind taking Bog Tatters over Welkin Tern, AEther Figment, or Umara Raptor in UB, which is what four of the five opportunities for a tatters were in this draft. The roil vs. tatters P3P5 was close, and I said as much, but it's results-oriented and counterproductive to look back at a draft like this and say, "I should have taken all those Bog Tatters, then I wouldn't have lost to them."
Or.. you could do what your opponent did and take the Tatters for your black deck, knowing so many opponents will be playing black and allowing them to attack unhindered. Also making his deck worse against your's as instead of his unblockable Tatters he has to attack with his.. 1/3 Scorpion or whatever he picks up to replace it.
Tracking down "exact" payoffs for certain outcomes isn't as important as you might think. Game theory functions as long as you have a complete and transitive preference ranking over outcomes. In a game of Magic, it's pretty obvious you have a preference ranking of win --> draw --> loss. From there, you can assign and numerical values, and the equilibria will remain the same. (The values of a certain mixed strategy will change, but we are more interested in knowing that the equilibrium is to mix than tracking down what that exact mixing ratio is.)
Here, I have assumed that cracking a Terramorphic Expanse immediately is a worse play, but it is more important to identify your opponent as a good player or a bad player. My conclusions follow directly from those premises. Pete's conclusions follow directly from the premises if you believe that the timing of the fetch is more important than knowing whether your opponent is good or bad. The nice thing about game theory is that it spells out all these assumptions and tells you exactly what is required for your conclusions to follow from them.
Props on the use of game theory, I so love that stuff i made it my major!
In my opinion the advice is way more general than just "when to crack your expanse", the point is: making the play that is not "optimal" in the common sense will sometimes throw your opponent off guard. Example: tapping out before attacking when you know it won't matter (opp tapped out) so that the opp instinctively will believe to some degree that 1. you are a bad player, 2. you dont have a trick in hand, 3. if you dont play something before attacks next turns you may have drawn a trick.
I assigned the 80-20 split because I did not want to make things super-complicated by having nature draw from p and 1-p. The semi-separating equilibrium exists as long as there are more good players than bad players, which I believe is an accurate assumption. I chose .8 and .2 because they fit those requirements.
Yeah, brutal. And a triple B mythic planeswalker in the same draft. I'd still draft Ob for the money though. Tough noogies.
I like the effort, but you have to be careful here. I do not believe that this is properly modeled as a Bayesian game. To put it simply, there is no prior from nature known to all players. You have arbitrarily assigned an 80-20 split between good and bad players in nature and assume that both players have the same belief in this prior. Additionally, you assume that the true payoff matrix is known to both players. The analysis simply breaks down without these assumptions.