[Deck] Death and Taxes

[Vandalize]

I'm still playing RW Death and Taxes. Magus of the Moon is too good to be on the bench.

Lands [23]
4 Wasteland
4 Rishadan Port
3 Cavern of Souls
2 Karakas
4 Arid Mesa
4 Plains
2 Plateau

Creatures [26]
4 Mother of Runes
4 Stoneforge Mystic
4 Thalia, Guardian of Thraben
3 Flickerwisp
2 Recruiter of the Guard
2 Phyrexian Revoker
2 Sanctum Prelate
2 Thalia, Heretic Cathar
2 Magus of the Moon
1 Mirran Crusader

Spells [11]
4 AEther Vial
4 Swords to Plowshares
1 Batterskull
1 Sword of Fire and Ice
1 Umezawa's Jitte

Sideboard [15]
2 Path to Exile
2 Ethersworn Canonist
2 Rest in Peace
2 Council's Judgement
1 Wear/Tear
1 Leonin Relic-Warder
1 Magus of the Moon
1 Sanctum Prelate
1 Phyrexian Revoker
1 Gideon, Ally of Zendikar
1 Containment Priest

This list has been running so smoothly. Recruiter of the Guard printing was the best thing that could have happened, nobody wants to cash out for Imperial Recruiters. I don't own three pieces of Karakas, so I'm using 3 Cavern of Souls instead, but it has been working fine. Any comments or advice on the list?

[Kayradis]

So, got back into it.
Dusted my old list and ran it to a Shitty 1-2 result at the Weekly Legacy here. I'm rusty as hell, I'm not gonna lie.

Here's my list.
Need some massive updates.
Feel free to comment

7 Plains
4 Wasteland
4 Rishadan Port
2 Cavern of Souls
3 Karakas
2 Horizon Canopy

4 Mother of Runes
2 Serra Avenger
4 SFM
2 P Revoker
4 TGT
1 Fiend Hunter
2 Vryn Wingmare
2 Mirran Crusader
2 Mangara of Corondor
4 Flickerwisp

4 Aether Vial
4 Swords to Plowshare
1 Umezawa's Jitte
1 Sword of Fire and Ice
1 Batterskull

Sideboard is always changing.

[mykatdied]

So, got back into it.
Dusted my old list and ran it to a Shitty 1-2 result at the Weekly Legacy here. I'm rusty as hell, I'm not gonna lie.

Here's my list.
Need some massive updates.
Feel free to comment

7 Plains
4 Wasteland
4 Rishadan Port
2 Cavern of Souls
3 Karakas
2 Horizon Canopy

4 Mother of Runes
2 Serra Avenger
4 SFM
2 P Revoker
4 TGT
1 Fiend Hunter
2 Vryn Wingmare
2 Mirran Crusader
2 Mangara of Corondor
4 Flickerwisp

4 Aether Vial
4 Swords to Plowshare
1 Umezawa's Jitte
1 Sword of Fire and Ice
1 Batterskull

Sideboard is always changing.

I must say that running 22 land with the 6 "thalia" configuration. Do you ever have issues using swords to plow when you need it? Personally not a fan of wingmare. I would select either crusader or avenger. Depends on what is more prevalent, punishing fire or abrupt decay. Recruiter and prelate really are amazing. I would suggest you check into them

[RobNC]

I tried THC as a 3-of in Maverick last night, hoping for the turn 2 dream drop. She either got countered or was bolted on sight; the only time she stuck she got Terminused away. I will admit I have the "emotional attachment" that was mentioned two pages back, as I've been trying hard to make her work. 10-20% of the time she's insane and practically wins you the game, but the rest of the time she's a little underwhelming. Usually when I drew her I wished it was a different card, and I felt that way when I last played D&T a few weeks ago. The infect player I played was the only one who was scared of her, because he didn't have main deck removal.

I'm not going to try to convince anyone not to run her, though, and I'll probably still continue to run a copy or two myself. I'm just at the point now where I feel we have other better 3-drop hate pieces (Sanctum Prelate) or beaters.

[Secretly.A.Bee]

I think she certainly has her good matchups. I also think that in some others, she's the first thing you side out.

I have no emotional attachment to THC. I do think that against SnT, lands, the mirror, and most goyf decks, she is decent to broken, depending on what turn it is. She does die a lot, but I'm not sure that makes her bad, but rather a pretty good bait card.

Sent from my SM-N920V using Tapatalk

[Poron]

imho Thalia 2 is great with Thalia 1, mana dorks and Winter Orb

She can really shine the best in Maverick.

4 Deathrite
1 Noble
4 Gsz
4 Thalia GoT
3 THALIA, HC
4 Mother of Runes


Tarmogoyf, Dark Confidant, Abrupt Decay and Stoneforge Mystics for endless variants..

[Secretly.A.Bee]

If they play a Dryad Arbor, the amount of turn 2 thalia 2.0 goes up quite a bit. However, how does this pertain to DnT? Not trying to be a jerk, but this isn't useful for me. I have an event tomorrow morning and ifanyone would give an opinion or suggestion on my list, I'd appreciate it.

Sent from my SM-N920V using Tapatalk

[RobNC]

I didn't intend to derail this into a Maverick vs. D&T discussion for THC, I only mentioned it to make a point that my feeling of her, even on turn 2, is rather unchanged. Amazing some of the time, "meh" the rest of the time.

If they play a Dryad Arbor, the amount of turn 2 thalia 2.0 goes up quite a bit. However, how does this pertain to DnT? Not trying to be a jerk, but this isn't useful for me. I have an event tomorrow morning and ifanyone would give an opinion or suggestion on my list, I'd appreciate it.

Sent from my SM-N920V using Tapatalk

What's your current list? The one you posted the other day with two Mox Diamond?

[Secretly.A.Bee]

Yes.

Sent from my SM-N920V using Tapatalk

[zebotc]

Alright I've done over 100 test games and some number crunching and have created a list of data points that might be very helpful for us in figuring out how many of each card we should actually be running based on how much we want to draw it, how many multiples we want in a game, etc. I know you could use a hyper geometric calculator for this but they sometimes output data in harder to read ways that don't relate to magic. So I've created a list of sorts that shows the various probabilities out of 100 games to draw a given card as a 1 of, 2 of, 3 of, and 4 of. Hope it's helpful!


MTG Card Probabilities Based on Number in Deck

100 Trials (Games)

The Deck
- 50 Islands
- 4 Chandra, Torch of Defiance ---------- The 4 of
- 3 Vampire Nighthawk --------------------- The 3 of
- 2 Thalia, Guardian of Thraben ---------- The 2 of
- 1 Elvish Mystic ------------------------------ The 1 of

Turn 0 or Opener (7 beginning cards)

	Times 1 Was Drawn	Times 2 Were Drawn	Times 3 Were Drawn	Overall Percent of Getting at Least 1
4 of	44	2	1	47%
3 of	27	4	0	31%
2 of	17	3	0	20%
1 of	12	0	0	12%
Nothing Drawn	28	0	0	28%

After Drawing 6 Additional Cards (Around Turn 5-7) CARD NOT IN OPENER

	Times 1 Was Drawn	Times 2 Were Drawn	Times 3 Were Drawn	Overall Percent Drawing at Least 1 if it Was Not in the Opener
4 of	28	3	1	32%
3 of	24	3	0	27%
2 of	22	1	0	23%
1 of	11	0	0	11%

After Drawing 6 Additional Cards CARD ALREADY IN OPENER

	Times 1 Was Drawn	Times 2 Were Drawn	Overall Percent of Drawing at Least 1 Copy of a Card After it Was Already in the Opener
4 of	11	2	12%
3 of	6	0	6%
2 of	1	0	1%

Nothing Drawn At All After 6 Draws

Nothing Drawn After Opener

33 Games

33% of Games

Overall Percent of Drawing a Card in a Game

	Overall Percent of Drawing at Least 1	Overall Percent of Drawing 2 or More
4 of	79%	19%
3 of	58%	13%
2 of	43%	5%
1 of	23%	N/A

[iatee]

If you do use a hypergeometric calculator, you'd see that the chance of drawing one of your 4-ofs in your opening 7 is 40% not 47% and in fact, you'd only get to 47% with a 5-of. That's a pretty big difference, and pretty important since 'What's the chance of hitting one of my 4-ofs in my opening 7' is actually a number worth keeping in the back of your pocket. Most of the other results you got don't seem to be as far off but there's still no real reason not to just make the same table using calculated results, since right now you're just approaching the real numbers instead of just getting them...

[zebotc]

I used real life testing because even though in a game theoretically those numbers would be the actual chance I wanted to see realistically in a given set of games how the draws would pan out. Of course if I did over a million games yeah the numbers would be fairly close. But I wanted to test the hyper geometric calculations to reality and create a data set based on a given set of games. It shows that there is substantial variation in just what a hyper geometric calculator calculates which makes sense because we're dealing with probabilities. Your right though it's not entirely accurate but I think it servers as a useful example because instead of just using the actual calculated values it puts them to the test.

Something incredibly interesting I found was that running a 3 of vs a 4 of seemed to significantly reduce your chances of seeing multiples of that card in a given game (a 4 of is around 1.8-2 times as likely to be drawn twice). It makes sense that it would reduce it slightely but I was surprised how significant the difference was.

This leads me to some conclusions about the number of cards we should actually be running.

First ROTG- I still think running this as a 4 of is way way to much however, I could see running a 3 of with a specific list but I think a 2 of would be the perfect sweet spot because you don't want multiples and it's only good in some games.

Flikerwisp and ROTG - I feel like if you add 2 ROTG you could safely take out 1 flickerwisp because by the turn you need to draw it your chances of finding it are essentially the same if not higher and a four of seems like it's too much.

our sideboard may need to be addressed too, I'm starting to have a feeling running one ofs that are non tutorable might be a waste of a sideboard slot. Possibly the best option would be to put lots of tutorable cards in it, and cards as 2-3 ofs for consistency. 1 of pithing needle and cataclysm is starting to make less sense because youre only going to draw it in a small amount of games even post sideboard. But then again running one ofs has been beneficial in some games in the past but how beneficial has it actually been? I honestly don't know yet but it's an interesting idea.

[iatee]

The calculations are more reality than your simulation though. You can both calculate the real answer and calculate how far off the real answer results are likely to be if you run 100 hands at once. There was a microscopic chance that you ran the simulations times and drew Chandra 100% of the time. But had that happened nobody would have anything new to learn from that either. The chance to draw her in your opening hand is still 40%. And because you play one game at a time, that's true for one game, 100 games and a million games.

[zebotc]

The calculations are more reality than your simulation though. You can both calculate the real answer and calculate how far off the real answer results are likely to be if you run 100 hands at once. There was a microscopic chance that you ran the simulations times and drew Chandra 100% of the time. But had that happened nobody would have anything new to learn from that either. The chance to draw her in your opening hand is still 40%. And because you play one game at a time, that's true for one game, 100 games and a million games.

I agree with you that the calculator shows the actual probability of drawing a card. But I just thought it would be intersting to see it in a controlled test. It might be a good idea though to add the theoretical values to the table to show the variation that exists from the tested to mathematically calculated ones.

[Fry]

I don't think that the "real world" stats is going to do anything off of only ~100 games. It doesn't really add up. There aren't enough data points for it to be spot on correct. Correct for that set of games, sure, but not representative as a whole.

There are so many things that affect the randomness factor of drawing cards in the opening 7, shuffling techniques (sometimes lack there of), number of shuffle iterations, old sleeves that stick together a little bit vs new sleeves (especially significant with the old sleeves if the same card is next to each other.

While I think that your "data" points for your ~100 games is interesting and gives some moderately close stats, I think it's ultimately just something for fill the void of empty time. Doesn't really show the true probabilities once the exercise if repeated 1,000,000+ times.

[raudo]

I'm still playing RW Death and Taxes. Magus of the Moon is too good to be on the bench

This list has been running so smoothly. Recruiter of the Guard printing was the best thing that could have happened, nobody wants to cash out for Imperial Recruiters. I don't own three pieces of Karakas, so I'm using 3 Cavern of Souls instead, but it has been working fine. Any comments or advice on the list?

I would definitely try to add couple of Nahiris in the list. Lock and win.

[zebotc]

I don't think that the "real world" stats is going to do anything off of only ~100 games. It doesn't really add up. There aren't enough data points for it to be spot on correct. Correct for that set of games, sure, but not representative as a whole.

There are so many things that affect the randomness factor of drawing cards in the opening 7, shuffling techniques (sometimes lack there of), number of shuffle iterations, old sleeves that stick together a little bit vs new sleeves (especially significant with the old sleeves if the same card is next to each other.

While I think that your "data" points for your ~100 games is interesting and gives some moderately close stats, I think it's ultimately just something for fill the void of empty time. Doesn't really show the true probabilities once the exercise if repeated 1,000,000+ times.

It was just a thing I did when I had some spare time because I thought it would be fun to test the hypergeometric calculator to actual games, obviously the hypergeometric calculator is the most accurate representation of it but since my values were fairly close I figured it'd be interesting and possibly useful to post them just to give people a real world example of what could happen in a given set of games. I think if anything the data sets I found prove more that we should be using hypergeometric calculation in some situations when choosing card amounts.

Anyways regardless of those data points and what they represent, it really seems like we should consider more the actual probability of drawing a card when we put it in our deck. I saw some people saying it was crazy to go down to 3 flickerwisp for example (Which makes sense because in classic DnT doing that is almost detrimental because it's such a key card in our strategy), but honestly with the addition of recruiter it makes a lot of sense because your chances of getting a flickerwisp actually go up with recruiter and your normal chances of drawing specifically one naturally are honestly not that much different.

Now the amount of recruiters we should possibly be running - IMO 4 is to many, your chances of double drawing are incredibly high and you honestly only need one if any in a given game normally. So 1-3 would probably be alright with 2 being a good middle ground.

Then there is the thing with our sideboard and running non tutorable 1 ofs and if it's actually worth it. It's starting to seem like running random 1 of's with no consistency honestly doesn't help that much and the one random game that would be won from it could have been made up for by running more consistent SB options.

[Warden]

Interesting discussion that mirrors a conversation from my office. Honestly, there is a lot of good thought here. To throw my 2 cents in the ring: the quantitative approach (stats, hypergeometic distrib) needs qualitative feedback (stories, opinion on performance, real world experiences from IRL matches) to form a true assessment. You need both sides of the coin to get a clearer picture.

I love stats and % but don't buy into them as the be-all, end-all approach to running cards in MTG. Magic will never have a perfect list and even if you had the best "on paper" distribution of card slots, there are a myriad of events and circumstances that mitigate the "perfect" list.

There are 2 very big areas where hypergeometric distributions don't click with MTG:
1 - Dumb luck. Others have brought this up, but sometimes you just never see [insert card] all game or all match. Even when the math is in your favor. Mana screw/mana flood as big examples. "Where the hell is card X -- I have 3 in the deck and I still haven't seen one by turn 7" as yet another. Hypergeometric is correct to a degree, but there are times you run more/less of a card based on your own dumb luck/reaction to prior experiences. With respect to Recruiter and the other flex slots for DnT, some players had success with a toolbox. Others preferred 2x set of key cards (i.e. 2 prelate, 2 THC, 2 Avenger). I have not even touched upon the dumb luck of the opponent. Sometimes you lucky topdeck. Sometimes they can't find shit and you profit. The math/odds don't necessarily support why you dreamcrushed because "on paper" they should have seen an answer. Additionally, IRL too many random factors influence the % of drawing/seeing cards (including but not limited to inadequate shuffling/randomizing)

2 - Circumstances/Decision Trees. I mean this not only for cards with tutor abilities (fetches, SFM, Recruiter), but also for your opponent. When building a deck, you need to factor in "will this card be destroyed? Countered? Better in bulk or as a singleton?" It drives me crazy when players tell me "you don't need so many X in the deck because you [can get by/only need] 1 copy". Perhaps the one and only copy you draw in the opening 7 has a tendency to die/get countered/discarded. Now what? Perhaps you like a card in multiples. SFM as a nice example. If a second one shows up, I'm not usually upset.

The bottom line I'm trying to make is that we can definitely take some of the hypergeometric discussion but preface it with the need to think about qualitative feedback and better our discussion. AFAIK, some of the best feedback in the thread has been about the impact the new cards have had on the game. Was a card great? Would you have won easier/not lost if multiples hit the table (prelate I'm very curious about exactly this)? Maybe something was a dud. THC for some folks doesn't shine. I'd love to hear why. Maybe she clogs the hand with multiples (going with or against the odds of you drawing her). Maybe all THC does is come out to get countered or take a bolt in the face. Hypergeo doesn't give us the full picture. This is where we need civil and insightful discussions beyond "this card/list sucks" or "you disagree with me but I'll just keep beating my own drum".

[iatee]

Definitely agree Warden. I feel like THC is like the Donald Trump of DnT in that its introduction has really torn this thread apart and destroyed our civility.

I played 20 games yesterday with a friend on Mentorless hard control Miracles today, with the following list:

4 Thalia (baby Frank)
4 SfM
3 Revoker
1 Spirit of the Lab
4 Mom
3 Prelate
3 Recruiter
3 Flickerwisp
1 Banisher Priest
4 Vial
4 STP
Bskull
Sofi
Jitte
3 Cavern
3 Karakas
4 Wasteland
4 Port
9 Plains

SB:
3 Path
2 Rest in Peace
Relic Warder
2 Mirran Crusader
2 Ethersworn
Sword of War and Peace
Palace Jailer
Faerie Macabre
Ghost Quarter
Mangara

Friend is a very good Miracles player, I don't want to talk him up too much cause he reads this thread, but he's a tournament end boss type. I went 6-4 pre-board and 6-4 post-board, so 12-8 overall, but we both got the sense that the numbers didn't reflect how ahead I was pre-board (I started out 6-2 and got unlucky with my last 2 games) and how it was more even post-board (got luckier and had a decent number of Vial hands.) I would put the EW rate with this more at 65-70% preboard and 50% postboard assuming they have an average amount of cards to bring in and both players are playing competently.

Palace Jailer (SB experiment) didn't do anything but eat a counterspell twice, so I have nothing to say about it other than "It's bad if you hardcast it without a Cavern and they have the card 'Counterspell' in hand." I swapped it with THC (not the drug) after 5 games, which also ate a counterspell once and was good t4 one game with Karakas. Never had a desire to tutor for either of those in any of the 20 games.

Over those games I think I only tutored for:

Flickerwisp (when I could safely Flicker my Recruiter)
Another Recruiter (when I wasn't sure I wouldn't get wiped by a Terminus if I tried)
Prelate (often)
and I think a Revoker once

The other big SB card, Sword of War and Peace, felt mostly useless - with Prelate on 1 plus Moms already blanking STPs, I had no reason to spend a million mana to get the same effect. I wanted some generic anti-Miracles card for the board that I would be seeing regularly, but it just wasn't very good. SfM and the equipment in general felt less important because Chalice effects provide virtual card advantage and stretch out the game in your favor, so you don't feel pressured to win as quickly. They're not necessarily pulling ahead by just making their land drops.

I could see trying a 4th Prelate in the 75 and finding room for one Magus in the main since this build has an iffy g1 vs. Eldrazi.

[Luca Grease]

What destroyed the civility of this thread was the shit posting, really. People not having a clue about statistics making fallacious inferences, people parroting what they've read ad nauseam without bothering to articulate and justify their points, people CONSTANTLY discussing other decks, etc. I don't remember the quality of discussion in this thread ever being so low.

We might have different opinions about THC's worth in DnT, but those kinds of posts will only sabotage meaningful discussion about the card, and inevitably infuriate the reader (namely, me).

For example, I actually explained somewhat in detail why I thought THC to be quite good against Miracles, and I think it's no coincidence that the card performed well for you when you were able to resolve it. Keep in mind that most Miracle lists will play mentors, and THC is quite good against that card (assuming you are already in an aggressive position). A card's capability to turn a somewhat favorable position into a dominating one has often been dismissed as irrelevant or "win more", but I think that's flawed thinking. Against slow decks like Miracles, Shardless, or Stoneblade, we'll almost always take the initiative, while they have the tools to catch up and nullify that initial advantage. People also used to dismiss Mirran Crusader, because any connection with an equipped creature was enough to swing the game in your favor. I think we've all won enough games off of double strike + equipment to recognize that reasoning as not always valid.

I haven't tested SoWaP yet, but I imagine part of the allure of that card against Miracles is that it will allow you to swing past an army of monk tokens.

I think your assessment of the Miracles MU is pretty accurate. Even though the impact of our new cards will be somewhat lessened by them adapting their sideboards (and, in some cases, mainboards too), the result is still a net gain, pushing the match-up into firmly (though not overwhelmingly) favorable territory overall.

As for your list being pretty soft against Eldrazi g1: that's part of the reason why I've decided to go down to 2 Mentors. It is an amazing card and I don't think 3 is "wrong" by any stretch of the imagination, but I'd rather not feel so naked against a good percentage of the field, especially when some of those matchups are fairly even (Eldrazi, Maverick, mirror) and losing g1 because you drew too many dead/mediocre cards puts you in a pretty difficult position. However, the card is still close to an auto-win against 50% of the field, so I can understand someone wanting to polarize their chances a little bit... Having an extra 1 or even 2 in the board might be a good idea, though, if you can afford the spots.

Speaking of Eldrazi, I feel like the only way to make Magus doable in mono-white is to move up to 4 Caverns: I think this is a legit setup with a lot of potential, but it also places quite a few deckbuilding constrictions on your list: you're probably gonna have to move down to 2 Flickerwisps, and you basically can't play any non-human sb card requiring WW (Cataclysm, Gideon, Judgement). If you value the reliability of mono-white, THC is still pretty good at improving those percentage points against Eldrazi, while also providing value against virtually every deck Prelate is weak against. In general, lower raw lockout power than either Magus or Prelate, but broader and more reliable.

Finally, one question: is Mangara still worth it? I have decided to give it another chance in the mainboard by relegating the Banisher Priest to the side (I figured I could afford this because my maindeck is a bit more resilient to a fast Eldrazi start anyway), but I didn't really have many chances to evaluate it in the tournament I played with it...