I thought about our strategy. Like many people are saying, including myself, we are great against Aggro, Midrange, against Control its favourable to slightly less favourable, depending on archetype, but we have so many dead cards against Combo, or Spell based decks. I think therefore its not the best option to go for and grindy, slow MB. I kinda want to build aroung the idea of having something similar to a morph board, but wihout changing the strategy to much, making us vulnerable to certain other MUs, because the slots in the SB are all occupied. Therefore I want to go aggressive MB, or lets say some more aggression than we used to have, and potentially changing to our beloved grindy deck postboard. I think for this purpose YP is one of the best options we have. It has some downside if it gets drawn lategame, but I think he is ok, in game 1 we want to be aggressive anyways. But, its also important to be able to grind in Game 1 as well, if needed. Therefore I would always stick to Liliana, Discard and Lingering Souls.
One advantage of this is, that you basically always have the option to board out some YP for other hate cards, without diluting our grindy strategy, at which we are best.
I think the best way to achieve the strategy that you just talked about would be to include 2-3x Monastery Mentor on your list. It is probably the most aggressive card we have access to in the 3 cmc slot and he doesn't dilute the deck's ability to grind either. Sure, we don't have access to blue, but having access to burn and discard is still a great way to use him. I would drop 1 K-command and 1 Helix from your list to make room for him.
I'm very interested by the vision, the land-denial option, the control orientation and the pox use
I think that I'll test this...but in proxy for the moment (fulminator, bitterblossom and some lands not yet in my pocket).
An opinion ?
I've always been suprised that there aren't more people on this forumn using Smallpox. Mardu has access to some of the best colors to take advantage of the card. I'm running a more aggro-based variant of the deck you posted, using Young Pyromancer, Bloodghast, and Kalitas, Traitor of Ghet as my threats. I've had a lot of sucess with it so far playing it in my modern meta.
i would say the format is very much medium speed ... gbx is very popular... as are the chord decks.. so you definitely would need 1 for those matchups.... and cryptic isn't a bad draw against aggro since it helps turn the corner... i play 2 and have had good success with it but i probably play a different version than the rest....
Your list looks typical, although I would make a few changes. Geist kills quick once he hits. once I land him I do everything I can to clear the path for him to attack safely. and in a lot of cases I will attack in, knowing he will die just to deal 4 to their face. this is not a control deck and the lightning bolts can also help close out games as well.
I would cut a cryptic command to run the 4th geist. and cut ajani for the 4th lightning helix. you could also try boros charm instead of elspeth to speed your clock up if you'd like.
I don't really think the 4th Geist is necessary. Depending on how you build this deck, it can really be a control deck. The list that @nomemorial posted shares 70/75 cards with the typical UWR control build. I think you're right that there are two different ways to play the deck, which is why it is so flexible compared to a pure control build. You can go in gun's blazing against other control and combo decks or you can play control yourself against aggro and midrange. For that reason, I think 4 Geist's is never a good idea because it takes away from the flexibility of the deck. When you are forced to play the control role, sitting with two Geists in hand is really a burden.
I'm surprised people are suggesting cutting Cryptic Command. The meta looks like it is slowing down, and it is one of our best cards in slower match-ups especially against BGx. It's probably our biggest source of card advantage as well. I agree that it's bad against aggro, but at least then the bounce + draw or tap + draw is relevant for Geist getting through. I'd say it's a 1-2 of, but never more.
So just to make sure I get it: now that you have fixed the model, the result is 4 cantrips equal a land in terms of hitting land drops right?
But are we talking about 4 generic cantrips, or Serum Visions? In other words: does the model now include the scry 2?
Yep! That is right. 4 generic cantrips is equivalent to an additional land in deck in terms of hitting land drops. Serum visions is a little bit better than that, and is equivalent to about 1.5 lands.
The model does account for the scry 2. That part was correct beforehand, I just had to tweak the "draw a card" math a little bit.
On another note, has anyone else tried TiTi in the sideboard? I've found it to be wonderful in some match-ups, but really not great for others. I am still not sure if I would rather just have sweepers over him. TiTi is house against elves and little zoo, but against affinity just isn't fast enough. I think I need to get my hands on a copy of Avacyn as well.
Edit: by the way, for Delver decks, I think there's something your model doesn't account for: mulligans. If you run a Delver deck with 0 cantrips, you can't keep a one lander. You need to mulligan. But you can keep a one lander with Serum Visions. So cantrips increase the number of keepable hands. This however isn't true for our Geist deck. 'Cause, at least IMHO, you can't keep a one lander, even with Visions in it. This deck still needs 2 lands in the opening hand to work.
My model can't account for the scry part of the mulligan, but it can account for mulligans otherwise. It would just decrease your cards drawn by the number of times you mulliganed. But now that we know the percentages, you are still ~90% likely to hit two lands by turn 2, which is much higher than a Delver deck.
So I've been lurking for quite some time... and I think I'm finally making the switch over to midrange. I was trying to go a prowess/tempo build, but my meta is very competitive, and honestly it gets eaten alive Lol so I hope this thread is active, because I'm going to have a lot of questions and want a LOT of feedback
Welcome Rutherjc19! I remember seeing you around the Jeskai Prowess forum. I made the switch to midrange a few months ago and haven't looked back.
There must be something wrong with the model. The results on Delver decks show it better than the result on our deck. 10 cantrips + 18 lands is not the same as 0 cantrips and 19 lands. It just can't be.
You are totally right, the results surprised me as well. I've played with SV in the past and it did seem to do more work with hitting land drops than 1%. I think I made one error in my model and will be correcting for it once I get home. I need to be adding the probabilty of drawing SV and Remand turns 1 and 2 into the "Cards Drawn" variable. I'll let you know once I've edited the post above.
EDIT: Fixed the math in my equation above, sorry for the confusion. @TheAller was definitely correct that something wasn't quite right. With the fixed maths, the results show that 4 cantrip cards are equal to a single land in terms of hitting land drops. Thanks for tuning in and I hoped you find the information useful!
I’ve come up with a good statistical model of how Serum Visions and Remand affect a deck’s ability to hit land drops. This was a fun problem to work on, I hope it helps and makes sense. If you want to skip to the results, go straight to the TL;DR section.
Derivation
First off, It’s very straightforward to figure out how likely you are to draw a certain number of lands by a certain turn. You can do that simply with combination operators:
Where C(n,k) is the combination operator, landsDrawn is the number of lands you have drawn, cardsDrawn is the number of cards you have drawn in the course of the game, and landsDeck is the number of lands in your deck. This is useful information because it can tell you how likely you are to hit a certain land drop by a given turn. For example, if Deck A runs 22 lands, it is 87% likely to hit the second land drop by turn 2. If Deck B has 24 lands, it is 91% likey. Checks out so far.
Serum Visions and Remand have some effect on these percentages because they act as cantrips and SV lets you scry 2. Before we start on how these affect the probabilities, let's list the basic assumptions:
1. You always cast both cards if you draw them
2. You always scry non-land cards to the bottom of your library and you always leave scryed lands on top
Without these assumptions, modeling their effect on lands drops would be much more difficult. With the first assumption, our model only really becomes useful after turn 2, since we need 2 mana to cast either card. We’ve already seen that both Decks A and B are likely to do this, so let’s move on to the effect of each card.
Serum Visions effectively acts to thin out your deck. It’s a cantrip, so you automatically draw a card, but by assumption 2, you will also move non-land cards to the bottom of your deck. For example, if you cast SV and scry into 2 non-land cards to the bottom, you will have thinned out your deck by 3 cards total. If you scry into 1 non-land card and 1 land card, you will have thinned out the deck by 2 cards. In the math, this modifies the number of cards in your deck:
Cards in Deck w/ SV = Cards in Deck - (Probability of Drawing SV)*(SV Thinning Function) [2]
Where,
SV Thinning Function = (1 + Chance of Scrying into two lands + Chance of Scrying into one land) [3]
The probability of drawing serum visions can be solved with Equation [1] by replacing some variables. The equation for SV Thinning Function is a bit more cumbersome to type out, but follows the logic stated above and is a function of lands in deck, lands drawn, cards drawn, and cards in deck.
Remand is much easier to account for, since it is a simple cantrip. You will always draw just one card when it is played. This also modifies the number of cards in your deck:
Cards in Deck w/ SV and Remand = Cards in Deck - (Probability of Drawing SV)*(SV Thinning Function) - (Probability of Drawing Remand)*(1) [4]
Both Serum Visions and Remand also increase the number of cards you have drawn in a game. Subsequently, in the math the probability of drawing either card has to be added to the number of cards drawn in a game:
Cards Drawn w/ SV and Remand = Cards Drawn + (Probability of Drawing SV) + (Probability of Drawing Remand) [5]
(Side note: This last formula is actually a bit tricky, because it is recursive. The probability of Drawing SV and Remand increases with the number of cards you have drawn, which increases based on your percent chance of drawing either card, which increases with the number of cards you have drawn... etc. We would need to use an iterative method to solve this, but for simplicity we will just use the first term)
If we substitute Equation [4] and [5] into Equation [1] for the value 60, we have a function that tells us how often we will hit a land drop based on the number of cards drawn, lands in deck, and if we are running SV or Remand. Awesome!
TL;DR
So what does this mean? We can input some additional information to compare Decks A and B:
We already know that our model is only useful after turn 2, so let's look at how often both decks can hit their third land drop by turn 3, on the play:
Deck A: 80% (as opposed to 72% without SV and Remand)
Deck B: 83% (as opposed to 79% without Remand)
How about the fourth land drop by turn 4, on the play:
Deck A: 65% (as opposed to 54% without SV and Remand)
Deck B: 69% (as opposed to 63% without Remand)
All of these percentages go up if you are on the draw.
WayTL;DR
Running 3 Serum Visions increases your percent chance of hitting land drops 3-4 on time by a more than 4-5%. A full playset of remand increases your chances by about 4%. Each land increases this percentage by 3-4%. This analysis was strictly based on how these cards affect land drops, they obviously have their own very useful abilities. What we did see is that if you are running 4 cantrips, they will have the same effect on land drops as a single land would. This is why delver-style decks can get away with running a fewer land than necessary. Any less cantrips than that shouldn't be considered as having a significant effect.
For those interested, I have the calculator programmed in Mathematica. I can look into making some sort of web based form of it.
You're right that Serum Visions and Remand helps... luckily we have math to find out just how much! Let me do some number crunching and get back to you. I've always been curious about this myself, I used to run Serum Visions in my list for a while.
Well I like to run the less amount of lands I can in my decks in general, to try and get all the juice I can later on instead of hitting useless lands. I noticed I was flooding sometimes. You will always have the occasional game where you get stuck to 2 lands and never find the third or stuff like that. Even if you run 23 or 24. But by running 22, at least you decrease dramatically the risk of mana flooding too badly.
To be more specific:
* I've never been on 24 for this deck, always run 23. I think 24 is really too much. Jund runs 24 with a higher curve, no Visions, no Remand, 5 utility lands;
* If you look you'll see many of the people who run 23 lands in this archetype run a singleton Ghost Quarter or Tec Edge. That's not really like having 23 lands, more like 22,5. And I don't like the singleton Quarter in this deck, think the downsides are more than the upsides;
* 3x Visions is what seals the deal for me. 22 lands without Quarters or Tec Edges + 3 copies of Visions and a fullset of Remand has proven to be enough to hit the land drops;
* Running 4 copies of Colonnade instead of 3 follows the same logic of having "as much juice as I can";
I want to agree with you on these points, but I think 22 lands is too few for one important reason: hitting land drops on time. At 22 lands, you statistically won't be hitting your 4th land drop until turn 5 or later. This is true with or without Ghost Quarters. It also seems like 22 lands and 4 Celestial Collonades are mutually exculsive. You will have that slow starts more often, and you are less likely to get to 6 mana in the course of a game.
This deck should be running at least 23-24 lands to make sure we hit our lands drops on time, even if that leads to flooding out sometimes. I personally don't mind flooding out since it usually means that Celestial Collonade will come online faster, and that you can make stronger plays with a Snapcaster in hand (i.e. flashingback a Cryptic or a Sphinx's Rev).
I have a land drop curve generator at home that I can post later for anyone that's interested.
I like your list a lot, I'm going to test something like that but cutting a path and the sphinx's revelation for the fourth remand and mana leak (I never liked sphinx outside commander). I'll try the thing in the ice MB since my meta is mostly aggro and the geist of saint traft on the SB.
Thanks! If you use Thing in the Ice mainboard, you should also swap the Eiganjo Castle for another Collonade or fetch as to not tip off your opponent as to what's on your sideboard. I wouldn't recommend cutting Sphinx's Revelation since can be a very strong finisher, but that is a pretty minor change.
You should just use the Archangel Avacyn you pulled in place of Ajani Vengeant. I will also be making that swap if I happen to pull her.
Because TiTi can win games after it sweeps the board. It fixes the one problem our deck has when playing the control side - closing out games. I tested with this card all weekend, and it makes the aggro matchup closer to 70% when played with 3-4 in the deck. It is a bad card when there is plenty of spot removal, but that's why it is a great compliment to Geist.
I'm not saying to not run Timely Reinforcements and Anger of the Gods, they are both great cards. But, try running 1x Anger, 1x Timely, and 3x Titi in your sideboard and you will not be dissapointed.
What do you guys think of Thing in the Ice as a sideboard card? I am currently trying out 3 of them to swap out with Geist in the aggro match-ups. It seems like Geist and Thing are complementary sideboard partners. Geist is great in the control/midrange matchup where Thing is great in the aggro matchup.
My list is rather threat-light, running only 10 creatures mainboard, so it might not be a great option for people running a higher amount. I'm using a few sideboard slots that would have been taken up by sweepers for the Thing.
Another epiphany: Grim Lavamancer. The deck has no real way to utilize all the cards we discard and fetchlands we crack. Lavamancer is also great against Company decks which I think will be very popular post rotation.
Grim Lavamancer is an awesome idea. I like it as a 2 of in the sideboard, since he isn't relevant in every match-up.
You should reconsider Indulgent Aristocrat as a 2-3 of on the mainboard. He doesn't seem like much at first, but has single-handedly won a few matches for me. He allows us to race other tribal decks in size of our creatures because he has the same awesome synergy with Bloodghast that Kalitas, Traitor of Ghet does. At worst early game, he eats a removal spell without huge loss in tempo. At worst late game, he eats a removal spell and we sac him in response, giving all creatures +1/+1 counters. A 1-drop lord that can impact the board right away is pretty neat.
I want to play Vampires, not an Aristocrats deck with some Vampires I have no intention to start collecting BW Vampires / Aristocrats / ...
I rather help out with the Aggro / Tempo list, as that list seems way more tribal at this point. My lists where merely some quick decks I've came up with (as an example), they weren't exactly heavily thought over pieces of work. The suggestions that are given at this point seem good and constructive.
I think you are right that we don't want to be playing an Aristocrats deck with Vampires in it. @ThatOneNinja has a really good analysis of how to build this deck a page back - we should cut all of the gimmicky synergies for raw power. But, I am not sure if that raw power will take form in a tempo deck right now. The power density for Vampires currently loaded in the 3 and 4 drop slots, pushing us towards midrange. The best version of this deck will obviously not be as fast as Merfolk or Elves, but it won't be as slow as Jund either. We're probably going to land somewhere right in the middle.
Stromkirk Noble suffers the same fate as Geist of Saint Traft - he is amazing when the board is wide open, but can be easily stonewalled. I personally like Indulgent Aristocrat in this slot.
I agree that only testing can determine the best build. A few of us are doing our testing on Cockatrice or XMage for the decks, you should join in.
I think the best way to achieve the strategy that you just talked about would be to include 2-3x Monastery Mentor on your list. It is probably the most aggressive card we have access to in the 3 cmc slot and he doesn't dilute the deck's ability to grind either. Sure, we don't have access to blue, but having access to burn and discard is still a great way to use him. I would drop 1 K-command and 1 Helix from your list to make room for him.
I've always been suprised that there aren't more people on this forumn using Smallpox. Mardu has access to some of the best colors to take advantage of the card. I'm running a more aggro-based variant of the deck you posted, using Young Pyromancer, Bloodghast, and Kalitas, Traitor of Ghet as my threats. I've had a lot of sucess with it so far playing it in my modern meta.
Would you mind posting your list?
I don't really think the 4th Geist is necessary. Depending on how you build this deck, it can really be a control deck. The list that @nomemorial posted shares 70/75 cards with the typical UWR control build. I think you're right that there are two different ways to play the deck, which is why it is so flexible compared to a pure control build. You can go in gun's blazing against other control and combo decks or you can play control yourself against aggro and midrange. For that reason, I think 4 Geist's is never a good idea because it takes away from the flexibility of the deck. When you are forced to play the control role, sitting with two Geists in hand is really a burden.
I'm surprised people are suggesting cutting Cryptic Command. The meta looks like it is slowing down, and it is one of our best cards in slower match-ups especially against BGx. It's probably our biggest source of card advantage as well. I agree that it's bad against aggro, but at least then the bounce + draw or tap + draw is relevant for Geist getting through. I'd say it's a 1-2 of, but never more.
Yep! That is right. 4 generic cantrips is equivalent to an additional land in deck in terms of hitting land drops. Serum visions is a little bit better than that, and is equivalent to about 1.5 lands.
The model does account for the scry 2. That part was correct beforehand, I just had to tweak the "draw a card" math a little bit.
On another note, has anyone else tried TiTi in the sideboard? I've found it to be wonderful in some match-ups, but really not great for others. I am still not sure if I would rather just have sweepers over him. TiTi is house against elves and little zoo, but against affinity just isn't fast enough. I think I need to get my hands on a copy of Avacyn as well.
My model can't account for the scry part of the mulligan, but it can account for mulligans otherwise. It would just decrease your cards drawn by the number of times you mulliganed. But now that we know the percentages, you are still ~90% likely to hit two lands by turn 2, which is much higher than a Delver deck.
Welcome Rutherjc19! I remember seeing you around the Jeskai Prowess forum. I made the switch to midrange a few months ago and haven't looked back.
You are totally right, the results surprised me as well. I've played with SV in the past and it did seem to do more work with hitting land drops than 1%. I think I made one error in my model and will be correcting for it once I get home. I need to be adding the probabilty of drawing SV and Remand turns 1 and 2 into the "Cards Drawn" variable. I'll let you know once I've edited the post above.
EDIT: Fixed the math in my equation above, sorry for the confusion. @TheAller was definitely correct that something wasn't quite right. With the fixed maths, the results show that 4 cantrip cards are equal to a single land in terms of hitting land drops. Thanks for tuning in and I hoped you find the information useful!
Derivation
First off, It’s very straightforward to figure out how likely you are to draw a certain number of lands by a certain turn. You can do that simply with combination operators:
Where C(n,k) is the combination operator, landsDrawn is the number of lands you have drawn, cardsDrawn is the number of cards you have drawn in the course of the game, and landsDeck is the number of lands in your deck. This is useful information because it can tell you how likely you are to hit a certain land drop by a given turn. For example, if Deck A runs 22 lands, it is 87% likely to hit the second land drop by turn 2. If Deck B has 24 lands, it is 91% likey. Checks out so far.
Serum Visions and Remand have some effect on these percentages because they act as cantrips and SV lets you scry 2. Before we start on how these affect the probabilities, let's list the basic assumptions:
1. You always cast both cards if you draw them
2. You always scry non-land cards to the bottom of your library and you always leave scryed lands on top
Without these assumptions, modeling their effect on lands drops would be much more difficult. With the first assumption, our model only really becomes useful after turn 2, since we need 2 mana to cast either card. We’ve already seen that both Decks A and B are likely to do this, so let’s move on to the effect of each card.
Serum Visions effectively acts to thin out your deck. It’s a cantrip, so you automatically draw a card, but by assumption 2, you will also move non-land cards to the bottom of your deck. For example, if you cast SV and scry into 2 non-land cards to the bottom, you will have thinned out your deck by 3 cards total. If you scry into 1 non-land card and 1 land card, you will have thinned out the deck by 2 cards. In the math, this modifies the number of cards in your deck:
Where,
The probability of drawing serum visions can be solved with Equation [1] by replacing some variables. The equation for SV Thinning Function is a bit more cumbersome to type out, but follows the logic stated above and is a function of lands in deck, lands drawn, cards drawn, and cards in deck.
Remand is much easier to account for, since it is a simple cantrip. You will always draw just one card when it is played. This also modifies the number of cards in your deck:
Both Serum Visions and Remand also increase the number of cards you have drawn in a game. Subsequently, in the math the probability of drawing either card has to be added to the number of cards drawn in a game:
(Side note: This last formula is actually a bit tricky, because it is recursive. The probability of Drawing SV and Remand increases with the number of cards you have drawn, which increases based on your percent chance of drawing either card, which increases with the number of cards you have drawn... etc. We would need to use an iterative method to solve this, but for simplicity we will just use the first term)
If we substitute Equation [4] and [5] into Equation [1] for the value 60, we have a function that tells us how often we will hit a land drop based on the number of cards drawn, lands in deck, and if we are running SV or Remand. Awesome!
TL;DR
So what does this mean? We can input some additional information to compare Decks A and B:
Deck A: contains 22 lands, 3 serum visions, 4 remands
Deck B: contains 24 lands, 0 serum visions, 4 remands
We already know that our model is only useful after turn 2, so let's look at how often both decks can hit their third land drop by turn 3, on the play:
Deck A: 80% (as opposed to 72% without SV and Remand)
Deck B: 83% (as opposed to 79% without Remand)
How about the fourth land drop by turn 4, on the play:
Deck A: 65% (as opposed to 54% without SV and Remand)
Deck B: 69% (as opposed to 63% without Remand)
All of these percentages go up if you are on the draw.
WayTL;DR
Running 3 Serum Visions increases your percent chance of hitting land drops 3-4 on time by a more than 4-5%. A full playset of remand increases your chances by about 4%. Each land increases this percentage by 3-4%. This analysis was strictly based on how these cards affect land drops, they obviously have their own very useful abilities. What we did see is that if you are running 4 cantrips, they will have the same effect on land drops as a single land would. This is why delver-style decks can get away with running a fewer land than necessary. Any less cantrips than that shouldn't be considered as having a significant effect.
For those interested, I have the calculator programmed in Mathematica. I can look into making some sort of web based form of it.
I want to agree with you on these points, but I think 22 lands is too few for one important reason: hitting land drops on time. At 22 lands, you statistically won't be hitting your 4th land drop until turn 5 or later. This is true with or without Ghost Quarters. It also seems like 22 lands and 4 Celestial Collonades are mutually exculsive. You will have that slow starts more often, and you are less likely to get to 6 mana in the course of a game.
This deck should be running at least 23-24 lands to make sure we hit our lands drops on time, even if that leads to flooding out sometimes. I personally don't mind flooding out since it usually means that Celestial Collonade will come online faster, and that you can make stronger plays with a Snapcaster in hand (i.e. flashingback a Cryptic or a Sphinx's Rev).
I have a land drop curve generator at home that I can post later for anyone that's interested.
Thanks! If you use Thing in the Ice mainboard, you should also swap the Eiganjo Castle for another Collonade or fetch as to not tip off your opponent as to what's on your sideboard. I wouldn't recommend cutting Sphinx's Revelation since can be a very strong finisher, but that is a pretty minor change.
You should just use the Archangel Avacyn you pulled in place of Ajani Vengeant. I will also be making that swap if I happen to pull her.
Because TiTi can win games after it sweeps the board. It fixes the one problem our deck has when playing the control side - closing out games. I tested with this card all weekend, and it makes the aggro matchup closer to 70% when played with 3-4 in the deck. It is a bad card when there is plenty of spot removal, but that's why it is a great compliment to Geist.
I'm not saying to not run Timely Reinforcements and Anger of the Gods, they are both great cards. But, try running 1x Anger, 1x Timely, and 3x Titi in your sideboard and you will not be dissapointed.
My list is rather threat-light, running only 10 creatures mainboard, so it might not be a great option for people running a higher amount. I'm using a few sideboard slots that would have been taken up by sweepers for the Thing.
Grim Lavamancer is an awesome idea. I like it as a 2 of in the sideboard, since he isn't relevant in every match-up.
You should reconsider Indulgent Aristocrat as a 2-3 of on the mainboard. He doesn't seem like much at first, but has single-handedly won a few matches for me. He allows us to race other tribal decks in size of our creatures because he has the same awesome synergy with Bloodghast that Kalitas, Traitor of Ghet does. At worst early game, he eats a removal spell without huge loss in tempo. At worst late game, he eats a removal spell and we sac him in response, giving all creatures +1/+1 counters. A 1-drop lord that can impact the board right away is pretty neat.
I think you are right that we don't want to be playing an Aristocrats deck with Vampires in it. @ThatOneNinja has a really good analysis of how to build this deck a page back - we should cut all of the gimmicky synergies for raw power. But, I am not sure if that raw power will take form in a tempo deck right now. The power density for Vampires currently loaded in the 3 and 4 drop slots, pushing us towards midrange. The best version of this deck will obviously not be as fast as Merfolk or Elves, but it won't be as slow as Jund either. We're probably going to land somewhere right in the middle.
I think you've listed our best three turn one plays. Thoughtseize, Inquisition of Kozilek, and Faithless Looting are the best cards we have available to our colors. We definitely want to loot turn 1 if we have a Bloodghast or two in hand.
Stromkirk Noble suffers the same fate as Geist of Saint Traft - he is amazing when the board is wide open, but can be easily stonewalled. I personally like Indulgent Aristocrat in this slot.
I agree that only testing can determine the best build. A few of us are doing our testing on Cockatrice or XMage for the decks, you should join in.