Saturday, January 29, 2011

Vs. Non-Aggro -> Play Aggro?

The other option, when faced with people who you can't control very well, and who have pretty good answers to your primary plan (Slime, Nature's Claim, Primeval Titan, Valakut, OR Negate, Cancel, Celestial Purge, Revoke Existance), is to go with a secondary plan.

The secondary plan being to kill them before they get going.

In at least one of the games I'd be going first, and counters are a pretty bad answer to a resolved threat.  Also counters to non-creatures(Spell Pierce, Negate, Dispel) are a pretty bad answer to creatures.

Unfortunately, as far as burn goes, I am already using quite a bit of the relevant burn.  Anything I might bring in at this point is of the "2 damage for 1 mana" variety, which is really not very good at killing people.

Burst Lightning is okay against control, because you do it on the end of their turn with a full hand, and either way you're probably good to go.  But cards like Galvanic Blast, Punishing Fire, Forked Bolt, and the like are pretty uninspiring.  Lava Axe as a sorcery doesn't accomplish much.


Artifacts, like Sphynx-Bone Wand tend to fall victim to similar removal that is already being brought in against the PA.  Acidic Slime, Nature's Claim, Negate, Revoke Existance, Naturalize, and so on, all also kill SBW.  Some cards, like Kor Firewalker, War Priest of Thune, Flashfreeze and Dispel are avoided, but the 7cc of this monstrosity makes it prone to things we can normally play around, like Mana Leak and Spell Pierce.

This is a pretty sweet artifact though, and it falls into the same vein as the deck in general.  Cards like Steady Progress, Staggershock, Surreal Memoir, and draw spells work quite well with this.  It just costs seven freaking mana.

So, it falls to creatures to try and answer this.

The most obvious is Kiln Fiend.  It comes down early, gets a massive 3 power per spell, and kills quite quickly.  It is vulnerable to some of the same removal as PA (Ratchet Bomb, Flashfreeze, Celestial Purge), but not alot, especially if I am going first.  The turn 4 PT can't kill the KF without a Khalni Heart Expedition, so I might have that turn to kill them, with a bounce on the KHE or PT.

The next best options to compliment that are Chandra's Spitfire, Goblin Guide, Plated Geopede and Calcite Snapper.
The 3cc options are pretty slow (slow to come down -> More counterable, slow to kill).
Geopede and guide do damage fairly fast, but not fast enough to kill on their own without some serious help from specific spells/lands (and no blockers)

Kiln Fiend + Goblin Guide put up some pretty respectable T4 kill numbers with the old baseline deck (Mana Leak still in).   With an even more dedicated plan, the numbers might get even better.
(+4 KF, +4 GG/Spitfire, +4 Burst Lightning, +3Twisted Image, -4 PA, -4ITR, -4 Deprive, -3CTM or CTM).  Screwing around with that will give me something to do on the bus ride to work today...

Vs. Non-Aggro -> Play Control?

I have been trying to take the control role against control decks, and ramp decks, in games 2-3.

In game 1, control's creature removal is dead, and their draw (Wall of Omens, Spreading Seas) are often sorcery-speed cyclers.  Their removal for my win condition is near-zero.  This often means a win game 1 against control, because in many ways I'm a better control deck game 1.

Ramp doesn't play pure creature-removal, or bad card-draw, and is fairly non-interactive on it's way to killing me with a Titan or Trap.  (problems I don't have good main-deck answers to)  This often leads to a G1 loss, since I don't have good answers, and they don't worry about having answers.  I am not a very good control deck game 1 (Titan->Trap ->Loss, Titan -> no counter -> loss, so many times 2x threats = loss).  The only way I win game 1 is to "aggro" them out with a turn 4-5 kill off an early PA + charge with staggershock, etc.

The control matchup feels like more of a coin-flip, once they bring in Purge/Memoricide for dead removal, and I bring in counters.  It is pretty easy to lose to discard -> Grave Titan, or "activate collonade on 12 land".

The ramp matchup feels nigh unwinnable with the current strat.  I can counter all I want, and they can still just play valakuts and kill me.  (especially when my clock goes infinite at 4x staggershock as my offense)

So if I am going to go control against these control + ramp decks, I need a answer to their non-counterable threats (MBT + Tectonic Edge), a way to deal with their removal for my threat post-side (Nature's Claim, Celestial Purge, Discard -> Memoricide, Ratchet Bomb)

Sideboard consequences:
I typically take out Burst, Bolt, and bring in 8x counter.
ITR is not awesome against most ramp.  ITR is probably manditory in some quantity against control, because it is defensive against removal for the PA, and offensive against answers like Leyline.

Having a planeswalker (Jace B.) to grant recurring card advantage and/or kill opposing Jaces could be necessary, in place of something?

I have a ton of sideboard slots I can dedicate to these match-ups, but I'm not sure what to take out after the initial 8.

Vs. Aggro = Comfortable in the control role.

I know what I want to do against aggro.

I remove their early threats.  Get card advantage with my 2:1 burn spells and Treasure Hunt.  Go 1:1, or get mana/quality advantage with my filter/cheap spells.  Eventually land a PA, and burn them out.

Recurring 2:1's with Call to Mind, preventing big swings with ITR, countering hard to deal with spells with Deprive (and potentially MBT post-board).

I feel very comfortable in the control role here.  There's no way for me to out-aggro Memnite, Goblin Guide, Lacerator, etc, and I have good answers to them as part of my main plan.

Maindeck/Sideboard consequences:

Playing Arc Trail main (which seems as good or better than Burst Lightning in any aggro matchup) means I don't need any sideboard slots for this matchup.  The Arc Trails are the first thing to come out vs control/ramp, so having a "more generalized" Arc Trail maindeck, and siding into Arc Trail, or siding Arc Trail into Pyroclasm for aggro matchups seems like a waste of 4 slots.

So Pyroclasm really feels like it can/should be benched.  It is too reactive, and is no more of a 2:1 than Staggershock or Arc Trail.  The scenarios where it is a full on blowout are few and far between, and it can never go to the head.

There may be some argument for putting the new 3-damage Pyroclasm (Slagstorm) in the board, especially if the green leyline, or more token-based strategies become prevalent.  The 3 damage vs 2 is not terribly relevant, and the 1RR vs 1R casting cost does not favor Slagstorm very well.  However, the 3 damage to the head option makes it look pretty good main deck.  It also doubles to kill titans, and is a non-targetted way to kill some bigger things (especially doubled), like the Trool, Gaea's Revenge, Sphinx, Frost Titan, and so on.  But we'll talk about that later.

So, in general I feel pretty good about the aggro matchup.  They can put incredible pressure on early, but I can't think of a single deck where they have no top-end threat, or no recursive threat, or where ITR is bad, where I would be willing to take out ITR + counterspell, for 4 creature sweepers.  So there is no room to play all the red spells, so the only thing to do with sideboard slots for this matchup is to upgrade existing creature kill with better creature kill.

/shrugs to that.

Monday, January 24, 2011

MTGO Metagame

http://www.mtgstats.com/

As of today, Valakut, UWControl, UBControl are 50% of the field.

I feel very good about my Main-deck (4x Arc Trail, 4-x some-Counter, 4x Staggershock) beating aggro, and beating control game one.

I need a sideboard strategy that beats control post-board (I am 1-2 twice with the new decklist, winning first, losing next two), and a sideboard strategy that can beat ramp from my main-deck.

I can dedicate something like 15 sideboard cards to these two matchups.  I need the SB games to be crushing for ramp, and control.

Threats from control are:
Cellestial Collonade -> Tech Edge/Demolish/Spreading Seas
Creeping Tar Pit -> Tech Edge/Demolish/Spreading Seas
Jace the Mind Sculptor -> Jace Beleren/Burn/Sphynx
Grave Titan -> ?? (I did beat 3 GT's in one game, but that was a rarity)
Other Planeswalkers (Elspeth Tirel, Gideon Jura) -> ??

Threats from Ramp:
Primeval Titan
Summoning Trap
Gaea's Revenge/Terrastodon/Avenger of Zendikar
Valakut, the Molten Pinnacle

Saturday, January 22, 2011

Treasure Hunt + Halimar Depths = Good?

I've done some math on Halimar Depths + Treasure Hunt, and these were the results:

Halimar Depths, chance of revealing n land, using my deck as an average layout.(*)
0: 20.8%
1: 44.2%
2: 29.0%
3: 5.9%

So, in the 0 case, we hold our Treasure Hunt, or time it to draw the good card at a irrelevant time.
In the 1 case, we get a draw two for sure, and can set it up to draw the land + non-revealing card this turn, next turn, or go for the ancestral recall two turns from now.
In the 2 case, I typically put the non-land on top, then cast TH the next turn for draw 3+.
In the 3 case, either you draw 4+ that turn, or you draw 3+ the next turn.

TH is exceptional in the 35% case where you draw 3 or more. (or 4 or more, depending).

The 1 land case takes up a huge chunk of the probability matrix, and gives you the most flexibility in your options.  If you need the land and can draw now, you can do that, or you can set up the draw-2 (reveal less dangerous) next turn, or you can go for the gusto.

The 0 land hand (21%) is the one where TH is the "worst", but your current situation is probably the best. (Drawing 3 spells)

Now, lets look at some basic See Beyond(SB) math.
Number of lands in the top two cards:
0: 35.5%
1: 48.8%
2: 15.6%

Treasure Hunt is exceptional in the 0 case, which is a similar 35% that you'd see for TH above.
TH is "comparable" in worth to the 1 land on top case, in that you have to wait longer for SB to give you the 2 cards, in roughly the same percentage of the time.
 
What is our scenario in the bad case for See Beyond though.  If SB is bad, we draw two land, shuffle the land we had in hand into the deck, and still have two land left over.  This puts us in topdeck mode.

In the "bad" case for TH, we're drawing three useful spells without even casting anything, so we just don't cast our TH for a while, and get more use out of it later.

So, TH seems clearly superior in every case where we have a Halimar Depths anywhere in our future.  (Thank you Deprive)

===========================

What about if we don't have a Halimar Depths?  What is our expected return off a blind TH?

Well, Treasure Hunt draws us one action card always, but the trick is how many land do we get.
This comes down to the chance that the top card is a land.  This is about 40% with unknown card distribution.

But we also have the chance that if the top card is a land, that the next card is also a land (roughly 40% * 40%), or another 15.6%.  Then we have the chance that the next card is also a land (another 6%), and the next card is also a land (2%), and so on.

This adds up to a little over 1.6 cards expected per Treasure Hunt.

Or, one "action" card, and .6 lands.

How does this compare to a "blind" See Beyond?
From the above chart:
Number of lands in the top two cards:
0: 35.5%
1: 48.8%
2: 15.6%
Presuming we shuffle the land back in, and don't have another land to shuffle in, this gets us an average of:
15.6%: 1 land
48.8%: 1 card
35.5%: 2 cards - our worst card at the time. (worth half a card lets say)

This puts our average draw at .16 lands + 1.02 cards  This seems inferior to the blind TH, with the only advantage being 2% more cards, and hiding what cards you drew.

If we presume that we do have a 0-worth land to throw away, then we get:
15.6%: 1 land
48.8%: 1 card
35.5%: 2 cards
This puts our overall result at 1.55 cards + .16 lands -.355 land
This translates to a conversion of ~.5 of a land, into .55 of a card.  Which is pretty good.  Depending on the game state, we may take that trade.  Early game, probably not, late game, probably.

==========================

Secondary effects of both spells:
Treasure Hunt reveals your draws.
This is awkward in a deck with counterspells, but not fatal.  It is worst when you draw a counterspell and it is your only card (~11% of the cards in the deck are counterspells), and have either no mana to play that card, or it is the only card in your hand, and you can't recover from the board position with a counterspell.

As a mitigating factor, See Beyond also can't recover by drawing a counterspell in the second situation, they can just hide that they drew it.

Very high level players can probably get some good usage out of this information, but the average player typically doesn't foresee what is going to happen with that card (bouncing own PA in response to removal for example), or it doesn't matter, because so many of the cards in the deck do the same thing.  (CTM especially!)

If I reveal a Mana Leak (which is no longer in the deck), that is easily played around.  If I reveal a Deprive, they may presume I also have Mana Leak (which any sane player would be playing now?), which means I get some good value despite what is normally the worst-case scenario for Treasure Hunt.

See Beyond shuffles your deck.
Normally, you'll have ~1 bad card scried to the bottom, and the one bad card in your hand you are shuffling back in, but you can still draw these cards.  You have a continual ~4% chance each turn to draw those cards that you'd "dealt with" each turn.

Not ginormous, because you are eventually going to have to shuffle your deck anyway, with Scalding Tarn, just to get land out to play spells, but it is there.

The more Foresees you cast, the more Preordains, etc, the more important this factor gets.  (When playing Foresee, I have scried ~8-10 cards to the bottom without clearly winning.  It sure sucks to shuffle those back in.


=========================

Overall Conclusions:

So, the Blind TH, vs the no-land SB seems to be a win for TH.
The Blind TH, vs the one-land SB leans slightly towards the SB (but with slightly reduced card quality going forward)
The HD + TH scenario clearly beats the SB scenario. early, late, and mid, because you end up with more cards in hand (that you can't waste draws on later), and your mana development isn't hindered.


If you think about it, the worst TH is not much worse than the average SB scenario.  66% of the time, you'll do no better than drawing one action card with SB, and some of the remainder, you will just be upgrading one card to another.  Treasure Hunt always nets you even in cards, and guarantees a minimum of one good card, which See Beyond can't promise.


See Beyond and TH are not too far apart, there are some big overlaps where they do the same thing, and where one just has a slight edge over the other.  I would just say that SB has more variance, which is probably not that useful.  So, I'm sticking with my 4x TH + 4x HD, and trying out 4x Deprive to compliment them.


=========================
(*)Average layout...
We presume that the deck starts at a ratio similar to 24 land, 60 total cards when we start looking at cards off the top of our deck.  In reality, this will never be the ratio, but we have to start somewhere, and we can't draw general conclusions about how many land or non-land we'll have seen.  This is a known problem in this type of analysis, and it probably can't be helped.

Decklist 1/22/2011

 4x:





7x Mountain
9x Island

Sideboard:

3x
Brought in vs.:
Vengevine (better than Deprive)
Ramp (Eldrazi G, Valakut)
Vampires (Bloodghast - dodges Inquisition, occasional trap cost, better than deprive)
Control - Hard counter. (better than burn)









4x
Brought in vs.:
Ramp (for burn).  Man, this seems terrible...
Mono Red, but probably not (over deprive?  that seems bad)
Elves (but probably not)

 Can't I do better than this?








2x
Great against control.  Kills planeswalkers good, good after a T6-8 counter war.  Dodges their answers (Flashfreeze, Ench removal, Planeswalkers, Negate, Spell Pierce, Dispel).  There is no answer to both Sphynx and Ascension.











 4x
 Great against aggro.
 Since I'm taking Burst out against everything but aggro, and probably taking Burst out against aggro for Arc Trail, why am I not main-decking this?












 2x
 Bringing this in vs. Valakut, and some control matchups.
 Killing manlands is awesome.
 Best answer to Ratchet Bomb.
 Best answer to Valakut (other than Spread?)
 Good against equipment.











(*) Cards that I am not entirely sure about.
Burst could be Arc Trail, or Mana Leak.
Call to Mind could be Foresee
Deprive could be Mana Leak (and most sane people would have that)

Flashfreeze + Demolish could be Spreading Seas, since it does similar things.  This is worth discussing.  I really don't like the fFashfreeze side.

(**)Some notable people have said Treasure Hunt is a terrible card. (GerryT, some people who have talked to GerryT)

I need to do a post on that, as well as some math, because TH really feels good, and compliments Deprive, and compliments HD, which are both pretty good cards on their own.

See Beyond Anti-compliments Preordain, Foresee, Halimar Depths.

I can't ignore the tested comments of some strong notable people, but I also can't ignore my findings, so that feels like it deserves a post of it's own.

Considering Some Maindeck Changes - Feedback Appreciated.

 
2 Call To Mind + 2 Foresee vs. 4 Call To Mind.
Argument for Foresee:
=Foresee is my best topdeck late game.  It lets me look at 6 cards this turn in a mad hunt for a particular card (often PA), which is probably 15% of my deck at that time.=If you don't have a PA, this hopefully finds one for you, and then you cast it + do things next turn.

 =Foresee is basically a doubled Preordain.  One card, for 4x the mana, and 2x the effect, a pretty awesome deal.  It would be like paying 4 mana for 6 damage, a spell that doesn't exist.  Lava Axe is 5 mana for 5 player-only damage, and there isn't a spell that does more than 4 general damage, that does more damage than it's mana cost (outside aoe).

 =Foresee is always straight-up card advantage.  It costs one card, and draws you two.

 =Foresee is often much much more than straight-up card advantage.  In the long run, it is similar to drawing about 4 cards in a conventional manner. When you consider that lands are often 100% dead when you cast Foresee (your most expensive spell), any land you don't draw is effectively a card you didn't have to draw for 0 upside.  The average Foresee of bottom two lands, draw two cards, is a draw 4.  The "good" Foresee of bottoming 3-4 land, or bad spells, is a draw 5-6.  The "bad" Foresee of putting one land on the bottom and drawing two business cards, with another business card on top, is a draw 3, and that is about as bad as Foresee gets.

 Argument against Foresee
=4 mana is alot to pay, and it tends to tap you out, so you are hoping to do things next turn.
This is alot less actually relevant if I have 0 counterspells in my deck, but it still maintains the percieved relevance to the opponent, and lack of bluff-ability.

 =Casting a Call to Mind for a Preordain, and then NOT casting the Preordain with 2 mana up is a extremely strong counter-bluff.  Casting a Foresee with 6 mana up is as good a counter bluff as Foresee can do.

 =If you do have a PA, Foresee likely doesn't put any counters on it, but may set you up for putting counters on it next turn.

 =4x Call to Mind makes charging PA off CTM fairly trivial.  CTM for CTM, CTM for anything, PA charged - for only a cost of 6 mana (or 2x 3 mana), and no card disadvantage.

 =2x CTM means you have to be very careful using your CTM mid-late-game.  Your second CTM MUST target your first CTM, or you will never get any more. (which may be good enough for a kill-play, but probably not for common usage)

 =CTM for a Preordain is a mini-Foresee. (with a similar casting cost).  You can turn a CTM into a mini-foresee, you can't turn a Foresee into a mini-CTM.  I'd almost certainly rather HD + CTM + TH, than Foresee, despite the Foresee probably drawing me one more card, and being easier to set up.

 =CTM likely just straight up gets you what you want (as long as it is not a PA), for one mana less.

 =Having 4x one-spell is better for charging PA than having 2x 2 different spells.  (more redundancy), and CTM is a card like Staggershock which can be a one-card full-charge of a new PA.

 What does "the community" think?
The community is probably using 4x Foresee, and 2-3x Call to Mind, and losing to aggro (or using SB slots for Pyroclasm, and still losing to aggro).  It takes away from the Staggershock slot (which is great against all matchups, and for charging PA)

 Bleh.

I can't imaging wanting to resolve more than two Foresees (or one doubled Foresee) ever in a game, so having 4 is just terrible.  The second Foresee is pretty much always either overkill, a (double)mulligan, non-castable, or too slow to actually cast.  (One might argue that the first Foresee is also too slow)
The first Foresee is pretty good, so I can see going to 3, but 4 is just way too much.  (Especially with the 23 land the community plays)

 Super-late game, any draw spell is great, with Foresee being the best, and Treasure Hunt probably being the worst (since it can draw you 1x counterspell/ITR/burn spell, and that's it).  Call to Mind gets you your second best one back, or can do things like charge up a PA or just kill the opponent.  Doubled Foresee is near-guaranteed kill with a active PA, double CTM is also near-guaranteed. 
I'm not sure where to go with this.  It's pretty close, and there are definate advantages of 4x CTM.  I guess I'll just think about what the game state would be if I had a CTM instead whenever I draw a Foresee (and record it)

















Counterspell(s) vs. Arc Trail
Whatever the winner of the below counterspell-off, it still might be better to just play Arc Trail main.
This is more about philosophy than capability, to some extent.  They are two different outlooks on how the deck should win.
Counterspells try to slow things down a bit, draw the game out, and eventually cause a win via doubled spells.
Arc Trail functions as main-deckable Pyroclasms, which are incredible against some decks, while still getting damage in to the opponent's head or planeswalkers.
I have never seen a blue player +2 their little jace, so Arc Trail is just as good as Burst Lightning, Lightning Bolt, or Staggershock at killing that cloaked fiend.  Arc Trail kills off a ton of aggro starts.

 1 Toughness guys:
Lotus Cobra
Birds of Paradise, Llanowar Elf, Sylvan Ranger, Arbor Elf
Unleveled Joraga Treespeaker
Pulse Tracker, Bloodghast, Viscera Seer (Can't sac to counter the spell), Vampire Hexmage
Steppe Lynx, Plated Geopede
Kor Duelist, Elite Vanguard
Squadron Hawk, MemniteBlight Mamba, Ichor Rats, Ichorclaw Myr, Necropede
Spikeshot Elder, Goblin Bushwhacker, Goblin Ruinblaster, Kuldotha Rebirth tokens
Herebaz Druid, some other terrible allies.
Liliana's Spectre, Nantuko Shade
Essense Warden, Lone Missionary
Perilous Myr

2 Toughness guys:
Every other creature in a R, W, or B aggro deck.  About 50% of the rest of the playable Green creatures (Not Vengevine, Wolfbriar Elemental, Overgrown Battlement, Titans, etc)

 Notably lacking:
It's pretty fair to say that neither Staggershock, nor Arc Trail help kill resolved Titans of any stripe.  4 damage is unlikely to happen in conjunction with these 2 damage bursts, to take these bigger creatures down.
If there aren't any creatures in play, I can always 1 myself, to 2 them, which is not amazing, but I imagine it will be a oft-used option.

Arc Trail is like Staggershock 5-8 in creature annihilation, and staggershock is pretty awesome for creature annihilation despite the downsides it has over arc trail of:

3cc

Takes a turn to kill the 2nd one.

If the first copy of SS is countered on resolution, by the creature already being dead, sacrificed or whatnot, the second copy will not happen.

Arc Trail is a Sorcery instead of an instant, but it may save around 6 life against aggro decks compared to staggershock.  (Kill two creatures T2, which can't attack me on their turn, and one of which can't attack me the next turn).  This is especially relevant on the draw.

A previously mono-colored aggro deck (Vampires) in a color which should have the "best" removal (black), is splashing red in order to cast Arc Trail, Lightning Bolt, and have a non-tribal manland in their otherwise ultra-tribal, mono-Black-no-colorless-casting-cost deck.   This is a testament to the power of Arc Trail (and Lightning Bolt) as strong, cheap removal.

 Llanowar Elf into Elvish Archdruid(or anything else in their deck)
Pulse Tracker (or Viscera Seer (or Vampire Lacerator likely)) into anything
Steppe Lynx (or Spikeshot Elder) into anything
Lotus Cobra into Oracle of Mul'daya (or little Jace -1, or Cobra #2)
Joraga Treespeaker into Growth Spasm
Goblin Guide into anything not Goblin Guide

All of these starts are the starts that threaten death, and soon.  All of them are completely shut down by this 2 mana sorcery.  The only starts where Pyroclasm is incredibly better are Rebirth into Bushwhaker from goblins, or 3x 1-drop from vampires.  And Pyroclasm can't kill them, and can't kill Planeswalkers.
Anyway, Arc Trail seems like it's a good, pro-active spell, that shuts down aggro just as well as Mana Leak (because it does it after the fact, and handles 1 creature  better, because of incidental damage to them), and is better than Pyroclasm against non-aggro.

I am leaning towards swapping this in for whatever counterspell main-deck.  (along with a few mountains for a few islands)  I could also do something like swap it in for Burst Lightning, if I wanted to add it, and keep the counterspells.

Going counterless puts alot of pressure on ITR as a semi-answer.  For some problems, it's an acceptable temporary answer.  For others, only an active PA with CTM will move it into more of a continual solution.  (EoT Bounce whatever, my turn, they die is a common setup)
=============

Alternatively, we could find that Arc Trail is good, and still keep in (whatever) counterspell, and just replace Burst Lightning instead.

Burst often does 2 damage, but it has the advantage of doing it at instant speed, which matters for:
Joraga Treespeaker (ties up their turn 2) (And killing the T1 JT)
Cellestial Collonade (4 toughness, "immune" to sorcery)
Bloodghast (Killing in response to land play + haste on their turn)
Goblin Guide (Getting the attack-proc and still killing it)  (And killing the T1 GG)
Goblin Bushwhacker (prevents 2 damage, which the Arc Trail will probably prevent more by killing another goblin)

Arc Trail does cost 1 more mana, so it is a bit harder to put together that finishing turn, combining (multiple) Call to Mind(s), and copied DD.  Staggershock is an instant, so it isn't horrific to have to cast the Arc Trail on your own turn.  It's definately better than main-decking Pyroclasm, because it is never dead.

Either one is sided out vs. non-aggro, so it makes some amount of sense to play the one that is better vs aggro, rather than the one that is better in general, and just not have the sideboard aggro slots.




=====================

Alternatively alternatively, I recently read some oldschool PA articles, that said Burst Lightning was better than Lightning Bolt, because there were more 4 toughness targets, than three toughness targets.  (aka all targets are 2 or 4).

It is true that there aren't a ton of 3 toughness targets, but there are some:
Steppe Lynx, Plated Geopede on their turn (even + Fetchland).  Pretty much any scenario where you are going to target a PG or SL as a sorcery with BL, Arc Trail is better (and Arc Trail is just better in these scenarios anyway)
Elvish Archdruid + Leyline, which is huge for allowing your spells to not be dead.
Tempered Steel + Anything (not likely)
Vengevine (not awesome)
Bloodthrone Vampire + a creature (moving out of favor lately)
Big Jace + Brainstorm (happens all the time)

The big deal is when it comes down to later turns, and you are either trying to kill them, trying to double-bolt to kill a titan, or combo them out with Call to Mind (when you don't have a ton of mana).  DoubledBolt-DoubledCTM-DoubledBolt is 12 damage, plus whatever else you CTM'ed, that can easily lethal someone out.  Do the same thing with Burst Lightning, and it is 8 damage (or 8+4 = 12 if you have a turn to kill?)

I can't immediately dismiss this, but it doesn't seem like a awesome idea now.






Mana Leak vs. Deprive vs. Mindbreak Trap

All of these:
++Counter any type of spell
It is important/critical to be able to counter:
Creatures: Primeval Titan, Grave Titan, Frost Titan, Sun Titan, Acidic Slime, Kor Firewalker, War Priest of Thune, Stoneforge Mystic.
Non-Creatures: Jace, the Mind Sculptor, Koth, of the Hammer, Elspeth Tirel, Memoricide, Mind Sludge, Dark Tutelage, Ratchet Bomb, Leyline of Sanctity, Luminarch's Ascension, Sword of Body and Mind, Disenchant-type effects.

 --Are pretty bad at charging PA.

 --They can't be cast at will

 --They don't draw cards, or kill your opponent.

 --They don't deal with already-resolved cards (without an Into The Roil)
 --Are bad against Ricochet Trap (even doubled)   

Mana Leak:

++Cheapest - castable on turn 2

 ++Easiest to cast (Least U).  Only have to leave 2 mana up (1U) so you can cast more sorceries. (Which tend to have U-costs)

 ++Can still be relevant with a active PA

 ++Some people don't play around it, especially doubled. (moreso in TPR?)

 --Can become less relevant late-game.

 --Easiest to play around

 --Can be terribad in some matchups (control, ramp), due to ease of playing around, and slowness of the matchup in general (when PA is trying to play "control")

 --Mana Leak "forces" you to counter some spells you might not, because them resolving makes Mana Leak irrelevant (2nd Overgrown Battlement, Harrow, KHE, Everflowing Chalice)
 Deprive:++Cheap - 2cc.

 ++Hard counter, at every game stage.

 ++Best(?) late-game counter, because it lets you reuse Halimar Depths, hopefully combined with a Treasure Hunt, or uncracked/unplayed Scalding Tarn, which can easily set up your next turn.

 ++Acceptable in counter-wars.  
They will often cast a important spell with 3 extra mana (to pay for your Mana Leak), while holding some counterspell, often Mana Leak.  If they are casting a Jace, this puts them at 7-8 lands before they do anything.  You only need to be at 5 untapped land to Deprive, and pay for their Mana Leak.  This means you could have cast a preordain, or played a Halimar Depths, or maybe even cast a Treasure Hunt or PA that turn.  Sometimes it is a Negate or Cancel they are holding instead of a Mana Leak, but no counter works there.  (And Deprive leaves up land to cast another counter, and lets you reuse HD even if they re-counter.

 ++People often play around Mana Leak, which gives you more turns to draw your Deprive, especially in slow matchups.

 ++With an active PA, you get two hard-counters, while only having to pay the pick-up-land cost once.

 ++If you are playing counterspells vs. aggro, you are often in a topdeck situation, and Mana Leaks are terrible in this scenario.  Deprives are perhaps the best counter in this scenario.

 00There are some potential bluffing advantages for (having to) leave up UR (and not UU), and having them either cast a important, or non-important spell, and you "not having the leak". (while still holding a counter).  They'll likely be convinced that you don't have Mana Leak (the only counter that a sane person would play), and get blown out with a counter on their truly important spell later.  But that still means you couldn't counter their semi-important (or really important) spell with your UR.

 00If you don't have another land anyway, Deprive lets you "use" your land drop, without giving away information about your non-land status to your opponent. (Making draw spells like TH and Preordain more likely to resolve)

 --Color Cost - UU is not-great.  Many of the sorceries are U-based, and trying to cast double-Deprive turn 4 is never going to happen.

 --Untapped UU turn 2 is hard to pull off.  It pushes your Halimar Depths to turn 1, which makes your Treasure Hunts worse.  (You can't use any of that HD knowledge to cast TH, and also leave up Deprive, while the cards are still on top with a T1 HD)

 --Tempo.  Picking up a land T2-4 is almost never going to be good.  If it's a HD, it MIGHT be neutral, but it will very rarely be a net + for you.  It is also pretty hard for it to be very useful vs aggro early.   You can always throw away a Mana Leak as a pseudo-burn spell T2 against an aggro deck, but you can't do that with a Deprive (as easily, or with as few consequences)
Mindbreak Trap:

 ++Counters non-"counter"able things, which are often disasterous to this deck when playing as control.
Bloodghast (Landfall - return to play from graveyard.  Haste if you are <=10 life.)
Vengevine (Return to play on second creature CAST in a turn.  Haste.)
Kuldotha Phoenix (Return to play on upkeep for (4) with metalcraft.  Haste.)
Emrakul, the Aeons Torn (Uncounterable - you lose, take an extra turn (~Haste).)
Gaea's Revenge (Uncounterable.  Haste (w/8 power, and non-targetability).)

 ++Can occasionally be paid for free, but almost never on your turn.
In counter wars, when they cast a spell, you counter, then they mana leak twice or something (Often requires you to have 2 counters anyway)
When they cast multiple creatures in the same turn (Vampires, WWQuest)
RUG: Preordain, Preordain/Explore -> some spell is not uncommon.  Sometimes that spell is even a Titan!
UB/BUG: Preordain -> Discard (that may not hit MBT) -> Grave Titan/Jace will be held off by the 0cc.

 ++Occasionally counters more than one spell.  PA Mirror
Long counter-war (Often requires multiple counters from you)
Overeager nuke-caster, or Trap-caster.  (They can always let instants resolve before casting the next one)

 ++Not discardable by Inquisition of Kozilek.  UB control often tries to strip your hand of permission with discard, before they cast their game-winning threat (Grave Titan, Jace, Memoricide).  They often play 4 Inquisition, and only 1 Duress (which still gets MBT).  Being able to keep your permission through 80% of their discard is good, you just need to be able to cast the MBT once they go for their game-winning threat.

 ++I did sideboard in my MBTs vs Vampires (Bloodghast) and Fauna Shaman (Vengevine) over Mana Leak.  Pyroclasm-like effects are okay, but not amazing against recursive threats.

 00Requires a combination of ITR to deal with many of these cheaper spells, but then deals with them for good.00 Average late-game.  Better than Mana Leak, worse than Deprive.

 --Most expensive, by 2x.
Hard to leave up while casting other (draw) sorceries.
Need 7 mana to counter a spell, and pay for the mana leak backup. (unlikely) (see deprive ++ scenario)
Many of the threats described in the general section above are 4 or less casting cost, or accelerated out on turn 4, which makes trying to counter any of them with Mindbreak Trap awkward at best.

 --Most color-intense (UU2 is no better than UU)

===============
Overall:

Mana Leak is best (very) early, and against impatient people.  It is the best spell to REPRESENT having, because it causes your opponent to give you more turns.  (A pseudo time walk, that you don't even have to cast?)  Mana Leak benefits quick-game play.  You can do stuff, and still slow people down, or represent slowing them down, and then they're dead before it matters that it isn't a hard counter.

Deprive is best in the mid-late game, where Mana Leak won't counter the spells, or the spells you want to counter are cheap.  Reusing Halimar Depths when you don't need the extra mana (often) anyway can be pure upside.  Deprive benefits patient long-game play.

 Mindbreak Trap is (only) good late, and plugs a few holes the deck can have, against specific cards.  (Most often when combined with ITR)  Generally, the more controlling an outlook (and the later we expect the game to end) the more necessary having some number of MBTs is.  Killing a ramp deck on turn 4 means you don't have to worry about their Gaea's Revenges.  Killing a Vampires deck on turn 5 means you don't have to worry about Bloodghast recursion.It seems like it's probably best to main-deck Deprive, represent Mana Leak, and have MBT in the sideboard. (For the specific matches where it is good).  Mana Leak vs. Deprive is tricksy business.

==============

Edit: I'm pretty sure I'm playing the Deprives main.  Valakut plan is still sketchy.  4xDeprive/4xFlashfreeze/3xMindbreak Trap/2xDemolish is kinda over the top, taking out 4xArcTrail/BurstLightning, 4x bolt, 1x Call To Mind?

Man, that just seems terrible.  Mycolith King Dome for a winning plan vs Valakut...Apparently editing this post totally broke the formatting of it.  Awesome.


I've effectively reposted this, because an edit totally destroyed all formatting irreparably.  I'll be deleting the other post.  I am not happy at all that editing destroys formatting, since I really love my editing, but copy pasting the text and re-editing works.  LAME.

Saturday, January 15, 2011

Hyper-Aggressive (Kiln Fiend) Mulliganing

7 Card Keeps:
Turn 4 wins:
Mt Mt ST - KF - LB SS - TH
ST ST - KF GG - LB - ITR ITR
Is ST - KF - LB - P - ITR ML
Mt Mt Mt Is - KF GG - TH
Mt Is Is ST - KF - SS - P
Mt - KF GG - LB - TH - ML ML
Mt Mt ST - GG - SS - P - ML (drew GG off T2 preordain)
Mt Is - KF - BL SS - ITR ML
Mt Is Is HD - KF GG - LB
Is HD ST - GG GG - BL BL


Turn 5 wins:
Mt Is Is - GG - P - ITR CTM
Mt Mt - GG - LB BL SS - P
Mt Is - KF - BL BL - ITR ML (drew another KF)
Mt Is Is HD - KF - BL - ITR(I think i just messed this one up, it should be a T4 kill)
Mt Is Is - GG - LB - TH - CTM

Turn 6 wins:
M - KF GG - P P - 2ITR

It does not look like any of these hands were particularly color-screwed, which is a problem I thought might occur with only eleven turn 1 red sources. We'll see whether this problem creeps up in the mulligan hands.

The KF hands just drew a bit badly to be five turn kills instead of four.  The GG hands tended to play out as 5 turn kills.  10 from GG plus 10 from other (with 5 turns of draw/mana) is more reasonable than 8 from GG plus 12 from other (in 4 turns of draw/mana)

If we mulliganed all the 5 + 6 turn kill hands (6), we'd expect to see similar results to the 6 card and worse mulligans, which are:
4 turn win: 4
5 turn win: 5
6+ turn win: 5

Which may or may not be worth the risk.  9/15 (60%) you're even or better.  11/15 (73%) you're even or worse.  40% you're worser, 27% you're better.  I'm not sure which of these four numbers is the one I should be basing this decision on.

The one land hands resulted in a turn 4, and a turn 6 kill.  The one that was a turn 6 looks slightly better to me at this time, but it really came down to whether they drew a land in the first two turns or not.  (T3 KF is often a T4 kill)  The T4 kill drew its land off the top, but still would have killed if it had drawn the second land t3.  The T6 kill didn't draw a land until T4. (and I got a bit impatient with burn, if I'd held it I'd have taken them to 1 life or killed on turn 5?)  The chance of not drawing a land in the top two (HD on the second turn doesn't help) is (30/53 * 33/52) = 35%, so the chance of drawing a land is 65%, which seems like it makes this a good keep.

=======================================================

6 Card Keeps:
Turn 4 wins: 
Mt HD ST - KF - SS - TH
Mt HD - KF - P TH - ML
Mt Mt ST - KF GG - LB

Turn 5 wins:
Mt Is Is ST ST - KF GG (Drew 3 land)
ST - GG GG - P - CTM
Mt Mt Is HD - GG - SS

Turn 6+ wins:
Mt Mt ST - KF - P TH  (Drew land or orange spells)

Here the one land keep didn't have a KF, so it seems like a good keep even now.  Worst case this is a T6 kill, and if I draw any burn spell in 4 draw steps it is a t5 kill.

Most of the delays in kill turn here involved keeping a hand of many land, with some action, and drawing more land.  With 3 land in hand, the odds of the first two draw steps being land are (21/53 * 20/52) = 15%.  If you have 4 land (or ST + 2 land, and agressively sac, which is what I was doing), it goes down to 14%.  Not as much of a drop-off as you'd think. 12.4% for 5 land (or 1 land + 2 ST, or 3 + ST).  You can't play around this 15% chance, but it can be discouraging.  It doesn't help that I have ~6 more cards that are just as good/bad as a land, so if you factor those in, it is more like 25% of drawing a 0-impact card in both of the first two draw steps.  That's pretty encouraging.
=======================================================

5 Card Keeps:
Turn 4 wins:
 Mt - KF GG - TH - ML

Turn 5 wins:
HD - KF KF KF - TH

Turn 6 wins:
Is Is Is - KF - SS (drew 3x KF)
Is HD ST - KF - ITR

50% turn 5 or better here, which is something.  But going to 5 is not great, because you have to include the 4 carders as well.  so going to 5 gives you
Turn 4: 1
Turn 5: 2
Turn 6+: 4
Which is only 42% turn 5 or better.

Obviously, these 5 card keeps are all pretty speculative.  They either have one land, and KF, or 3 land + KF.  Neither of those are very good times, and we see 2 additional 0-impact cards across these 25 cards.  (better than the 6/60 ratio at least?)  The turn 5 wins hit turn 2 and 3 KFs.  The turn 4 win was t2 KF (hit the land)  The T6 wins had either no action, or no land.  (emphasizing what the hand already lacked)

Still, it could be much, much worse than 42% T5 or better going to 5.
=======================================================

4 Card Keeps:
Turn 5 wins:
KF - LB - P - ML

Turn 6+ wins:
Is - SS - P TH 
KF - LB - TH - ML 

Obviously these are extremely speculative hands.  Two have no land!

I never went to three, but it is not terribly hard to imagine a turn 4 kill off Land + KF + something. (which none of these hands are)  T2 KF, (one card in hand), T3 preordain, DD spell (1 cards in hand, opponent at ~11), 2 draw + burn, or 2 burn and opponent is dead.

Probably better odds than the KF + LB + TH + ML hand had... (plan = draw two lands (25%) to get a t3 KF, then draw another land off the TH (~50%) to cast the LB (50%) and get in for 10 on turn 4.  Then draw draw spell (<20%) into burn spell (~50%) into opponent dead turn 5!  Overall, thats probably 0.5% to kill on T5 or better.  

KF + 2 Land + non-land-non-terrible  is ~(4/60 * 24/59 * 23/58 * 26/57) * 4* 3*2*1/2(order irrelevant) ~= 4%  You probably need to get lucky again on the T3-4 draw step, but a HD as one of the lands, or a preordain as the something could substantially improve your chances there.  Perhaps up to as high as 6-8% overall!  That's probably better than the KF-LB-TH-ML hand had.

The Preordain hand won on turn 6, the no land hand scooped on turn 5 with no potential damage to the opponent in sight.  The turn 5 hand having a preordain is a million times more likely to have any action at all than the Treasure Hunt hand.  Preordain gives you alot of outs, Treasure Hunt doesn't give you any more outs until you win the "two land on top" lottery.
 
=======================================================
Mulligans:
7 Card Mulligans:
6 land - BL
Mt Mt - LB BL BL - TH - ML
Is HD - GG - LB SS - P - ITR
Mt Is ST ST - BL - P TH
Is ST ST - BL - ITR ITR ML
Is Is Is Is Mt - BL - TH
Mt Is HD - BL SS - P - ITR
Mt Mt Is - BL BL SS - CTM
Mt Mt Is Is - LB BL SS
Is Is - KF KF GG - TH - ML
Mt - GG - BL SS SS - P - ML
Is - GG - BL SS - TH TH - CTM
Is ST - LB - P TH TH - ITR
Is Is - GG - LB LB BL - ML

6 Card Mulligans:
3Land - LB - 2ITR
No Land (reporting was lax in the first few hands)
Mt ST HD - BL - TH - ML
Is ST - LB LB SS - CTM
KF - BL SS - P - CTM
Is Is HD HD - SS - P
Mt Mt Is ST - SS - ITR


5 Card Mulligans:
Mt Is HD - LB - CTM
BL BL SS - ITR ML
Mt - SS - TH TH - ML

The chance of not drawing a KF in the top two (if you don't have one) is (49/53 * 48/52) = 85%, so a 15% chance of drawing at least one KF if you don't have one, in the first two turns.(21% in the first three, since that could be still fast enough).  So a no-draw, no-KF hand seems like a good ship.  Halimar Depths turn 1 basically gives you the first three draws in one step, but you also only get those draws for the first four turns, so it just gives you the 21% on turn 2, instead of across turns 2, 3, and 4.  (aka it makes you no less likely to be 100% screwed)

If you have a hand like Island-RedSource-Preordain, and want to see what your chances of aggressively hitting a KF turn 2 is (ship both if not KF to get a random card), you again have the 21% chance of hitting it on the turn 1 Preordain, but unlike Halimar Depths, you actually have your draw step turn 2 to hit it, and your draw step turn 3 may not be too late.  Also, unlike HD, you get new cards on T2, T3, not cards you already know won't help you.  (Like the "Sage Owl, see 4 lands, scoop" problem).
Therefore, if we have the P + U + R hand, and want to figure our KF chances, we have the top 3 (preordain pushes two + random draw), plus the next card (T2 draw) plus probably the next card (T3 draw) to try to hit the KF.  The odds of hitting a KF T2 become (49/53 * 48/52 * 47/51 * 46/50) = 72%, so a 28% chance of it working by T2 (34% by turn 3)

If you have double-Preordain + U + R, you're really hoping to hit the KF by turn 3, but then you have your 3 other cards, plus one preordain draw, to try to kill them on turn 4.
You have your same shot of hitting the KF on your first preordain + draw phase (28%).  If that doesn't work, you bring your total up to 50% of hitting the KF with the second preordain + third draw phase.

So, the double-Preordain + good mana + good instant+sorcery hand may be worth keeping.   Treasure Hunt math is harder, so I'm not sure how good those hands are, or how risky.  Some of those 7 card hands, with 2+ land, and 2+ draw spells might have been keepers, but it's so random.
=======================================================

Key: (and general order)
Land:
Is = Island
Mt = Mountain
ST = Scalding Tarn
HD = Halimar Depths

Creatures:

KF = Kiln Fiend
GG = Goblin Guide

Damage:
LB = Lightning Bolt
BL = Burst Lightning
SS = Staggershock

Draw:
P = Preordain
TH = Treasure Hunt

Other:
ITR = Into the Roil
ML = Mana Leak
CTM = Call To Mind

Call to mind gets an "other", even though it is similar to Staggershock:
-It does 2-4 damage, compared to SS's 4.
-It costs 3 mana, of lax color requirements
-It can give you the "rebound" on the same turn you cast it, instead of next turn
Because:
-It does 0 damage up front (though often 2-3 on the "rebound")
-It requires 4 mana to cast two spells in the same turn
-You have to pay for the "rebound" the next turn, which can interfere with casting another 3cc spell, or just from colored mana troubles going double-bolt, or double-blue spell.

It is definately better than ITR or ML for goldfishing purposes.  ITR could have infinite value, bouncing a wall or KHE (Time walk, as an instant?), or could have 0 value.

Hyper-Aggressive (Kiln Fiend) results.

To recap, the strategy was:
+4 Kiln Fiend
+4 Goblin Guide
-4 Pyromancer Ascension
-2 Foresee
-2 Mana Leak (Into the Roil, and Call To Mind, can remove walls, KHE (slow them down))

This was 100% goldfish, playing first every game.  One can only assume that my Turn X kills would only be better on the draw, because I would have one more card to double-bolt my opponent with, and so on.

This plan would be in place for:
Valakut
Maybe MGE (They could take their mana dudes out, for threats?)
Non-White, Non-Red Control.  (Kor Firewalker, Celestial Purge are TERRIBLE with Kiln Fiend.  White could have Leyline/Ascension in theory, which KF is great against, but you can't count on that)

This is probably 33% of the metagame, so having 8 sideboard slots against these matchups is possible.

I took fairly detailed notes on this, so I have some good results to share.

Overall Results:
Turn 3 kills: 0 (GG -> KF -> Preordain + 2 1cc burn)   I did put them to one once (GG -> KF -> 2xPreordain + Bolt), which could be a kill if they have a pay-fetchland.
Turn 4 kills: 14
Turn 5 kills: 10
Turn 6 kills: 4
Turn 7+ kills: 2

Turn 4 kill analysis:
Ending life totals for the opponent were:
0 |||||
-1|
-2 ||
-3
-4 ||
-5
-6
-7 |
-8
-9
-10 |||

Of the "close" wins (-3 or worse):
KF only: ||||
2*GG ||
 GG+KF ||
In these scenarios, they probably can't have a 0/4 wall, but can have any low-toughness blocker. (because I can staggershock it, instead of them)

Of the blowouts: (-4 or better)
KF only: || (Both of the -4's)
GG + KF ||||
In these scenarios, they can block a KF once, or a GG twice, etc.

In the Turn 4 kills, I mulliganed:
0 times: 10
1 time: 3
2 times: 1

T4 kill conclusions:
Having the Kiln Fiend is pretty important.  Double Goblin Guide only accounted for 2 of these 14 wins in this category.  If the GG's had been Arc Trails(A multi-function spell that I may end up maining), at least 4 of the GG+KF draws would have still been wins, and 1-2 more would have put the opponent on 1 life.  Obviously, the double GG wins would not have been wins as double arc trail hands.

I mulliganed at least once in 14 of the hands, so a mulligan has a 4/14 (28.5%) chance of being a turn 4 kill.
Of course, this also means that I didn't mulligan 16 times, and only won 10 of those on turn 4.

In some of these(typically the -10s), I did have ITR or Mana Leak "backup" for the win.  I only counted ITR for pumping KF if I could bounce my own Goblin Guide and replay it.  (for a net swing for 6 that turn, ug?)  I never counted Mana Leak as a spell I could cast.

Turn 5 kill analysis:
Ending life totals for the opponent were:
0
-1 |
-2 |
-3 |
-4 ||
-5
-6 |
-7 ||
-8
-9 |
-10 |

"Close" games (-3 or worse)
2*GG |
1*GG |
KF + mull to 4 + no land|

Blowouts (-4 or better)
2*KF ||
GG+KF ||
2*GG ||
KF |

In the turn 5 kills I mulliganed:
0 times:|||||
1 time:|||
2 times:|
3 times: |

Double KF draws tended to push out to 5 turns (with a blowout) because not enough instants and sorceries were drawn.
Single KF draws that didn't kill in 4 turns had similar problems.  Sometimes you draw alot of ITR or Staggershock (which, while good, does not put people away like 2-3x 1cc burn + draw)
Alot of these games involved preordaining past 3cc spells, or 2cc no-impact blue spells., or drawing more land than was optimal.  Kicked ITR on turn 4 for a land (as the only instant in my hand) was often the best I could do.

4 of these t5 wins had the opponent at 3 or less life at the end of turn 4.  Only two had the opponent at 10 or more life at the end of turn 4.  (those being 15-25 point double-KF swings)

Many of the double GG draws were single GG draws, that hit a second GG off a turn 2 preordain.

All of the hybrid GG games would have been equal wins with Arc Trail over Goblin Guide.  The pure GG wins were often 7 card hands, that would have to mulligan with something like Arc Trail, Staggershock*2, Preordain, Mountain, Island, Bolt.  (I kept most hands with playable GG turn 1, and these tended to be turn 5 or so kills.

Turn 6+ Kill Analysis.
My second game I mulled to 5, with 3 islands, a staggershock, and a KF.  Probably should have mulliganed that away.  Of course, I drew 2 more KF's, and no mountains.  My "opponent" was at 5 life on turn 7.
My 11th game, I mulliganed to 4, with Staggershock, Island, Treasuire Hunt, and Preordain, which drew into a turn 6 kill.
My 12th game, I mulliganed to 3, and never cast anything over 6 turns.
My 18th game, I started with KF/GG/Mountain/Preordain*2, Into the Roil*2 (perhaps a mulligan in retrospect?).  Dropped the GG turn one, KF landed turn 4, kill turn 6.
My 22nd game, I mulliganed to 2 mountains, scalding tarn, kiln fiend, preordain, and treasure hunt, which should be a good hand.  I drew only land (and ITR/Mana leak pushed down with preordain probably), so the kill escaped me until turn 6.  (1 life on turn 5).  Normally this hand would be an easy turn 4 or 5 kill.
My 25th game, I mulled to 5, with Island, Halimar Dephts, Scalding Tarn, Kiln Fiend, Into the roil, which is about as good as you can get on 5 cards.  I drew only land, to push the kill out to turn 6 (3 life on turn 5)

Mulligans were 2, 3, 3, 0, 1, 2 at this level, which is still pretty impressive, since only twice did i go past turn 6.
Still, this is 6 total fail hands, out of 30 games (20%), which is more than I'd like.  The 5 turn wins might not be real wins, because turn 5 may be after Primeval Titan comes down on their turn 4.

========================

Goblin Guide Openers:
For hands that had a castable T1 Goblin Guide:
T4 kill (With KF) |||||
T5 kill (with another GG drawn t3) ||
T4 kill with double GG, multiple burn ||
T5 kill with KF + GG drawn t5 |
T6 kill with KF played t4 (1 land opener) |
T5 kill starting with 2*GG |
T5 kill, Mountain, Halimar Dephts, KF |

4 of the 5 initial T4 kills would be the same kill with Arc Trail.  It's hard to say how many of the others would also be T4-6 kills, since I was always pushing Mana Leak, or ITR to the bottom, and those are likely to be Arc Trail slots if I didn't have GG.

Many of these hands would be mulligans without GG, but that's not necessarily a loss.  Most of my mulligans were T5 wins anyway.

So, GG definately did something, I just don't know how awesome it was in general.  Probably have to run more testing with Arc Trail, to get an accurate mulligan/scry representation.

Opening hands/Mulligans:
I have records of all the hands I kept, or mulliganed, in shorthand form.  I'll put these in later, but it is 2am, and I am very tired.  I did mulligan very aggressively, and did not keep most any hand that required drawing a specific color of mana to do something.

I had a fair amount of no-land hands.  I also had a fair amount of mono-blue-lands hands.  If we made this swap "for real", adding 1-3 Mountains for Islands would be a good call.

It is very possible that there are some hidden trends of what hands win and lose by T4, 5, 6, so looking at that should be useful.

Overall conclusions:
Goblin Guide may not add alot of percentage points to this matchup, over the anti-aggro Arc Trail  (a main-deckable Pyroclasm (like Staggershock), that could replace Mana Leak in the stock list)

I would have to mulligan quite a bit more to guarantee a Kiln Fiend if no GG was present, but AT goes quite well with KF, unlike GG.  (GG goes better with itself)  Arc Trail isn't exactly amazing vs Primeval Titan.  (But could lead to a burn-out with KF + ITR)

I definately wasn't missing the PA with this post-board plan.  I never played more than 6 or so spells, so the odds of even charging the PA with that are pretty low, especially since I'd have to take out some charging spell to put in the non-synergistic (with KF) PA.

~45% Turn 4 win is probably worth investigating further.  Another 33% on turn 5 might be fast enough, especially if they keep a hand that is good vs my game one deck.  (Acidic Slime is super-irrelevant G2, as is Nature's Claim, Gaea's Revenge, etc)

I'd rather not devote 4 (more) sideboard slots to this matchup if I can double-purpose those against some other matchup as well.  4 arc trail (for 4 mana leak) main would do some interesting things to my matchups, but it's hard to say whether it's good or bad without testing.  I've been thinking about some other changes, but that's another post.

Monday, January 10, 2011

Some RL testing - Thanks Charles!

A friend of mine who has been out of magic for a while, but was slinging turn 1 Juzams back in the day came over and played a few games with my terrible proxy decks.  The goal of which was trying to help me get a more realistic handle on the Valakut matchup, since i'm 0-n(matches), 0-2n(games) against that recently.
All games were post-sideboard, with the Valakut deck having 3x Acidic Slime as their anti-PA, and taking out 3x Pyroclasm (which do nothing against my deck).  I brought in my ultra-controlling anti-ramp sideboard of 3x Mindbreak Trap, 4x Flashfreeze, 1x Volitions Reigns.
Findings from my point of view:
1) I had no chance of victory without starting with at least one counterspell.
I kept some speculative Ascension, or Treasure Hunt hands, and they turned into alot of lands, or burn spells that didn't "get there" as it were.
2) Many of my game-wins involved casting two or three counterspells on turns 4-6.
The Valakut deck never got more than 4 threats attempted, and MBT eliminted alot of the explosiveness of the deck.
3) Some of the wins revolved around a "missed" summoning trap.
Getting a wall, or a Slime, is not what the deck wants to be doing.  I'd say roughly half of the traps (that resolved - 50% of all traps?) missed.  This is probably lucky on my part.
Of course, the Valakut player won a game or two where they missed on their trap, despite missing (7 non-threats to the bottom, check)
4) Several of the games would have been won by the other deck had the other deck gotten one more turn.
This is a testament to the explosiveness of the Valakut deck, and the methodical nature of the PA deck.  One Titan resolves, and it may be game over right then. (Turn 10).  That said, 2nd Staggershock to the face to charge PA, followed by Bolt + Bolt, or Bolt + Call to Mind is game over.  (4 + 4 + 6 + 6).
5) Volitions Reigns seemed pretty bad.
In all scenarios, I'd rather this had been Demolish.  I was never in a position to cast VR on a creature and not be dead anyway, or have things happen in response that effectively killed me.  VR would probably be alot better against MGE, where they cast their Eldrazi T6, and I take it on my T6, rather than the Valakut plan of casting PT on their T4, and me being dead, or casting it on T10, and me being dead to Valakuts that turn.  The in-between case rarely happens.  (The PT-ing player can often just kill the PT, or me, with the double-Valakut fetch from casting the thing)
The best thing I stole with VR was Valakut.  It might have won me that game, or it might have pulled a bunch of instant speed fetching + losing.
6) I lost several games where the Valakut player never resolved a "Threat".
In retrospect, I needed to counter the Acidic Slime that killed my PA, and just PA them out, instead of letting that resolve, taking 4-6 from AS and 15 from Valakuts.
The "just play lands out" with a Valakut or two in play was often good enough to kill the 11-counter deck.
7) Mana Leak is pretty bad if they are patient.
The 9-mana PT + Summoning Trap hand makes Mana Leak look really bad.  Mana Leak + Flashfreeze, or Mana Leak + Mana Leak can't solve both problems, so you have to have Mindbreak Trap, or you lose.  (They could not have the Trap, but they can just wait until they do, or have Valakuts, etc)
Basically, a patient Valakut player is pretty hard to beat, because they have inevitability in a few different ways.  (Mana Leak irrelevance, Valakut, Trap, or any of the threats resolving)
The best use for Mana Leak was countering a T3 Harrow, or T2-4 KHE.  If KHE resolves, Mana Leak is 100% irrelevant. (And they get to 2 mana on the play before you get 2 mana for Mana Leak)
I have to have it, to enable [Mindbreak Trap] OR [2x 2cc counters] Turn 4, to prevent dying to the Threat + Trap.  Drawing it turn 5+ is pretty close to dead.
8) I altered my Halmir Depths playstyle in this matchup. 
I played it more turn 1 in these games than I probably have in the hundreds of games I've played online.  (and there, I am playing it in desperation to find a land)
The reason was, I was never going to cast a counterspell (or hold one up, intentionally) turn 2.  I am going to cast Treasure Hunt, or Pyromancer Ascension turn 2, and hope to be holding 2 counters, or a MBT, turn 4.  I do fairly frantic tap-outting turns 1-3, so I can be "set up" for the turns that "matter" (aka where the threat that kills me resolves or doesn't)
Normally, I hold HD for turn 3, or turn 5, or some late turn, because I want to both maximize the yield of Treasure Hunt, or Scalding Tarn, and keep making land drops.  I still think this is a good strategy for the general match, because I may need to counter turn 2-3, and playing HD doesn't do anything about drawing those next few cards.  I just draw them in a different order.  So being able to get rid of them entirely with Scalding Tarn, or set up a big draw-2 or draw-n with Treasure Hunt turn 5 or 7 is a much better use of the card in general.  I just don't have that luxury against Valakut, and the spells I'd be countering aren't happening until turn 4 anyway.
========================================================
Observations from Charles's Perspective:
1) Frustration off the start, since I was playing a presideboarded control deck against a deck with few threats.  Playing a land (or a ramp spell) and saying go isn't "exciting", and when your biggest beater is a 2/2 for 5, things tend to drag out a bit.
2) Many mulligans from the Valakut deck, but it won games where it mulled to 5 (and I've lost those games online too).
3) As the afternoon wore on, the Valakut deck started winning more and more, and Charles exclaimed "I think I was just keeping bad hands".
The starting 7 Charles liked the most was 2 Valakuts + some junk.  That was a T8 win or so, which was about as fast as any game ended.
========================================================
This sideboard plan is trying to play the control role, and near the end of our testing, Charles made a suggestion that would be Heresy to the general player.
"What if I took out Mana Leak for 'something else'" (says Charles, not knowing the format).
That got us thinking, and Deprive seems like a non-horrific option.
Upsides:
-Still good after T3
-Reuses Halimar Depths (which are pretty awesome, with Treasure Hunt, Scalding Tarn, Preordain)
Counters small stuff late.
-2x Hard Counters off PA is just as good as 2x Mana Leak, and sometimes better.  You only have to pay the cost (which includes returning the land) on the one you actually cast, not the copy.
-People play around Mana Leak.  The only thing better than them playing around mana leak (and giving you extra turns) is not actually having Mana Leak in your deck.
-Early problems can be solved with the red half of the deck (Bolt it), let the blue solve the late game (via saying No, and meaning it)
-Deprive is only dead when any counterspell would be dead.  Mana leak is almost certainly dead after turn 7 or so.  (Doubled Mana Leak can still counter things, but I'd rather double Deprive, and reuse HD T7)
-Late game, I am often bouncing some dire threat that has a low casting cost (Kor Firewalker, Ratchet Bomb, Luminarchs Ascension, Sword of Body and Mind).  Being able to bounce it EOT, and counter it for real the next turn is good.
Downsides:
-Tempo hit.  Picking up a land is rarely purely good.
-Harsher casting cost.  Some games have only U3, and double-deprive T4 vs Valakut is non-plausible.  (Deprive-Flashfreeze is possible though, for the 2x counter mana leak + flashfreeze would give)
-You still have to pay the casting cost (aka return a land) even if it is countred.
-UU is harder to hold up on turns where you cast sorceries (your blue draw spells), which probably means you'll hold off on your draw spells until turn 5 or so (which means you have to hope you draw lands naturally?), or put down the counter shields.
Mindbreak Trap is also non-terrible.  I often side in MBT for Mana Leak in some matchups.
Upsides:
--Deals with Bloodghast (+ Bounce), which is a "recurring" problem.
--Deals with Vengevine
--Deals with Kuldotha Phoenix (Seen in casual room more than TPR)
--Deals with Eldrazi in a way that doesn't shuffle other answers back in.
--Semi-relevant to exile vs Sun Titan
--All the upsides of a pure counter, especially T4+
--Trap cost occasionally relevant
----counter wars
----final turn burn-out
----ww quest glink hawk/memnite turn.
----multi-preordain/explore turn from a UG ramp deck. (UG, BUG, RUG)
--Better than Mana Leak vs ramp.
--Not hit by Inquisition of Kozilek. (Hit by Duress, but IoK is more common)
Downsides:
--I do often counter things T2 in a general matchup. (as a proxy burn spell most of the time)
--Holding 4 mana up is alot more obvious/hard.
--More likely than either other option to be semi-dead against aggro.  (They have threats in play, you have counters in hand...)
--Even worse than Mana Leak to reveal off a Treasure Hunt.
If I went with either of these, I'd probably still do 4x Flashfreeze, 3x Mindbreak Trap (or Deprive) in the side. for ramp.
This is the control route.  It may be the best one, but both games we played after implimenting the "Mana Leak = Deprive" rule were mono-land Valakut Burn-outs.
========================================================
Possible alternative strategy: AGGRO!!!1!!one!
From online experience, the only G1's I've won vs Valakut were games where I won T4-5 on the back of an active ascension, 2x Staggershock to the face, and other burn/recursion.  I don't think I've ever won a game where I killed a Primeval Titan, or where a Summoning Trap was played for the trap cost. (G1)
It could be a good idea to have a sideboard strategy that emphasizes this (winning?) methodology
Such a plan would look like:
+4 Goblin Guide (8 damage by t4 if in opening hand and castable)
+4 Kiln Fiend (basically an ascension for my bolts/burn.  Better than ascension for Burst, Staggershock)
-8 of:
-4 Into the Roil (can bounce Wall/KHE to give another turn to swing)
-4 Mana Leak
-4 PA, to blank the Disenchant sb plan?
-2 Foresee (Too slow?)
-2 Call To Mind (probably not)
This is a very all-in-ish plan, that may stress the mana base quite a bit, depending on the -8.  A normal turn 12 has me with 3 mountains and 7 blue sources.  Red mana T1-2 is not seriously optimized.  And, if they ever get a Valakut off, or cast Pyroclasm (while I have creatures) or keep in bolts (fairly unlikely) it could be a fast loss.  Creatures also don't work terribly well with pyromancer.  They don't charge it up, and don't benefit from it being charged.
To that end, siding out the PA could be a reasonable plan.  It leaves in 12x burn, 8x ultra-aggressive creatures, 4x bounce (or counter) to remove blockers/reset KHE, and 8x 1-2cc draw spells.
Siding out PA also eases the mana stress.  (Taking out "red" spells for red spells is less stressfull on the must-dump-hand strategy than taking out 8x blue for 8x red)
Goblin Guide draws them to threats, but if they are still alive T5, they've probably won anyway. When I was testing KF before, I put it in for something terrible like Burst, instead of something terrible like Mana Leak, PA, or Into the Roil.  The hand of PA + Kiln Fiend + 4 land + something is pretty sad.
A T1 Goblin Guide means I only need 12 damage from other cards by turn 4 to have won, and I see at least 9 (8 + mountain) other cards on the play by turn 4.  (With no draw)
A T2 Kiln Fiend  means I need ~3 burn spells to win, or ~6 any other spell. (or ~4.5 spells total)
Burst does 5.
Staggershock does 10
Bolt does 6
The KF does 1 itself each turn.
I'd be ~58% to have at least one of the two in my hand off the start.
T1 Mountan/Tarn Goblin Guide (18)
T2 Land Kiln Fiend (16)
T3 Preordain into something else (4 to 7)
T4 Any spell (-10 to 1)
is pretty sweet.  Staggershock can fill it T3 as any two spells (with rebound on T4).  This plan uses 8 out of my 11 cards (3 land), plus a Preordain draw.
And of course the absurd:
T1: Goblin Guide (18)
T2: Double Goblin Guide (12)
T3: Double Bolt (0)
Or:
T1: Goblin Guide (18)
T2: Kiln Fiend (16)
T3:Bolt, Burst, Preordain  (-1)
8 of 9 cards (+ preordain draw) used on the play, with no two of the same card.  Bursts and Bolts are redundant in this draw.
This plan is about as far from casting Mindbreak Trap as is imaginable, but the Mindbreak Trap plan isn't winning any games against the Valakut land-draw.
Of course, any of these plans auto-loses to a castable pyroclasm on T2 (which might not happen, between ETBT lands, drawing Pyroclasm, and having a playable hand otherwise)
I'll run a few goldfishes with it, to see what happens, since goldfishing is 100% relevant with this sideboard plan.  Report to come later.
Historical weaknesses of this strategy are drawing PA + creatures, Not drawing T1 red, or just having bad hands (which mulliganing may fix?)

Thursday, January 6, 2011

Some more post-money-game analysis.

Was I happy to just be playing "for money"?

I read an article by PVDDR (a famous/good magic player), where he was listing out his strengths and weaknesses, and one of them hit me.  He said that one of his weaknesses was "he was happy to get top 8's", but then he'd end up never winning.  The top 8 was his goal, so he did that, and then didn't have the drive to go on and win.
(Whole article, for reference) http://www.channelfireball.com/home/pvs-playhouse-strengths-and-weaknesses/

I was pretty happy to be playing when I was playing.  I've played hundreds of test games in the TPR, and I finally got up the nerve to play "for real", and not feel like a jerk.  I jumped the hurdle, and feel like playing more (despite having a strong losing record), so maybe with time, I won't be happy to play, and my goal will be more to win.


I made some stupid mistakes.
I need to rewatch as many of those games as possible, today or tomorrow, just to observe what I was doing, and if there was anything I could have done better.   My best matchups were control and aggro, and I didn't have a winning record against either.  I didn't win a GAME against ramp.

I may have gotten overconfident.
Its hard to believe that I may have been overconfident playing these games, when at no point during the 20 games did I have a winning record.  But I felt like I was going to win (or should win) a particular match/game, and that led me to do things in a non-amazing way.  Assuming that you're going to win is rarely useful, and I often did assume that I was going to win.

This especially comes into play with the next point.

I need to mulligan more.
If I feel like I'm in a good matchup, or if I feel like I'm in a bad matchup, or if I feel like I don't know what the matchup is, I never feel like I can mulligan.

If I feel like I have a good matchup, I feel like I can beat them without having amazing cards in my hand off the start.  When in fact, I need to have some amount of useful cards in order for it to be a good matchup.  So, the Halimar Depths + 4 land + Treasure Hunt + any spell hand, while awesome at playing out lands, and drawing alot of (land) cards, is going to result in a loss against an aggro deck.  (Even if I have 4 pyroclasm, 8 1cc instant bolts, and 4 staggershocks in my deck somewhere)

If I feel like I have a bad matchup, I feel like mulliganing puts me even deeper in the hole, when mulliganing is exactly what I need to do to have a fighting chance.  Keeping my first 7 and being near-guaranteed of a loss is no more likely to win a game than trying for 6.

If I don't know what the matchup is (game 1 for example), I tend to keep pretty much anything that is lands and spells.  I think this is okay, but once I understand the meta a bit more (or know what decks people are playing), this will come up less often.

When I rewatch the games, I'm going to pay particular attention to my mulliganing.  I could not have had worse luck on keeping an anti-aggro hand and having them play control, or vice versa.

I need to concentrate while I am playing.
This is an obvious one, but one I feel like I need to state anyway.  This is the time when I need to be paying attention, if no other time.

Edit: Turns out, of  the 47 games I played in these 20 matches, about 8 were recorded.  So that took me about half an hour to go over, and I wasn't able to learn much.  Looks like a problem with MTGO.