Tortoise Format: Challenge #3

[combo_dude]

Rules and History

The Tortoise Format is a deckbuilding challenge that involves creating a deck of 7 cards.  The objective is to have the deck's optimal win be as slow as possible.

Here are the rules:

- 7 card decks, play starts with all 7 cards in the hand of a robot designed to play perfect Magic.  The robot will try to goldfish with the 7-card deck as quickly as he can.
- The Type 1 banned and restricted list are in effect
- You and your goldfish opponent do not lose as a result of not being able to draw a card

Rules introduced for format #2:
- Cumulative Upkeep is banned (*)

Rules introduced for format #3:
- Upwelling is banned (**)
- Wishes are banned (**)
- Ring of Ma'ruf is banned (**)

===================================
* Explanation of why Cumulative Upkeep is banned:
I ran the first Tortoise format on the old TMD, and we did not have cumulative upkeep banned.  Virtual, Cyberknight, and I collaboratively developed the following list:

1 Black Lotus
1 Taiga
1 Upwelling
1 Naked Singularity
1 Reality Twist
1 Sustaining Spirit
1 Cocoon

I can't remember the exact math, but this monstrosity takes trillions upon trillions of turns to win IIRC.  And of course, it was cumulative upkeep that was so broken.

===============================
**Reason why these cards are banned:
Upwelling is banned because it went in everything - there's no reason other than to encourage creativity here.
However, Wishes were broken by Jacob Orlove with the following deck:

Burning Wish
Mirari
Sand Silos
Leviathan
Phyrexian Altar
Wizard's School
Disturbed Burial


"Sideboard":
1 Golden Wish
1 Soldevi Digger
1 Battle of Wits
205 compeletely irrelevant sorceries (I can make a list if you like)

Optimal win (correct me if I'm wrong here):
play depletion land
accumulate mana
tap for 19, tap school for B, play leviathan and burial, sac leviathan to get R
R+4 = wish + mirari for B. wish + G. Wish
repeat 205 times, eventually discarding 200 irrelevant sorceries

Golden wish turn:

tap for 424, play leviathan, sac for W
tap school for W, golden wish+mirari for digger and battle
play Digger and battle
Return 200 cards to your library.

Next upkeep: win

By my count, that's 4338 total Blue mana needed, and an additional 207 turns for wishing (because you can only use the school once/turn). Add in one turn to play the silos, and one to win on, and that's 4547 turns in total.

This got worse when yours truly worked the Ring of Ma'ruf/Mirari/Golden Wish combo into this deck:

MD:
Wizard's School
Sand Silos
Leviathan
Phyrexian Altar
Golden Wish
Mirari
Disturbed Burial

SB:
Ring of Ma'ruf
Battle of Wits
Morality Shift
200+ enchantments

...which was even slower than Jacob's deck.

As a result, a 3-card SB limit went in. So then this happened (this was Alfred's deck, which I improved upon):

Black Lotus
Upwelling
River Delta
Phyrexian Altar
Breeding Pit
Mirari
Golden Wish

SB:
Pyromania
Steamclaw
Ring of Ma'ruf

...which still breaks the Wish engine in half. As a result, SBs are still around - there's just no way to get at them.


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

So, onto the new challenge! Remember, your deck must be as slow as possible.

Here is the slowest I could come up with this time:

[card]Crystal Vein[/card]
2 [card]Timberline Ridge[/card]
[card]White Mana Battery[/card]
[card]Soul Foundry[/card]
[card]Peacekeeper[/card]
[card]Epic Struggle[/card]

Off we go again! Please post how a deck works, no matter how fast or slow it is, and ideally a guess at the speed.

My deck plays WMB sacrificing Crystal Vein, gets up to 4 mana to play Soul Foundry imprinting Peacekeeper, then plays Test of Endurance before creating sufficient White mana to get up to 20 Peacekeepers through time (with upkeep!)

This is therefore:
Turn 3: WMB hits play
Turn 5: Charge counter
Turn 7: Soul Foundry
Turn 9: Charge counter
Turn 11: Test of Endurance
.
.
.
.
(The 60 mana to play 20 Peacekeepers is already dealt with by the mana sources, plus 20, 19, 18...turns of upkeep which is a LOT of mana, even though the WMB and a land/turn can contribute to the Peacekeepers. You need 3 mana, then 5, then 7..., then 41 on the last turn (= 1, 3, 5, 7...counters), which is a total of 1763 mana (1681 counters). As there's no opportunity to charge it once the Peacekeeper engine is going, you need to do this before - and it takes 2 turns to charge once!

I make that (1681*2)+11+1 (to trigger Test) = 3148 turns. Surprised it takes that long, and I'm shattered so I need to check my maths, but that's quite a lot...

Beat/correct that!

Adam

EDIT: After a couple of complaints, I unbanned storage lands. Go crazy.

Bear in mind that I don't have the final choice (realistically) about what's legal and what's not - it's thorme's brainchild, not mine - so if you can come up with a reasoned argument then I'm normally amenable. The only reasons that I'm making the decisions are that I posted it, and because I'm the one with access to the "edit" button on my post.

[Limbo]

Hmm, I'd say the first tortoise was the most challenging. Seems you are banning anything that makes the format interesting (cumulative upkeep anyone?). What's next? Ban regular upkeep as well?

[Jacob Orlove]

Actually, banning regular upkeep would be interesting too.

[combo_dude]

Okay. Here's JP's response to the first deck:

Let's see:

Just for Sustaining Spirit you would need to generate 2^1+2^2+...2^24 mana at the rate of one mana per turn. In order to generate that much white mana, you would need to keep Reality Twist around for that many turns, which would require 3^1+3^2+...3^23 turns. And in order to generate that much blue mana, you would need to keep Naked Singularity around for that long you would need 2^1+2^2+...2^23 turns (only 2^x because you only need 1UU to upkeep Reality Twist.) And in order to keep Naked Singularity around that long, you would need 3^1+3^2+...3^24 turns. And to cast everything you would need 1 turn for Upwelling, 5 turns for Naked Singularity, 1 turn for the colorless mana in the CC of Sustaining Spirit, and 2*(2^1+2^2+...2^24) turns for the colorless mana for both Reality Twist's upkeep and Sustaining Spirit's upkeep.

The whole idea behind this is that we've already completely broken cumulative upkeep, because as soon as anyone says "how about a deck with cumulative upkeep" someone else will say "hey, take a look at this, it crashes computers!"

Same with the lands - I'm banning them because they were everywhere. I'm trying to encourage people to use a different approach in this one - one that doesn't revolve around the usual ways of making it kill in ridiculous numbers of turns. I'm banning the stuff that we'd decided to eliminate at the end of the last tortoise thread.

Post a storage land deck or an Upwelling deck if you want - I just think the other approaches are more challenging. It's not like this is a real format is it?

[virtual]

If your opponent has a chance to make a play, will he play optimally (the best way for him to live longer)?  Or is he working against your goal of taking a long time to win, and will he make decisions that let you win faster...

*gears turn*

I think he should play optimally for himself to live the longest, and assume that you will play optimally to try to kill him...

rulings? thoughts?  

IF he had a chance to try to kill you... would he try to?  This could get very complicated....  Like 3cb calculations... but for TONS of turns =D

-Virtual

[Jacob Orlove]

Isn't this against a goldfish opponent?

[Alfred]

See, the problem with banning the storage lands and/or upwelling is that they STIFLE creativity, rather than encourage it. This is due to the fact that more slots will be devoted to ways of generating mana, rather than actually creating a cool combo. I think ban upwelling, or ban storage lands, but for god's sakes, don't ban both! BTW, I'm so glad that tortoise is back.

[thorme]

BTW, I'm so glad that tortoise is back.

My baby is back!!!

I must get to work...

[combo_dude]

See, the problem with banning the storage lands and/or upwelling is that they STIFLE creativity, rather than encourage it. This is due to the fact that more slots will be devoted to ways of generating mana, rather than actually creating a cool combo. I think ban upwelling, or ban storage lands, but for god's sakes, don't ban both! BTW, I'm so glad that tortoise is back.

That's a fair point. I'm more inclined to take the storage lands off the list, rather than Upwelling, as Upwelling still goes with Black Lotus+depletion land as near-automatic components.

[virtual]

My question specifically has to do with if I have a forbidden orchard.  

If i use it 20 times with no blocker myself... should I be considered dead?

If i use it once, and attack with a 1/1, will he block?  

Is there some algorithm for determining optimal play (attacking/blocking), or is it NP complete...

Now everyone can steal my tech, but maybe that gives some insight into why I asked...

If he blocks, you get more flexability in the cards that you can use.  X/1's bein aren't nearly as impossible to play as before...  Or even a huge creature that he can chump every turn because you give him a creature...

Can I get some opinions?

-Virtual

[Alfred]

I think that the goldfish opponent should just attack with the tokens at every available opportunity, even if you have a blocker that will kill it. This way you avoid the "best case possible attack" scenario that you are describing, also, the opponent should never block, because he is intent on killing you. This would also avoid the scenario of using a forbidden orchard to create chump blockers for your creatures. The goldfish opponent should just simply have the mindset: "Must kill opponent" with no secondary motives, so there is no either/or dilemma. I think that this should be the ruling on the matter because it's simple, and it involves no higher thought or logical conundrums that would lead to confusion when using such a card.

Also, I move for unbanning the storage lands. They are one card that solves mana generation problems, and allows for 6 other cards to be used to create combos. More card combinations = more fun.

[wonkey_donkey]

Here is the slowest I could come up with this time:

[card]Crystal Vein[/card]
2 [card]Timberline Ridge[/card]
[card]White Mana Battery[/card]
[card]Soul Foundry[/card]
[card]Peacekeeper[/card]
[card]Epic Struggle[/card]

Off we go again! Please post how a deck works, no matter how fast or slow it is, and ideally a guess at the speed.

My deck plays WMB sacrificing Crystal Vein, gets up to 4 mana to play Soul Foundry imprinting Peacekeeper, then plays Test of Endurance before creating sufficient White mana to get up to 20 Peacekeepers through time (with upkeep!)

This is therefore:
Turn 3: WMB hits play
Turn 5: Charge counter
Turn 7: Soul Foundry
Turn 9: Charge counter
Turn 11: Test of Endurance
.
.
.
.
(The 60 mana to play 20 Peacekeepers is already dealt with by the mana sources, plus 20, 19, 18...turns of upkeep which is a LOT of mana, even though the WMB and a land/turn can contribute to the Peacekeepers. You need 3 mana, then 5, then 7..., then 41 on the last turn (= 1, 3, 5, 7...counters), which is a total of 1763 mana (1681 counters). As there's no opportunity to charge it once the Peacekeeper engine is going, you need to do this before - and it takes 2 turns to charge once!

I make that (1681*2)+11+1 (to trigger Test) = 3148 turns. Surprised it takes that long, and I'm shattered so I need to check my maths, but that's quite a lot...

Beat/correct that!

I'll correct it - hopefully beat it too sometime soon, but that's still being worked on for now. How do you get 1691 counters? I make it 440 mana needed to drop 20 peacekeepers - at least, that's what 3+5+7+9...+39+41 adds up to. Either I've misunderstood how you're winning, or you've got a figure from somewhere.

I'd probably rather keep storage lands unbanned - it's true that you can get some fairly slow kills without them, but they are fairly few and far between. Anyway, most decks will then start using mana batteries or something as a replacement, which is not necessarily any better. May as well keep both around to give more choice, if not a more creative way to solve the problem.

Tom

[virtual]

I personally think that the goldfish opponent should attempt to not die.  As a normal goldfish opponent would.  

Lets say you're goldfishing oath, and for some reason you use verdant force as your oathed up critter.  You need to calculate the extra turn lag for the block... and I think that is standard goldfish procedure.  

I guess the question is... does the goldfish opponent know your motives.  And if so... does he work against them.  

Does it make sense to have a "goldfish opponent" act differently, depending on your personal goals for the game...

I think that optimal play by the goldfish opponent should be assumed, where optimal play means that he plays to live the longest that he can, or kill you...

This makes calculating certain scenarios a bitch, but I guess that's what I get for trying to use forbidden orchard.  

For ease of use, we could just say he is a non-offensive opponent, and he always will chump block.  

Or we could just say that he never blocks...

I don't like that one personally, because it avoids my tricks, and, well being creative and trying to work with constraints like that is pretty fun.

-Virtual

[Jacob Orlove]

We're assuming that your deck is played optimally to win as fast as possible; I think we should assume the same for the goldfish opponent, and give them perfect information. They shouldn't try to speed up your kill, though--even if they can't win (which they'll know due to perfect information), I think they should play to stay alive as long as possible.

[Nibble]

I'll correct it - hopefully beat it too sometime soon, but that's still being worked on for now. How do you get 1691 counters? I make it 440 mana needed to drop 20 peacekeepers - at least, that's what 3+5+7+9...+39+41 adds up to. Either I've misunderstood how you're winning, or you've got a figure from somewhere.

I believe this discrepancy comes from the fact that he's using the Battery for all his mana. I'm not sure of the exact math, because once you stop adding counters to the Battery and using it to cast/pay upkeep on Peacekeepers, you can also be using your land to pay upkeep or Foundry costs and all that.

Well actually, I've got nothing better to do, so I'll give it a shot. The turn 1-11 startup is definitely fine. The way I look at it, the Battery produces a 'free' W per turn, not requiring a counter. Effectively, this reduces Peackeeper cost to 2 per turn. Your Ridges will produce 2 every other turn, so the cost to make a Peacekeeper alternates between 0 and 2 counters each turn. Total: 20 counters to make the Peacekeepers.

Then add the upkeep costs. Land and free WMB mana is already accounted for, so these need to be paid in full. Over the 20 turns of production, this will be 0 + 2 + 4 + ... + 38. Total: 380.

The total number of counters needed, then, is 400. Because of the nature of the Ridges, each counter takes essentially two turns to create, so that's 800 turns of charging.

It would seem, then, the final time for winning would be:

11 turns startup + 800 turns charging + 20 turns production + 1 turn activation = 832 turns.

So, yeah, where did 1600+ counters come from?

[combo_dude]

, and I'm shattered so I need to check my maths

Maybe "check" is wrong. "Completely rehabilitate so I know how to add things up" is perhaps more accurate.

I was probably using the {n(n+1)}/2 formula to add numbers up and working an extra factor of 2 into the equation, at a guess.

I would definitely rule that the goldfish is working to kill you. It's like your opponent's deck is also handled by the same robot (i.e. perfect information) - the difference being that his hand consists of 0 cards, not 7.

@thorme - any word from TEH CR34TOR himself?

[CaptainPlanet.dec]

Goblin Bomb
6 Mountain

It is possible to get every flip wrong for 80923673123612371236126253231 277321653217 turns. And then get the rest right. Beat that.  :lol:

[fadeblue]

This looks like a real interesting challenge... Hmm...

Yeah, how does the random rule work? I was considering Mana Crypt in a Test of Endurance deck, but wasn't sure what counted as "optimal win."

Also, about how many turns is normal for Tortoise decks? The one I have right now is only 828 turns...

[combo_dude]

Random effects I will assume go in the pattern of you lose one, then win one, then lose one, then win one etc. - in the long run it would even out, so have it as an every-other-turn thing.

And it varies as to the speed of wins - 828 is pretty good though. As I said, post the deck even if it's "only" a couple of hundred turns (!) - it's not like there's an incentive to hide any tech or anything, so letting the community work on it makes sense.

[raye]

Goldfishes don't fight back.

You'd give them tokens, and they would sit there and look sweet, without blocking, attacking, or doing anything, unless forced to do so by other forces, such as Lure.

Why add unnecessary rules?

-Kowal, still hijacking Raye's computer

[combo_dude]

We're assuming that your deck is played optimally to win as fast as possible; I think we should assume the same for the goldfish opponent, and give them perfect information. They shouldn't try to speed up your kill, though--even if they can't win (which they'll know due to perfect information), I think they should play to stay alive as long as possible.

I think that optimal play by the goldfish opponent should be assumed, where optimal play means that he plays to live the longest that he can, or kill you...

The goldfish opponent should just simply have the mindset: "Must kill opponent" with no secondary motives, so there is no either/or dilemma. I think that this should be the ruling on the matter because it's simple, and it involves no higher thought or logical conundrums that would lead to confusion when using such a card.

Because it makes no sense at all to be trying to win when you've put yourself in a position where any rational human being - or even an irrational human being - could kill you that turn.

Besides, I think a land that produces a mana of any colour without a drawback, as would be the case on Forbidden Orchard, hasn't been printed for a reason, if for the sake of some slightly different formats. Let's keep it that way.

[CaptainPlanet.dec]

Random effects I will assume go in the pattern of you lose one, then win one, then lose one, then win one etc. - in the long run it would even out, so have it as an every-other-turn thing.

And it varies as to the speed of wins - 828 is pretty good though. As I said, post the deck even if it's "only" a couple of hundred turns (!) - it's not like there's an incentive to hide any tech or anything, so letting the community work on it makes sense.

So then you could never actually win with Goblin Bomb. Damn.

[fadeblue]

Okay, here's my 828-turn deck:

teferi's isle
crystal vein
blue mana battery
test of endurance
caribou range
celestial prism
capsize

1) isle
2) vein
3) battery
5) prism
charge the battery for 12 counters
31) tap isle and battery for 15 blue, use the prism to get W, capsize the prism with buyback, cast prism, use prism to get W, play test of endurance

It takes another 26 turns to play caribou range on the isle. After that, it takes 26 turns to gain 1 life from the caribou range, and then 1 turn to win, so that gives

31 + 26 + 30*26 + 1 = 838... I guess it was 838, not 828.

But I came up with an even better deck after that:

teferi's isle
crystal vein
white mana battery
caribou range
altar of shadows
leshrac's sigil
tendrils of agony

It starts off the same. You get range out and start creating lots of caribou tokens (it takes 4 turns to produce 1 token). After you have enough, you play Altar and start killing off your tokens until you get 36 charge counters on the Altar. The next turn, you play Sigil 9 times and then Tendrils for the win.

Due to mana burn (and you can only sink 2 each turn into the Battery), you need to produce lots of extra tokens for life - 359 of them, by my count. Range comes out on turn 7, and after that it's 4 turns for each token, and you need 359 + 36 = 395 of them, so at turn 1587, you're ready to start killing your tokens. After that, it takes 75 turns to play Altar and to kill 36 tokens (including the 1 turn to play Tendrils). This gives a total of 1662 turns.

EDIT: Rechecking my numbers, it looks like it's only 349, so reduce that number by 40 turns = 1622.

EDIT 2: Never mind - I'm a fool. It started out as a Drain Life deck, which is why I was thinking you couldn't spend the mana, but with Sigil, you can sink most of the mana each turn. I still think there's potential with Altar of Shadows though.

EDIT 3: I just thought of using Slaughter instead of Sigil. It means fewer turns to charge up the Altar, but it costs an extra 4 life per use (so it's more efficient to take mana burn than sink the mana into Slaughter, until the winning turn). That means 108 turns of life-gaining. I haven't calculated the new total yet, so if anyone else wants to, that would be great.

[Limbo]

Although I haven't decided on a kill condition yet, I was thinking along the lines of this as a really, really, really slow mana engine:

City of Traitors
City of Traitors
White Mana Battery
Soul Foundry
Peacekeeper

This is just as slow as wonkeys engine to create mana with the mana batterys, but leaves room for 1 more card, so somehow a slower kill.

Phyrexian Altar is quite nice to create mana of any color with the peacekeeper, now all we have to do is find the most mana intensive non-attacking kill we can find. I think Soul Burn is quite slow, but is there something slower?

Edit: Funny, I am getting dragged in the Tortoise format again. Lets see if I can help to break this one as well.

[CaptainPlanet.dec]

Although I haven't decided on a kill condition yet, I was thinking along the lines of this as a really, really, really slow mana engine:

City of Traitors
City of Traitors
White Mana Battery
Soul Foundry
Peacekeeper

This is just as slow as wonkeys engine to create mana with the mana batterys, but leaves room for 1 more card, so somehow a slower kill.

Phyrexian Altar is quite nice to create mana of any color with the peacekeeper, now all we have to do is find the most mana intensive non-attacking kill we can find. I think Soul Burn is quite slow, but is there something slower?

Edit: Funny, I am getting dragged in the Tortoise format again. Lets see if I can help to break this one as well.

High Market
Test of Endurence

Thats soooooooooo slow.

[ReAnimator]

I think Soul Burn is quite slow, but is there something slower?

   
Magma mine?
brush with death?
 
also is gemstone array slower than a mana battery since it doesnt tap to add one mana on its own?

[CaptainPlanet.dec]

I think Soul Burn is quite slow, but is there something slower?

   
Magma mine?
brush with death?
 
also is gemstone array slower than a mana battery since it doesnt tap to add one mana on its own?

Yes, I also think Gemstone Array is slower. I've been experimenting with it.

City of Traitors
City of Traitors
Gemstone Array
Soul Foundry
Peacekeeper
Peacekeeper
Epic Struggle

or

City of Traitors
City of Traitors
White Mana Battery
Green Mana Battery
Soul Foundry
Peacekeeper
Epic Struggle

[fadeblue]

Teferi's Isle+Crystal Vein+Battery/Array is still much slower than City+City+Battery/Array. With the Isle you can only charge once every 2 turns.

[jro]

Here's something I spent way too much time thinking of.  The detailed turn structure is in the code boxes below.  The general idea is to use those awful sac lands from Masques (in this case, Saprazan Skerry) combined with Soldevi Digger as your only mana source, so you only have mana available every other turn.  I use Implements of Sacrifice to ramp up to 3 mana once every 6 turns, which lets me put a charge counter on Iceberg.  To get the best effect out of that tedious process, you want to cast something with very high CC that isn't a threat to your opponent, hence Dragon Tyrant.  He's castable because his cost only has RR (which you can get from Implements), but his upkeep requires RRRR, which you could never get.  So if he's going to die, you may as well Soul Net him, and use Test of Endurance for the win.

I count 1698 turns, but I could very well be wrong, because I'm very tired.  The setup might be a few turns shorter optimally, but it's the "inner loops" that really matter.  Hopefully, my shorthand is relatively clear.  Most of the key plays involve using the Skerry during your upkeep to dig something, very often the Skerry itself.

Code:

CARDS
Saprazan Skerry
Soldevi Digger
Implements of Sacrifice
Iceberg
Dragon Tyrant
Soul Net
Test of Endurance

BEGIN SETUP
turn 1 skerry
turn 2 digger
turn 3 dig skerry, draw skerry, skerry
turn 4 implements
turn 5 dig skerry, draw skerry, skerry
turn 6 skerry->implements->iceberg.1 //with one charge counter
turn 7 dig skerry, draw skerry, skerry
turn 8 dig implements, draw implements
turn 9 dig skerry, draw skerry, skerry
turn 10 implements
turn 11 dig skerry, draw skerry, skerry
turn 12 iceberg.0->implements+skerry->test of endurance
turn 13 dig skerry, draw skerry, skerry
turn 14 dig implements, draw implements
turn 15 dig skerry, draw skerry, skerry
turn 16 implements
turn 17 dig skerry, draw skerry, skerry
turn 18 skerry->implements->iceberg.1, iceberg.0->soulnet
END SETUP

BEGIN LIFE GAIN LOOP
 repeat 30 { // need to gain 30 life
  BEGIN ICEBERG CHARGE LOOP
  repeat 8 { // charge up iceberg for tyrant and soul net
   turn 1 dig skerry, draw skerry, skerry
   turn 2 dig implements, draw implements
   turn 3 dig skerry, draw skerry, skerry
   turn 4 implements
   turn 5 dig skerry, draw skerry, skerry
   turn 6 skerry->implements->iceberg.++
   } //end 8 loop, iceberg has 8 charge counters
 turn 1 dig skerry, draw skerry, skerry
 turn 2 dig implements, draw implements
 turn 3 dig skerry, draw skerry, skerry
 turn 4 implements
 turn 5 dig skerry, draw skerry, skerry
 turn 6 iceberg.1->implements.R+skerry->tyrant
 turn 7 tyrant dies, iceberg.0->soulnet->gain 1 life, dig skerry, draw skerry, skerry
 turn 8 dig tyrant, draw tyrant
 } //end 30 loop
END LIFE GAIN LOOP

TOTAL TURNS = 18 Setup + 30 * ( 8 * 6 + 8) Life loop
            = 1698

You could, of course, charge the iceberg all at once if you preferred.

Some similar strategies I tried involved Animal Boneyard, Rabid Wombat, and Astrolabe, and Necropolis, Iceberg and Animal Boneyard.  I also tried something with Caribou Range or Farmstead, but it never came together.

[Limbo]

Improved idea:

Black Lotus
Jeweled Amulet
Gemstone Array
Lava Tube
Soul Foundry
Peacekeeper
Epic Struggle

With Lotus + Lava Tube you play the Gemstone Array. Play Jeweled Amulet. Now you can start charging the Gemstone Array with the Jeweled Amulet + Lava tube, adding 1 counter every 4 turns, due to the delay of the lava tube (yay). To win, we need to play 20 peacekeepers, a soul foundry and an Epic Struggle, so this would take (approximately):

Casting gemstone array + untapping Lava tube again         : 2
Charging for Epic Struggle + Soul Foundry         2x 4x3 +2 : 26
Casting of 20 Peacekeepers                                  20x4x3 : 240
Paying Peacekeeper upkeeps 4x(2+4+6... ...+38) = 4x380 : 1520
Correction in last 20 turns for Lava Tube                     4x4 : -16
Total                                                                               : 1772

I think this calculation is correct, and slowest thus far. Anyone know a way to slow this even more?