1. This guide refers to something that happens just after level 1000.
   
2. It reveals things that happen en route to level 10,000 (aka 10K).
   
3. It reveals some (but only some) basic information that can be difficult to find.
   
4. SD2 is a forgiving game. Any team, any gear will work, because every failure gives you essence to strengthen your party. I never changed my build to adapt to a Dark Lord; I never had to; I just upgraded my gear and relics until I could bull past them after losing a few times.
   
5. I suggest it is more fun to figure things out for yourself. Muddle through like I did. Get to at least level 10K before you compare your team build with this one. (I'm past 60K and still finding mistakes and things to tweak. Apparently I'm not good at optimization.)
   
6. Scripting, theory-crafting (or deliberately NOT theory-crafting), fine-tuning and a bit of automated grinding for better gear is where the fun is in SD2. If you resort to guides like this to design your build for you instead of as an aid to improve your own working build, you've lost the plot along with the fun.
   
7. At present (v 1.05), the last new gear is introduced at 10K. (This spoiler is that there's no spoiler.) So 20K - after you've collected and considered the 10K+ gear - is a good place to stop and take stock.
   
8. This build is not optimal at 400K. It isn't available until after 10K.
   
9. While the concept of a build to maximize essence return should remain viable from 10K to ~400K (as Gold Find comes into its own), the gear will probably change with future updates (1.06?). I particularly look forward to expanded craftables.

You too can use my practical advice to help you go forth and conquer all the stuff that isn't in the game yet and which will render this build obsolete when it gets here.

At 20K I changed my party and order to Dark Lord, Thief, Nurse, Darkmage, Darkmage, Darkmage.
I'd thought it was too early to drop my Mystic, but the Healing Fairy had no trouble keeping me going.

There is 1 chance in 4 for portals to appear on levels ending in 3, 6 and 8, with equal chance the non-?????? choice will be fairy, mineshaft, treasure or warp. So 53.3 levels per fairy on average. It is unlikely your team will go great stretches without seeing a Healing Fairy: 50 levels: (15/16)^(3*50/5) = .380 chance of no fairy, 100: .144, 150: .055, 200: .021.


The idea for this build is to use short-animation attacks to clear floors quickly to optimize essence collection to improve the team as fast as possible.
-The Dark Lord kills bosses and Janitors and damages other enemies.
-The Thief steals essence, the Nurse heals, the Darkmages kill.

The Dark Lord's Shadow Slicer ability does .5x attack per party member against every enemy.

Dark LordThiefNurseDarkmage
This party never stops to mine ore. If it isn't essence, it isn't worth it.

The Dark Lord's Dark Savior passive can save 1 ally per battle from death, so we gear him to survive an ambush, the Mages to maximize damage and the Thief and Nurse to strike a balance.

All gear below (that has levels) should be upgraded to Level 10.
Half the gear must be crafted. Gear made with Uranium needs mining.



Essence Find is the chance that the first enemy has essence to steal.
It will be colored in purple if it does.

I originally gave each Darkmage a Magnificent Robe (+33% to magic attack).
I originally gave the Thief Ossein Armor.
They now have Radioactive Suits, which do slightly more damage but need piles of Uranium.
(See the Mining section.)


Only Default, Laser & Skull Bash attacks get the 50% back attack bonus.


I saved their gear loadouts.
Click on the current party to get to the party screen, click on a party member to get their info, click on the clothes-hanger at the bottom to save their current gear set.


I connected everybody to new scripts using the new gear.
(SD2 only allows 10 scripts - not even two teams! - so I had to delete some old ones.)

Click at the bottom of the Arena, then on Soda Scripts, to design class scripts.
Click on the blue book at upper left, then Characters, to assign a script to a class.

I started with the essence fairy who, sadly, adds to essence find rather than essence recovered and is thus little help if you're already near the maximum EF.
I also tried the hit point whale, the evasion chameleon and the damage reduction rock.
At some point, Cliff may be necessary to prevent the Dark Lord from being one-shot in ambushes.

The current best pet for this build is Cassiel, who boosts magic attacks by 8%.
She adds maybe 7% to run lengths (which also adds ~7% to time taken). QV the Run End section.

I have no idea what's optimal; this is my best guess and works okay for me.

I am running Relic of Attack = Relic of Magic Boost = Relic of Darkmage = Relic of Dark Lord.
At 100 above that I run Relic of Health.
At 100 below I run Relic of Physical Boost = Relic of Thief = Relic of Nurse.

I set Relic of Mana to flat 500 and Relic of Gold Find to 200.
(I have since started increasing them - especially the latter - but 500 and 0 are plenty to start.)

All maximizable Relics except Dungeon Keys (100 of 300 is already ridiculous) are maxed.

I set "Allow keys to be used for:" to Healing Fairy only.
Healing Fairies have HP and MP Heal set at 70%.
I set "Chance to pick unknown path when no key used:" to 0%.

Fairies use time but a drought can wreck your run.
I suggest setting MP Heal higher while your Mana Relic is low.

Looking at hourly numbers for the run from level 1 through 38K, I found that essence collected per hour rose until roughly 2/3 of the maximum dungeon level reached, then leveled out. The last 33% of the levels took 40% of the time but gathered 71% of the essence.
If I expect each run to go about 10% farther than the preceding one on average, I should start a run at roughly 3/4 of the current record depth achieved to maximize essence collected over a given run time. (Adjusted lower or higher so a run doesn't die just after I go to bed.)
If succeeding runs go 50% farther and If 2/3 in is the best place to start a run, then runs should start at the floor the previous run ended.


(I really should do another data collection run with the current, more-reliable build, but it takes so dang long and next time even more of the numbers will be low-res.)

Ambushes can occur on floors ending in digits 2, 3, 4: 4 enemies; 6, 7, 8, 9: 2 enemies.
Ambush chance is about 4.4% per floor overall.
Ambushes don't seem to occur closer than maybe 9 or 10 floors apart.
There is one attack per ambush.
Attacks have an even chance of striking every position except double chance at the first.

This build dies soon after it hits the threshold where its damage can't kill every enemy unit outright.

To maximize efficiency, minimize time spent limping along after this threshold is reached. To wipe the party as fast as possible, put the unit most necessary for its survival in the position where it will be killed the quickest.


Here's a way to calculate when a run will end, using numbers from a previous run:
Dark Lord: Phantasmal Claw: 8.5K damage, Darkmages: Noxin: 17K.
3*17 + 8.5 = 59.5K: total minimum damage per enemy mob.
level * 7/9 = enemy hitpoints at a given level (3x that if it's a boss (irrelevant here)).
59.5/(7/9) = 76.5K: the approximate level where their hitpoints will pass that damage.
Level I'm at: 72.5K, leaving 4K levels.
4K ÷ 680 levels per hour run speed: 5:53 hours remaining. Call it 6 hours to go.

You can see the Claw and Noxin numbers by watching your run. You can make a bad (high) calculation of floors per hour by clicking the Gold/Keys/Essence strip at bottom left to bring up the run info screen which lists floors traversed and time spent as its last two items at the left.

Speed varies in stages over the course of a run, decreasing every time the enemies pass a threshold which requires more hits for your party to remove them. If you make only relic investments while keeping your team, scripts and gear constant, you will have the same speed in floors per hour in each stage of your runs and for your runs overall. Damage and hit points will increase, the lengths of the stages will increase, their start and end levels will increase, but the speed must be the same.
So you can get far more accurate estimates if you know know which stage you're in and what the speed from that point was for comparable previous runs.

My last run before I decided to speed up my essence build was levels 17001-23999 and recovered 97K essence per hour.
My next four runs while this build was under revision were 1-38005 @ 138K EPH, 29991-50701 @ 260K EPH, 38001-76471 @ 357K EPH and 59991-96003 @ 516K EPH.

The last one was my check run after v1.05 dropped, with upgraded Pearls and the Thief automagically opening chests quickly. The party was exactly as described here except I forgot Cassiel and hadn't yet swapped the Dark Lord's Copper Bracelet for the chicken fetish.
My threshold was 126.4K, but I lost the Dark Lord in the Low 90s and the rest of the party at 96003 from having too many asleep. I'd guess that's what killed the DL as well.

My final run died early too. Stats:
Floors 96001-118999, harvesting 21.6M Essence and 66.8B Gold in 30:41:44.
749 floors, 704K essence and 2.18M gold (converts to 311K more E) per hour.
//edit 28/8: I'll try again, without upgrading relics, just short of 119K, to find out what happened.
I'd guess the DL died in ambush due to lack of HP, then the rest failed at the Janitor.

//edit 28/8: I'm now doing a mining run without Geodes to double-check the effect of Ore Find at 150 instead of 220 and will add that result to the mining section.


I stopped my last mining run when I had 33 Uranium ore to finish the Rad Suit for the Thief.
1501-13150: 33 U, 197 Pt, 1.7K Au in 18:51:49.
618 floors per hour, 1.75 U per hour, 2.8 U per 1000 floors. Well fewer U than UsUal.


So long as you start at the same relative position in your runs:
Essence increases linearly with level, so EPH must increase proportionately as you invest the essence that allows your team to succeed with ever longer runs at ever higher levels. OTOH - as noted in the previous section - run speed in floors per hour should remain the same.

I have outrun the content, so my immediate plans are to wait for SD2 to catch up.
I am currently doing a final 1.05 check run and then will retire from the field while we wait for a content patch.
Ad interim, I will grind three more Rad Suits -- one for the Thief and two spares (I'll leave the Nurse with Ossein). They really only mattered for the Darkmages.
Then I might see how far I can go with Necroshine's 6-Junky Challenge (with the sound off).

More distant, post-content plans might include comparing builds at 400K (using net Essence, including Gold converted in the Magic Well), dropping the Nurse or Dark Lord, trying out revised classes and new gear, trying new challenges.

If you want to progress your SD2 party as far and fast as possible, don't stop to mine.
The gear I'm after will help my party going forward, but not as much as the essence I could have gathered in the same time. On the other hand, if you don't want to get too far ahead of the content; if going as far and fast as possible isn't your main goal (though it's the only goal in view until the next installment of said content drops) -- well, then, you may as well stop and smell the roses/ solve some scripting puzzles/ grind4gear.

I've mostly stopped progression, so I decided to make some Radioactive Suits.
Radioactive Suits give 218% damage reflection and 14% attack increase. The former can occasionally speed you up slightly and entertainingly when a janitor cleans himself up, but the latter is the reason to get them; +14% to base damage is better than +33% to magic damage (albeit not by much and not soon).
They need 25 Uranium Bars apiece, so making suits for a full party needs 25*3*6 = 450 Uranium Ores.
At 3-4 U ores per 1000 levels, that's over 100,000 levels to grind. At 2-2.5 Uranium Ores per hour, that's over a week in real time. Good thing this game will grind itself in the background.

Ore rarity doesn't vary with dungeon level. This is overridden by some lode types being disabled until floors from 500 to 1500.

I can tell when my party dies (no more faint "tink"ing in the background) and don't mind that it does; it's past the point where it's most efficient. If you want a longer-lasting team, start early (floor 1501) and swap the last Miner's Geode out for a Doodle Charm to prevent sleep. Otherwise a 000 Janitor (license to massacre) WILL wipe your whole party eventually.

My Miner League Team is 5 Miners and a Huntress (last so she wastes attacks the least).

Gear
*Don't be like me. Remember to upgrade your gear.

Miner Huntress

Edit 18/8: I have changed this to harvest all ore that appears alongside enemies by transmuting the last enemy instead of the first. Suppose 3 enemies + 1 gold: 4 attacks to mine the gold, 2 to kill 2 enemies, leaving a last enemy that gets in one attack and is then transmuted and mined. Ore occurring in with enemies is rare and mining is done at relatively low levels, so this is still safe
This should add about 1% to ore recovery from the previous setup.


This team's optimal UPH return is in the stage beginning at the start level of 1501.
It can go a long ever-slowing way past that, so quit and restart when it matters to you.

I had thought that the Miner's Excavator passive - "higher chance of mine shafts" - meant that some floors would have mineshafts instead of enemies. From my data, I now think it increases the chance of the Mineshaft choice when "Choose a Path" occurs by ~8% per Miner from the base 25%. It does not increase the chance that "Choose a Path" happens in the first place (I'd prefer both).
The other options are presumably reduced proportionately. If my 5-Miner team has a 65% chance of Mineshaft, the remaining 35% must be divided evenly between Fairy, Treasure and Warp, so their chances have dropped from 25% to ~12%.

I set "Allow keys to be used for:" to Healing Fairy and Mineshaft.
I set "Chance to pick unknown path when no key used:" to 100%. The Huntress' Leader passive means that only 20% of the "??????" fairies will be bad. This mostly compensates for halved known-good fairies.
That enables a few warps. Average warp length seems to be (4+5+6+7)/4 = 5.5. .35/3*1/4*3/10 = 7/800 chance * 5.5 * 100 = ~4.8 floors warped ahead per 100 levels.
Mine shafts from paths take two levels. 20.9/4 = ~5.2 levels skipped for that.
4.8 + 5.2 = 10 floors skipped per 100 levels is the same as saying there are, on average, only 100/110 = ~90.9 floors landed on IN 100 levels.
Ore deposits mixed in with enemies is ~2%, all harvestable = ~20 per 1000 levels.
Ore from transmutation = 1 per normal level (8 of 10) * .909 = ~727 per 1000 levels.
Ore from ?????? shafts= 7/800 chance * .909 * 4 ores per shaft = ~32 per 1000 levels.
Ore from unlocked shafts = .65*1/4*3/10 * .909 * 4 ores per shaft = ~177 per 1000 levels.
(So, the extra 4 Miners and Huntress? -Only incremental help in ore count.
They do replace lost Fairies and chew up ore quickly to compensate.)

Now, what we really want is Uranium. (Why can't my Miners only harvest U? Union rules?)
Chance of U per shaft (says Reddit) is 1/250 (/4 because only 1 will be U) = 1/1000 per lode.
Suppose the chance of U from transmutation is the same. Ditto lodes starting amidst enemies.
(I finally saw one. My team transmuted the enemy in front, harvested that ore and bypassed the Uranium. Argh!) (Since fixed, BTW.)
Then we should average about 20 + 727 + 32 + 177 = 956 ore lodes per 1000 levels passed.
So we should mine about 1 Uranium Ore in that span.
But I'm actually getting very roughly 3-4 (lows of 0 and a high of 7 in 1000-level spans).
Why? -I'm guessing the team's 220 Ore Find. Ore type supposedly improves with Ore Find.
Danged if I see how to calculate how much it helps in any reasonable time, though.
-Maybe I'm getting 220% more Uranium Ore?
-Maybe "Spawn Rate: 250" DOES refer to lode and not shaft, so the base rate is 4 per 1000 and Ore Find is doing nothing?

Suppose that the chance that a lode is U is 1 in 1000 independently of adjacent lodes. If one lode in a mine shaft is U, then there is .999^3 chance that every other lode there is NOT U ...which leaves ~3 chances in 1000 that another lode COULD be U. ~1 shaft in 250 has U, so I'd just have to watch 250,000 mineshafts and see if ~3 of them have at least 2 green ores to check that this theory is true.
If the chance is 3 times that, I should expect to see one double-U lode in about 10K shafts.
I'm obsessive enough to feel the pull, but not dumb enough to succumb to it and start watching.

Originally posted by Necroshine 4 Aug @ 9:14pm:

How far can your 6 Soda Junkies make it?

I will add the requirement of no levels in the Soda Junky Relic for my run.


Obvious build is 6 Claws is obvious, but I elected to undertake the eructation selection.
Big Burps are slow (and loud), but the build is sturdy. It turns out a sixpack of pre-mastery bottle abusers can go a long way indeed if you have enough essence invested in other useful relics.

DR + Overkill + Fairies FTW!


GearExcept:
Script

Imagine living in a world where the loudest proudest lysenkoists proclaim that carbon dioxide is a weapon of mass destruction.

BUG: A 68:52:35 run will show as 20:52:35 in the dungeon completion screen because the clock wraps.

BUG: If a transmuted ore deposit is formed from an enemy above a natural lode, the team will mine the transmuted ore and incorrectly advance to the next level without collecting the natural deposit. If formed below, both will be collected. EXCEPTION: This once worked correctly when two of my party were asleep from Me Help.

POSSIBLE BUG: I noticed in my earliest, low-level SD2 runs (just before version 1.04) that a Thief could attack twice and get essence both times, even though the second time there was a message saying she had failed.

BUG: On the Adventure Progress pop-up, sometimes the inner halves of the green inventory scroll buttons pass clicks farther inside onto an inventory item. (Also, it'd be nice if inventory wrapped.)

POSSIBLE BUG: 1.) Base relative chance for Uranium/Platinum/Gold Ore is 14/20/35. 2.) Ore Find increases the relative chance of discovering rarer ores. 3.) Actual relative amounts of U/Pt/Au Ores found with Ore Find of 220 are very roughly 1/4/40. 4.) At least one of these statements must be wrong. 5.) I know: I'm probably misunderstanding something.

Feature Request: Scripts: Strongest, Lacks Essence.
Feature Requests: Scripts: Janitor rank, Ore rank.

I feel the rank system is the best way to test for enemy categories.
Rank = Ore/Norm/5Boss/0Boss/00Janitor/000Janitor/0000Boss/..., rather than individual tests such as Is/Isn't Ore, Is/Isn't Janitor, etc..
It's simpler, more consistent and less duplicative. It's sufficient.
It also allows the possibility of subcategories later: Rank = Ore.U.

OTOH, the tests you do have should be more complete.
If I want to reduce the HP of the enemy with the most, so that my AoE units have a chance of finishing the enemy off sooner and with less mana, I currently can't (except by chance).
If I want the Dark Lord to kill an enemy which doesn't have essence when another enemy does (without killing both), the only way to do it is to have him (effectively) pass on his first turn.


Feature Request: Scripts: Pass.
No action, no animation, no time taken.
Replaces time-consuming and useless placeholder actions such as Default Attacks on Ore when the rest of the team is Miners.


Feature Request: Party Saves, to replace and improve Coat Check, Script, Pet and Dungeon Settings saves.

I have several different parties that I run to do different things. I want to try more, but there are only 10 Script save slots, only 1 Gear save slot per Class, et cetera. I would rather not have to make notes for every dang thing I try just because there's no room to keep it.
I would like to save Parties instead, with class, gear and script for each individual position and pet and dungeon settings for the party as a whole. Only inventory and relics would be shared.


Feature Request: Slower Progression.
Feature Request: Difficulty and Progression Speed settings.

I like easy games, but everything went by too fast in SD2.
-Reduce base chance to get an item by 90%. (With a larger Item Find effect so both IF and crafting become important.)
-Reduce essence recovery by 50-80%.
-Have slightly faster enemy HP & ATK growth (9/7 & 1/7 level instead of 7/9 & 1/9? - or more).
-Have stronger ambushes. Let two ambushers attack (on opposite sides and both only the adjacent or opposite position or side, so nobody gets attacked twice). Give the Huntress the abilities to passively go first in an ambush and delay traps a turn, and to actively cancel a negative effect and remove a trap.
-Traps as below.

Games with "width" - more dimensions/more systems - have a steeper learning curve, are more prone to exploits and are harder to balance than games with "height" - more stuff.
I do not think the first 10K of SD2 should be harder or more complicated, just slower.
Introduce complexity by adding interactions later.
Allow higher (or normal) difficulty whenever the player chooses.


Possible Feature: Unique soda coasters - one assignable to each position - which provide or enhance some ability for the unit in that position. Saveable in party save. Unlockable?
Possible Feature: Position tests in scripts. (Place, alongside Enemy/Ally/Self).
Possible Feature: Traps.

I feel that position is a good way to add another dimension to combat.
-A Soda Junky in position #1 wants to turn an enemy, because the Junky in #2 can then backstab it. A Junky in #3 doesn't want to turn an enemy if he has a more damaging attack available, because then it would be facing the wrong way for #4's back attack.
-Some positions might deal more damage (or apply more aid) to other positions.
-Some positions might only attack other specific (melee: adjacent, ranged: across) positions.
-Some abilities might change positions like Big Burp does to facing or effect attacks/buffs do to status. (Suppose, e.g., you have a strong tough enemy that only attacks one position.)
-Some positions might interact with gear, abilities or classes in different ways.
-Traps attack a position.
.........Some timed, mostly instant
.........Consistent with dungeon section
.........Damage the same as enemy damage
...........or half that but causes effect
.........Mostly not simultaneous with ambushes
.........May be more likely in some positions


Feature Request: Player Info: Show actual damage done for possible attack skills (and heal strength for Nurse). Include default attack, critical attack and attacks from equipped gear.
Feature Request: Player Info: Show the chance for "chance of x" on default attacks.


Feature Request: Log upgrade to Pet Barn: more pet levels for pun-ishing/logarithmic increases in cost.
Feature Request: a magic well upgrade/option that automatically reserves enough gold to warp your party to your best level achieved, plus enough to hire an expensive party.
Feature Request: allow the tap + book to turn off Soda Junkies when other classes are enabled. Or go further: have those with booked reservations seated at the tables, while those at the bar or wandering around include more waiting their turn at the tables, Junkies and others (maybe even including locked, story-line or otherwise unavailable classes).
Feature Request: Make Beds (the Tavern reset) free.

A patrons refresh is needed to hire some parties and not others, is needed to hire parties with different classes but not those with the same, becomes ridiculously expensive ...but is free if you quit and restart. Hiring costs should depend on the level and class of those hired, period. (And be recovered on firing, which is usually caused by a unit walking under the cursor while you're clicking on a different target.)


Possible Feature: Relic Override: Temporarily set any relic to a lower level.
Possible Feature: Weekly Challenges
Possible Feature: Challenge Rewards
Possible Feature: Unique Items: E.g., Soda Coasters.
Feature Request: Modding.

Given specified restrictions on everything in a party save plus relics, accomplish some task.
Examples: -Mine a minimum amount of Uranium in a given time.
-How far can you go from 10K with a team with no two of anything?
-Beat a given boss with daggers, no mastery levels and 10K total relic levels.
-Survive ten turns in one battle before killing a given boss.
-Do more effect damage than attack damage in a run.
-Stone every enemy for 1000 floors.

When you can do those things, you have enabled (and might as well allow) modding too.


Feature Request: Craftable unlockable scriptable consumables.
Feature Request: Levelable accessories.


Possible Feature: A second mastery tier + subclasses. Max the base mastery level at 1K or 10K, then have multiple new masteries with their own special skill(s) from that point.
Semi-exception: give Master Soda Junkies a Burpquake skill that damages, reverses and affects all enemies, but have each Junky get his own single effect randomly from those available (slow, stun, burn, etc.) and pick the Junky you want. Grab one of each if there's some bonus or achievement for stacking effects.


Possible Feature: Allow a player to set the end level for a run as well as the start and automatically shut down the game if it's running in the background when the end is reached.
Show the dungeon completion screen when SD2 is next started.
Feature Request: Alert the player when a run finishes.

I'd like a fairly quiet one-time alarm and a "notification banner" or such-like pop-up notice; other people surely have different preferences.


Feature Request: Let us specify Shadow Slicer as such, not just "Strongest Group Attack".

This is my first guide (that I've published). I would appreciate some comments to let me know if there's some way this guide can be improved, or whether it worked as is. That will help anybody else considering it, too.
In particular, if you have a build that collects more essence per hour, I would be happy to learn about it (relic levels, starting floor, etc. constant in the comparison and no large difference in operating level such as a gold find build after floor 300K).