What Gold Is Actually Worth: Death-Screen Math
Look at any losing run's final screen and one number stands out: unspent gold. Two hundred, three hundred, sometimes more, sitting on a runner that died carrying four starter Strikes, with a shop three floors back where most of that gold could have become a better deck. Gold you are holding is not power. It is a promise, and death cancels it without a refund.
The way to fix the habit is to price gold in units of what it buys, and the current build gives you two public anchors. First, the shop's card removal service, whose ladder is the game's most stable purchase: under the Ascension 6 rule renamed Inflation in v0.102, a removal starts at 100 gold and each further removal costs 50 more. Second, shop relics, whose prices were cut by a flat 25 gold across the board in v0.103.2, a deliberate buff to relic shopping as a strategy. Every other purchase decision in a shop is a comparison against these two lines, and this guide treats them as the baseline throughout. If you cannot remember both numbers for your current run, you do not have a budget yet. You have a pile of coins.
The third conversion is the least obvious one: HP. Elites trade your health for relics, and gold can trade back, indirectly, by making you strong enough to take the aggressive path, or by buying potions that blunt a boss's kill turn. Seen this way, a good economy is not the biggest number on the HUD. It is the right number at the right shop, where right means assigned to a purchase you can name before you walk in. The potion usage guide covers the belt side of that conversion; this one covers the counter.
This guide covers the whole pipeline as of the v0.111 beta branch: what removal is actually worth, how to read shop stock, when to save for shop-exclusive relics like Membership Card, act-by-act budgeting, the event economy, and how Ascension's Inflation rule reshapes all of it. Every number below is anchored to a documented patch value, and anything the game has not pinned down in public data is described qualitatively rather than invented, because a made-up price is worse than no price.
Card Removal Is the Best Gold Sink in the Game
The single most repeated economy sentence in the community, and on this site's own beginner pages, is that removing a starter Strike or Defend at a shop is usually the best use of gold, ahead of buying cards. It deserves the repetition, because it is the one purchase whose value cannot be duplicated by playing well. Starter cards are the worst cards in your deck, they reappear every shuffle, and no card you buy fixes the hands they ruin.
The arithmetic of thinning is why. A starting deck plus early rewards runs well past twenty cards, and you draw five per hand. Every starter removed raises the concentration of the cards your build actually wants, which is the same effect as adding power without adding cards. The deck-thinning guide covers the theory; the shop's role in it is that removal is the only place gold converts directly into that concentration, at a price you can plan around.
Removal order is a deck-direction question. Remove the starter your engine is replacing: a Strength list built around Whirlwind and Demon Form still wants its Defends for the Knowledge Demon's damage race, so it cuts Strikes; an Exhaust Block engine heading toward Barricade and Body Slam eventually finds raw Defends outclassed by its own block generation, so it cuts Defends. If both halves still feel necessary, cut Strikes first, because in the acts where removal happens, dead defensive draws lose runs more often than weak offensive draws do.
One restriction changes the whole market: Afflictions, the StS2 replacement for Curses, are debuffs attached to your existing cards and cannot be shop-removed at all. That does two things to the economy. It makes starter cards the main legitimate removal targets, since the pool of junk you can actually delete is smaller than it was in StS1, and it makes any event or blessing that pays in Afflictions strictly worse than it looks, because the usual escape hatch, paying gold later, does not exist for them.
Free removal exists outside shops, and a good economy keeps a ledger of it. Amalgamator removes two starters at once, Doors of Light and Dark removes one targeted card, Wellspring's bathe removes one at the price of a temporary curse, and Dense Vegetation trades HP for one. Every free removal you take is a shop removal you no longer need to buy, which is why the event guide treats those nodes as gold. Runners who count both sources end their runs with cleaner decks at the same spend.
Reading the Shop: Pricing Patterns, Shop-Exclusive Relics and Bargains
A shop stocks cards, relics, potions and the removal service. Prices track rarity, which makes the interesting decisions the ones where price and usefulness disagree: a cheap common your deck specifically needs, sitting next to an expensive rare that fits nothing you are doing. Learning to spot that disagreement is most of the skill of shopping.
Start with the shelf that exists nowhere else. Three relics in the current data are shop-only: Membership Card, which halves the price of everything at every shop; Brimstone, +2 Strength per turn at the cost of all enemies gaining 1, a Strength Ironclad's dream; and Alchemical Coffer, which fills every potion slot at once, a one-stop armory before a boss. Because these cannot drop from elites or chests or events, seeing one is not a choice to defer. It is a budget event, and the right response is usually to have, or to immediately start building, the gold for it.
For normal stock, permanence is the pricing signal. Relics act in every remaining fight, so a generically strong relic, Lantern's first-turn energy, Pantograph's 25 HP at every boss start, Mercury Hourglass's chip damage every turn, tends to outvalue a single card of similar price unless that card completes an engine. Cards are bought to fill holes, not to fill space: Defragment for an orb deck after the StS2 rework made it the only stable Focus source, Catalyst as the Poison finisher, Corruption as the engine block of Exhaust Block. If you cannot finish the sentence "I am buying this card because my deck cannot do X without it", you are buying a card, and removal was probably better.
The bargain logic, in one line: compare every item to the removal price first, because removal is the baseline purchase, and only buy past it when the item beats the baseline. The v0.103.2 cut of 25 gold from shop relic prices moved that comparison in favor of relics across the board, which is worth remembering whenever a mid-price relic like Gremlin Horn or Red Mask sits next to a flashy rare card your build does not need. The flashy card is not competing with the relic. It is competing with what the relic's price would otherwise buy, and that arithmetic is usually unglamorous.
Membership Card Runs: When Half Price Compounds
Membership Card is the shop relic that makes future shops cheap: everything costs 50 percent less, everywhere, for the rest of the run. Its own description carries the strategic instruction, the earlier you buy it, the more it pays, because a discount is only worth the purchases it still has time to discount. That sentence is the entire investment thesis.
Run the shape of it. Buy the Card in Act 1 and every later shop in the run effectively doubles your gold: removals, relics, potions, everything, at half price, across a whole remaining campaign. Buy it in Act 3 and it discounts perhaps one visit, which is still often a fair price for the savings, but nothing like the compounding case. This is why the Card should be evaluated against your expected remaining shop count, not against this visit's stock. A mediocre Act 1 shop is still the best place in the run to buy the Card, because the shop's contents stop mattering the moment the discount outlives them. One sequencing note for planning: the Card discounts the removal service too, so on long runs the cheapest way to buy your fourth removal is to have bought a discount card three acts earlier.
Playing a Membership Card run is a discipline of its own. The Card rewards gold income and shop density, so path toward merchant nodes, keep your gold unspent past the removal baseline, and treat event gold, the small chest at Sunken Treasury, the rummage at Trash Heap, as investment income rather than pocket money. The economy inverts: standard advice says convert gold into power promptly, but a Card holder has a documented reason to hold, because every deferred purchase will cost half as much a floor later.
One interaction to know before you commit: Ectoplasm, the Ancient relic, grants +1 energy per turn and makes it impossible to gain gold at all. If you have already taken Ectoplasm, Membership Card is close to dead weight, and if you are holding the Card, an Ectoplasm blessing is priced very differently than it looks on paper. The two artifacts define opposite economies, one where gold is abundant and must be spent well, one where gold simply stops, and the Act-start Ancient choice is the moment to decide which economy you are actually playing.
Relic or Card? The Purchase Priority Ladder
Standing in a shop with 300 gold, use this ladder. The default rung is removal, because the community line is blunt: removing a starter beats buying a card in most decks. Above the default sit exactly two categories of purchase that justify spending the gold instead, and both are specific, never general.
Rung one: a card that completes or rescues your engine. This is the rarest and best purchase in the game, a Defragment landing in an orb deck that lost permanent Focus to the rework, a Catalyst arriving at a Poison Silent missing its finisher, Corruption closing the loop of an Exhaust Block list. Rung two: a build-defining relic, Brimstone for Strength Ironclad, Ice Cream for a deck that wants to bank energy across turns, Pen Nib alongside multi-hit attacks. Both rungs share one test: you can name the fight, or the archetype, that the purchase changes. If the answer is "it is generally good", it is not on the ladder.
Below the default sit the purchases that are usually errors. Generic filler cards bought because they are strong in a vacuum are the classic leak; the card-picking discipline that says skip rewards which do not serve your build applies doubly when the cards cost gold. Overpriced situational relics, and second copies of effects you already own, round out the list. The shop's stock is not your deck's wishlist. It is a market, and markets happily sell you things you should not buy.
A worked comparison makes the ladder concrete. Imagine 250 gold, a removal available, a strong common attack on sale, and Pantograph in the relic slot, with an act boss two floors away and your HP line thin. The removal is the default. The attack is the leak. Pantograph, in this exact situation, is rung two: 25 healed HP at the start of every boss combat is a named, immediate, fight-changing effect, which is what justifies paying past the baseline. Same shop, same gold, different run states, different correct answers: that is the entire skill, and it is why shop guides can only give you the ladder, never the picks.
Save for the Key Relic or Patch the Deck Now?
The recurring fork: you have a few hundred gold, the shop in front of you offers a decent common relic and a removal, and you know that Membership Card, Brimstone or some other run-defining purchase may appear at a later shop. Saving is how you end the run rich and dead. Spending is how you end it poor and alive. The resolution is not a rule but two questions, asked in order.
Question one: is the deck currently winning its fights? If yes, hold. Gold converts best into power when the purchase is decisive rather than incremental, and a deck that is already clearing its path does not need increments. If no, if the next elite or boss exploits a hole your deck demonstrably has, spend now: a Test Subject C10 fight that demands perfect Block in phase two, or an Owl Magistrate Verdict turn you cannot cover, converts an immediate purchase into survival, and survival is the prerequisite for every future bargain. Question two: how many shops remain on your planned path? Saving for a purchase that has no remaining shop to appear in is just losing gold slowly with extra steps.
Timing anchors make the decision easier. Pre-boss shops are the best spend windows in the run, because the next fight is known and its script is documented: Pantograph before any act boss, Alchemical Coffer to fill the belt with potions before a boss's kill turn, a final removal when the deck is otherwise finished. Mid-act shops favor saving, because the act's remaining threats are still varied and a specific purchase has fewer named problems to solve.
The psychological guardrail: keep a floor and treat everything above it as spendable. A sensible floor is one removal's price, held at all times, so that no shop is ever entered without the option of making the baseline purchase. Gold above the floor is free to chase the rung one and rung two buys from the ladder above. Runners who die richest are not the ones who saved for the perfect purchase; they are the ones who never defined what perfect meant, and so never had a reason to stop saving.
Act-by-Act Income and a Working Budget
Where gold comes from in the current build: normal combats early, which pay gold and card rewards and are the reason the Act 1 tips say to fight normals first; events, from Sunken Treasury's small chest to Trash Heap's rummage and the Battleworn Dummy's scaled payout; and the hidden Merchant fight's 300 gold, the largest single payday in the game. Elites, by contrast, pay in relics, not gold. Plan the two tracks separately: the relic track runs through combat, the gold track through hallway fights, events and shops.
A working budget by act, offered as heuristics rather than fixed numbers. Act 1: aim to afford one removal plus one modest purchase by the act boss, funded by hallway combats; since v0.101, floor 6 no longer spawns elites, so a combat-heavy opening is safer than it was at launch. Act 2: front-load your removals, because each removal costs more than the last, and the act's elites, a 200-HP Infested Prism among them, punish underpowered decks immediately. Act 3: convert. Savings have almost no remaining shops to be spent in, and the final boss pool rewards a finished deck over a rich one.
The Ascension 6 rule deserves its own paragraph inside budgeting. Since v0.102, Ascension 6 is named Inflation: shop removal starts at 100 gold and climbs by 50 per removal. On ascending runs this reframes the entire act structure: the first removal is a major purchase, the second is comparable to a relic, and beyond that you are pricing removals against build-defining buys. The consequence is that free removal events, Amalgamator, Doors of Light and Dark, Wellspring's bathe, Dense Vegetation, are worth more on A6 than on any lower level, sometimes more than the elite fights sitting on the same path.
Co-op budgeting differs in one structural way: shops are per-player, so each runner manages an individual economy with no shared wallet, and the hidden Merchant event does not appear in co-op at all, removing the 300-gold payday from co-op planning. Otherwise the same ladder applies, with one added wrinkle: because specific co-op cards pass between players, an engine-completing purchase for one teammate can be a team asset, which slightly raises the value of rung one buys in multiplayer.
Events as an Income Source: Chests, Tickets and the 300-Gold Merchant
The event pool is a shadow economy, and reading it correctly is worth more than any single shop visit. The showcase mistake is Sunken Treasury: the small chest pays a little gold, the large chest pays a lot plus a permanent curse, and the community takes the small chest because the large chest's marginal gold is most likely spent removing the curse it just installed. An income source that immediately generates its own expense is not income, no matter how big the number on the chest is.
Trash Heap's rummage option pays gold plus a unique card, the only event payout denominated in both currencies at once, which makes it the best pure-income node in Act 1b for a deck that wants both. Its sibling option, dig deep, trades HP for a unique relic from the legacy pool, which belongs to the relic track rather than the gold track, and the healthy runner should usually prefer it: relics act every fight, gold acts once.
Wongo's, covered fully in the event decisions guide, is a spending event rather than an income event, but its ticket option behaves like deferred income: one payment now, several random relics delivered after a few fights. Treat the ticket as a bulk discount on relics with a delivery delay priced against Act 2's elite wall, and buy it when you can both afford the price and survive the interval before delivery.
The Merchant fight is the ceiling of event income: 300 gold, every knock-off relic the fake merchant was selling, and the exclusive Merchant's Rug for completionists. The price chain is holding the Foul Potion from Potion Courier across an act, which costs a belt slot, and then winning a boss-strength fight on demand. Battleworn Dummy pays scaled rewards for a three-turn kill, which makes it the one event where burst archetypes, Claw zero-cost above all, convert deck speed directly into gold. Across all of these, the same bookkeeping applies as in shops: event income is only real once it has a named purchase attached to it, and a runner who cannot say what the 300 gold is for has not finished the economy lesson yet.
When Gold Stops Mattering: Ectoplasm and the Act 3 Spend-Down
Ectoplasm is the Ancient relic that states the economy's limits outright: +1 energy every turn, but you can no longer gain gold. Evaluating it forces you to decide what gold was worth in the first place. The energy is a permanent tempo advantage in every future fight, roughly a card's worth of headroom per turn. The gold loss is a handful of purchases across a run. Decks that are already functional, or Ancient boards where the alternatives are weak, should frequently take it; runners planning Membership Card shopping sprees should not, and the two cases are usually distinguishable by Act start. Between the extremes sits the ordinary case: a deck with one real hole, where Ectoplasm's extra energy quietly covers the hole's job better than the gold would have.
The general lesson Ectoplasm teaches is that gold has a half-life. Its value decays as remaining shops shrink: a removal bought on floor 4 improves dozens of fights, while the same removal on floor 40 improves a handful. By Act 3, gold should convert aggressively, into the last removals that finish the deck, into potions for the boss's documented kill turns, or into Alchemical Coffer if it appears and your belt matters for the finale.
A spend-down checklist for the approach to the final boss. First: will any remaining shop appear on my planned path? If yes, hold exactly what that shop's intended purchase costs and nothing more. Second: is the deck finished? If yes, gold is now potion money and nothing else. Third: which boss am I likely to face from the Act 3 pool, Aeonglass, Test Subject C10 or the Queen, and what does its script demand? The Queen's 99-stack debuff turn and C10's perfect-Block phase are both answerable with purchases made in time, and unanswerable with unspent gold carried into the fight.
The mirror image of the death-screen math from the top of this guide: the best runs end broke. Not because spending is virtuous, but because a runner that converted all its gold into deck quality, HP and belt contents, and had nothing left it wanted to buy, was by definition fully equipped for the fights it actually faced. Ectoplasm simply makes that ending mandatory; every other runner has to arrive at it voluntarily.
Ascension Inflation: Economy Habits That Survive A6 and Beyond
Ascension changes combat, but its sharpest economic intervention is Ascension 6, renamed Inflation in v0.102: shop removal starts at 100 gold and each removal costs 50 more than the last. Illustrate the ladder: on an A6 run, four removals across a campaign cost 100, then 150, then 200, then 250, a cumulative 700 gold. That is not a line item. That is the run's entire discretionary budget, and planning around it is what separates ascending economies from flat ones.
The adaptations, in order of impact. First, treat free removal events as premium nodes: Amalgamator's two-starter consolidation alone saves 250 gold of ladder by the time you would otherwise have bought your third removal, which is relic money. Second, move Membership Card up your priority list on any run where it appears early, because halving an inflating service compounds faster than halving a flat one. Third, lean on transform effects, Aroma of Chaos above all, which convert problem cards into new cards without touching the removal ladder at all.
Fourth, rebalance the combat tracks. Elites still pay in relics, and Black Star doubles that income when you own it, so the relic track through elites remains the strongest purchase-free power source on Ascension; but the gold track narrows, since more of every purse goes to inflating removals, which raises the relative value of event gold, chest nodes and the scaled payouts from Battleworn Dummy. On high Ascension the economy is less about accumulating gold and more about never paying retail for anything the event system can hand you free.
The long-run habits, the ones that hold at every difficulty. Enter every shop knowing your baseline purchase, which is removal, and any named target above it. Keep a mental ledger across floors: gold on hand, income expected from the path ahead, purchases planned. Never leave an act with unspent gold when a shop is still on the path, because acts, not floors, are the natural accounting periods of a run, and each boss you pass is a deadline the ledger has to clear. And when a patch shifts prices again, as v0.103.2 did for relics, re-derive the baseline before trusting any habit, because a budget built on last patch's numbers is just nostalgia with arithmetic attached. Runners who internalize the deadline stop dying rich, which was the whole point.