Don't Ask What a Sink Gives. Ask What It Eats.

There's a clean little framework making the rounds for idle-game economies. Three sinks — Progression, Engagement, Prestige — and for each one a row that says what it absorbs: permanent upgrades, refills and boosts, cosmetic items. It's a good starting map. But the word "absorbs" is doing something sneaky, and if you don't catch it, you'll mis-diagnose every economy you ever tune.

Here's the problem. A sink, by definition, is a thing that destroys a resource. That's the whole point of the word — it's the opposite of a source, which mints resources into the economy. So when a framework tells you a sink "absorbs permanent upgrades," it's pointing at the wrong end of the pipe. Permanent upgrades aren't what gets destroyed. They're what gets handed back to you. They're the output. The thing that actually disappears — the thing the sink eats — is whatever you paid to get them.

You already sensed this. "For prestige you pay levels. For playing you pay energy." That instinct is the correct one, and it's worth following all the way down, because it changes the framework from a labeling exercise into a diagnostic tool.

A sink is a converter, not a container

Strip an idle economy down and every sink is the same shape: it takes one resource in, deletes it, and emits a different resource out.

[resource paid]  →  ( SINK )  →  [thing you receive]
   destroyed                          created

The "Absorbs" column in the original framework describes the right-hand side — the reward. To make the framework operational you have to fill in the left-hand side: what resource flows in, and where did the player get it?

Once you ask that, you notice idle games don't run on one currency. They run on a small stack of fundamentally different resources, and each sink is tuned to eat a specific one. Mixing them up is how economies quietly break.

The four resources almost every idle game actually trades in:

And underneath all of them sits money (IAP), which can buy a shortcut into any of the four. That's the real reason monetization touches every sink at once.

Now let's re-read the three sinks with the left-hand side filled in.

Progression sink: you pay time wearing a currency costume

The framework says the progression sink absorbs "permanent upgrades" and tunes to player level. Reframe it:

• You receive: permanent power (higher production, faster numbers).

• You pay: soft currency.

• But soft currency is just time in disguise. You earned that gold by waiting — actively or idly — for production to tick. So the resource the progression sink truly eats is time, laundered through a currency so it feels like a choice instead of a wait.

This is the core treadmill, and it's a closed loop: you spend time to make currency, then spend currency to make time more productive, which lets you make currency faster. Cookie Clicker's buildings, AdVenture Capitalist's business upgrades, Clicker Heroes' hero levels — all the same machine. The exponential cost curve ("tunes to player level") exists precisely to keep the time-to-next-purchase roughly constant as your numbers explode. That constant interval is the pacing of the game.

So when the framework says it fails when content is maxed, restate it in resource terms: the sink ran out of depth while its source kept minting. Production never stops. If there's nothing left to buy, the soft currency has nowhere to go, it inflates into meaninglessness, and the number going up stops meaning anything. The symptom — "upgrade buys flatline" — isn't really about upgrades. It's a source/sink imbalance: the faucet is still open and the drain is capped.

The fix is never "add more upgrades" reflexively. It's "re-balance the source against the sink," which usually means the next sink in line needs to take over absorbing the overflow. Which brings us to the genuinely interesting one.

Prestige sink: the framework drew the wrong prestige

This is where your instinct exposes a real ambiguity. The word "prestige" means two completely different mechanics in this genre, and the framework only drew one of them.

Reading A — the status sink (what the diagram shows). Absorbs cosmetic items, tunes to player ego, symptom is "top 5% IAP drops." Here:

• You receive: status — a skin, a badge, a leaderboard rank, a visible flex.

• You pay: money (or premium currency, which is money).

• The resource being destroyed is cash, and the thing you get back has no mechanical function at all. Its only value is that other people can see it. That's why it "tunes to player ego" and why its ceiling is your top spenders: a status sink only works as long as there's a more expensive thing worth being seen owning. When the top 5% have bought the top item, the sink is full, and their spend drops because there's nothing left to signal with.

Reading B — the prestige reset sink (what you meant by "you pay levels"). This is the classic idle-game ascension loop — Cookie Clicker's "ascend" into Heavenly Chips, AdVenture Capitalist's Angel Investors, Clicker Heroes' Hero Souls, Egg Inc's soul eggs. Here:

• You receive: a permanent meta-multiplier that makes your next run faster.

• You pay: your entire accumulated progress. You hard-reset. Every level, every coin, gone.

• The resource being destroyed is accumulated progress itself — the most expensive thing in the game, because it's everything you've built. Nothing else in an idle economy asks you to pay that.

These are not the same sink. One eats money and pays you status. The other eats your whole run and pays you a multiplier. They tune to different motivations (ego vs. the dopamine of starting over faster), they fail for different reasons, and they have different symptoms. A framework that files both under "prestige" will send you tuning a cosmetic store when your actual problem is that nobody wants to ascend — or vice versa.

If I were redrawing the framework I'd split these into a Status sink (pays in money, returns visibility) and a Reset sink (pays in accumulated progress, returns a meta-multiplier). The reset sink is arguably the most important mechanic in the entire genre, and the original framework doesn't have a box for it.

Engagement sink: you pay energy, which is a clock with two faucets

The framework says this sink absorbs "refills and boosts" and tunes to session frequency. Your phrasing was exact: "for playing you pay energy." Energy is the resource being eaten here, and energy is special because it has two sources feeding it:

• Time refills it for free, slowly. This is the leash. The regen rate is literally what "tunes to session frequency" — set the leash short and players come back hourly; set it long and they come back daily. You are not designing an energy bar, you are designing a return cadence.

• Money refills it instantly. This is the monetization hook attached to the leash.

So the engagement sink quietly eats two different resources depending on the player: it eats time from the patient majority and money from the impatient minority, and "refills and boosts" is just the menu where time-payers and money-payers diverge. The boosts do the same thing from the other direction — they let you spend money to compress time, collapsing hours of idle production into one tap.

It fails when sessions shorten because the energy economy got mistuned: if a session's worth of energy drains in thirty seconds, there's nothing to do when the player returns, so the return stops being worth it. The "D10–D14 slide" symptom is the mid-funnel signature of this — players survive the first week on novelty, then drop out exactly when the energy leash has to carry the retention on its own and isn't pulling hard enough (or is pulling too hard and feels like a chore). The window is the author's chosen diagnostic, not a universal constant, but the logic holds: this sink is what's load-bearing in the second week, after onboarding stops doing the work.

The real lesson: "Fails when" and "Symptom" are always input-side problems

Look back at all three. Every failure mode in the framework is described in terms of the output — buys flatline, sessions shorten, top spenders cap. But every actual cause is on the input side:

• "Content is maxed" → the soft-currency sink ran dry while its time-source kept producing. Source/sink imbalance.

• "Sessions shorten" → the energy sink's regen rate is mistuned against session length. Conversion-rate mistuning.

• "Top spenders cap" → the status sink ran out of more-expensive things to signal with. The resource lost its meaning at the top end.

These are the only three ways any sink ever fails:

1. The source outpaces the sink (inflation — too much resource minted, not enough destroyed).

2. The conversion rate is wrong (the price is too high or too low for the pacing you want).

3. The resource stops meaning anything (you maxed the depth, so paying it no longer feels like a cost).

You cannot see any of these by looking at what a sink gives. You can only see them by tracing what it eats, where that resource is sourced, and how fast the source fills relative to how fast the sink drains.

The framework, redrawn

Here's the same three (well, four) sinks, with the left-hand side filled in:

Four resources, four prices, four distinct ways to break. The original three-box version collapses the two prestiges into one and never names a single price. That's the gap.

How to actually use this

When you sit down to balance an idle economy, the question is never "is this sink good?" It's:

• What resource does this sink destroy, and what's its source? Draw the faucet and the drain. A sink you can't pair with a source can't be balanced — it can only inflate or starve.

• What's the conversion rate, and what pacing does it imply? Cost curves aren't flavor. They are the clock of the game.

• When this sink fails, which of the three failure modes is it? Inflation, mistuned rate, or exhausted meaning — the fix is completely different for each.

This is exactly the kind of thing worth modeling rather than guessing at, because the failure modes are all about rates over time — how fast a source fills versus how fast a sink drains across a 14-day curve — and that's invisible in a spreadsheet snapshot. Wiring sources to sinks as nodes and watching the resource balance move is the difference between "I think the gold curve feels off around day 10" and "the production source out-mints the upgrade sink by 1.4× after the third tier, here's the exact tick it goes runaway." (That's the whole reason node-based economy simulation exists — but that's a separate post.)

The one-line takeaway, and the correction to the framework: a sink is defined by its intake, not its output. Don't ask what it gives. Ask what it eats — and where that came from.