What the simulation actually said.

This is not a bug and it is not a bonus we are handing out. It falls directly out of the pool rule: one twenty-eighth of the pool per shift, split pro-rata by ore. The numerator does not care how many rigs showed up; the denominator does.

The consequence is worth saying flatly. Anyone who buys in during the first week and extrapolates that rate is extrapolating a number that structurally cannot hold. The floor filling up is the mechanism working, not the mechanism breaking.

We considered smoothing it — scaling the payout with participation so early and late look alike. We did not, because every version of that we simulated either drained the pool in a month or made the first hundred players subsidise everyone after them. Both are worse than an honest curve.

Draw sets efficiency, but efficiency only decides the payout when grit is expensive relative to what a shift pays. In the simulated economy that condition held about two shifts in five.

The rest of the time Greedy’s raw ore advantage — ×1.30 against ×0.85 — put it ahead on absolute share of the pool, which is what actually pays. Neither trait is the right answer; the grit price is.

This is the clearest illustration of why we did not make one trait strictly better. If Thrifty won 61% of shifts and won them by a wide margin, draw would not be a decision, it would be a tax on reading comprehension.

The cheapest improvement available to a new player is not the press. It is reading which floor their rig’s seam matches and standing on it. That is free, reversible between runs, and worth nearly twice what the first merge is worth.

There is a counterweight, and it is the interesting part. Everybody chasing the matching floor makes that floor busy, which is exactly the condition that hurts a Contrarian rig and helps a Swarm one. The obvious play degrades as more people find it.

We spent longer tuning this pair of numbers than anything else on the site, because it is the one place where the right answer changes depending on what everyone else already did.

We tried 10%, which returns 10×, and it read as more “serious” in early drafts. It also meant that in the busiest simulated shifts the grit burn was large enough to become the dominant cost of playing, which changes what the game is about: from who dug the most efficiently to who could afford to run.

At 6.25% the toll stays a toll. The rig is the cost of playing; grit is the cost of turning it on.

The tradeoff we accepted: a lower burn means less deflationary pressure on $GRIT per shift. We would rather the token be supported by a game people keep playing than by a burn that prices them out of it.

A Nocturne rig and a Swarm rig between them are paid well at night and on a busy floor. Press them together and whichever cycle survives is the only condition you are still betting on.

In runs where the player held a mixed pair and the floor conditions were volatile, holding beat merging about a third of the time — and it beat it by more than merging beat holding in the other two thirds. The press is a bet that you know which condition is coming.

None of this is in the merge dialogue as advice, because it depends on your read of the floor and not on arithmetic. What is in the dialogue is the exact ore-per-token figure for your specific two rigs, before you press.

Contrarian and Swarm are reflexive: their condition is set by the crowd, so any edge they give is competed away as soon as enough people notice it. Steady depends on how calm its stock is, which you can estimate but not control.

Nocturne pays ×1.30 on night and ×0.80 on day, graded by how much of a shift is night: all of 00–06, a third of 06–12 and of 18–24, none of 12–18. That grading is fixed, published, and identical every day. Buy 00–06 every day, take the two thirds-night blocks when they are cheap, let the noon one idle, and the trait pays exactly what the table says.

The honest counterweight: everybody can read this note. If enough Nocturne rigs crowd the night block, the pool for that block is split more ways, and the edge shows up in the split rather than the multiplier. The multiplier is reliable; your share of it is not.

It was there because every site in this category has one, which is a bad reason to collect an address. There is no allowlist to be on and no drop to be early for. When the mint opens, the door is the mint page; after that it is the secondary market.

If a site asks you for a wallet address before there is a contract, ask what it is being written to. Ours had no answer, so we deleted it rather than write a better caption.

A date written on chain can only be pushed later, never brought forward, and anyone can read it without asking us. A date written in a tweet can be anything by Thursday.

So the rule we are holding ourselves to is simple: the opening is written on chain before it is written here. Until that transaction exists, the answer to “when” is that we do not know, and a rig bought today is bought before that day exists.

We are aware this costs us. A named date moves machines and “no date” does not. That is the trade, and we would rather make it than name a day we might miss.

The payout is the whole idea. A rig that pays in points is a rig nobody needs, so the game only exists where tokenized equity actually exists, and today that is a very short list of chains. We took the one where the tokens are real and accepted that the book behind them is not deep yet.

What that costs you, concretely: a claim in stock is a swap, you set the minimum you will accept, and on a quiet day it will not fill. The claim reverts and nothing is lost, but you did not get the stock either. The ETH payout is what actually catches you, and in the simulation it was the path most claims took.

The bet is that the book deepens. It might not. If it does not, this stays a game that settles in ETH with a stock ticker painted on it, and anybody who bought a rig for the equity payout specifically has a right to be annoyed. We would rather they be annoyed now, reading this, than after their first claim.

There is no version of this note that ends with a number going up. That is the point of the note.