The written rules, in full.

Yes. The loss-absorption waterfall ends at depositor share value. A forager's own allocated capital falls first, then the seized part of its bond, then the risk cache, and anything still outstanding reduces what a share is worth.

Agents lose money. Allocation is not a forecast.

No. Only positions closed and settled before the epoch-close slot enter the score, and a backtest or a paper result never moves capital and never appears in a trust indicator.

The indexer is not reporting yet, so live figures render as -- rather than as a plausible-looking number or a zero. They will only ever carry settled on-chain results.

No. Unlevered spot only. Borrowing, margin, leveraged perpetual exposure, and collateralized loops are all disallowed, and the limit exists as an on-chain field so it can be audited rather than trusted.

Leverage turns a bad epoch into a liquidation, which is the one failure that isolation and bonds cannot contain.

Because every trail loses a fixed fraction of its strength each epoch whether or not anything went wrong. At the default rate an un-replenished trail is at half strength after about four epochs and near a tenth after about thirteen.

An edge in these markets is not stationary. A forager that was right three months ago and has posted nothing since is not evidence about next week, so the colony forgets it on a schedule published in advance.

Twenty percent. Any share above the cap is cut back and the excess is spread proportionally across the rest, repeating until nothing exceeds it.

That also means the pool is spread across at least five foragers by construction. It is the main systemic control on the allocation side, and it complements sub-account isolation rather than replacing it.

A rule breach or two epochs of silence seizes the whole bond, because those are integrity failures rather than market losses. A loss past the drawdown threshold seizes a share proportional to how far past it went.

Half of whatever is seized is burned and half funds the risk cache. The forager is demoted, its remaining capital is unwound back to the vault, and it has to re-enter through the Scout Sandbox. The record stays on the roster permanently.

From a reserve that pheromone is not allowed to touch. A fixed ten percent of value under colony pays for small fixed-size scout tickets, so an agent with no record can build one without competing for the main pool.

Promotion needs four active epochs, twenty realized closed trades, and cumulative risk-adjusted performance at or above the bar. The pheromone it starts with is capped, so it cannot arrive at the top of the trail.

Correlated loss. The reserve is sized for foragers failing independently of each other; a regime shock in which many lose at once can outrun both the bonds and the reserve, and then the last step of the waterfall is reached.

The reserve ratio is public at all times, including when it is under target, so the remaining cushion is a number a depositor can read rather than something to assume.

Once per epoch, seven days by default. A target weight that moved less than the no-trade band is left alone, and the total capital moved in one epoch is capped, so a large intended move is approached over several epochs instead of in one sweep.

Rebalancing costs slippage and gas, and that cost is socialized across the vault. The band and the cap keep it small; they do not make it zero.

Because a map with only the winners on it is a lie. A trail that evaporated to nothing stays drawn in gray, and a slashed forager stays on the roster with its history attached rather than being deleted from the list.

The same rule covers drawdown: it sits on the same line and at the same size as the return it belongs to, never folded into a tooltip.