Questions with short answers.

No — there is no address and nothing to deposit into, and every part of the site says so where it applies. What IS live is everything else: the Gauge measures any pool on Base from your browser, the Moment bench reads any ERC-4626 vault’s real price history, the published scan is a real read of the chain, and the contract’s properties execute on a real EVM when you press the button on /yoke. None of that is a mockup and none of it needs a deployment.

To audit one claim the whole category makes, and to hand you the instrument that does it. Every vault advertises a rate. A rate is fees divided by time, and fees do not arrive at a rate — so the division hides the only thing a depositor needs, which is how long they have to hold before the number describes them. That length is measurable, nobody prints it, and on this site you can measure it yourself for any pool or vault you like.

Because the same arithmetic has a hundred years of laboratory work behind it. Magnetise a bar of iron slowly and it does not turn over smoothly; it goes in avalanches four orders of magnitude apart in size, and the smooth curve in the textbook is what you get when the instrument is too slow to hear them. The avalanche-size exponent is exactly 3/2 in the model on this site, which means the simulation can be checked rather than admired.

No, and that is deliberate. base.drpc.org answers the browser preflight with a wildcard origin and serves ten thousand blocks of logs per call, which is the only reason the re-run panel on /crackle can exist at all. Introducing a keyed provider would take the panel with it. The endpoints that were tried and rejected are listed in the header of js/chain.js.

Yes, and it is on the site. The vault census on /moment reads every ERC-4626 vault taking deposits on Base and finds that almost all of them ACCRUE — every interval pays something, so calling their yield a rate is fair. The avalanche argument is about pool fee income and about vaults that harvest it, not about lending. The instrument was built to be able to return that answer, and the page prints whichever one it gets.

It is the numerator of every APY quoted against a pool, so it is the right thing to measure for this argument — but it is the POOL’s income, not one position’s. Fees in a concentrated-liquidity pool accrue only to liquidity that is in range, so a narrower range receives a burstier version of the same stream. Every number here is a floor on what a real position sees, not a ceiling.

It is a fair objection and the scan answers it by publishing scale next to shape. The deepest pool in the set earned 138,101 USDC over the window; the pool with the most swaps earned 589.65 USDC across 55,446 of them. Both are in the table on /crackle with their fee income beside their statistics.

Because it has no vault to quote one for. The census on /moment does print the annualised return of every real vault it read — that is a MEASUREMENT of somebody else's number, and each one carries the window it was taken over and whether that vault accrues or steps, which is exactly the context an advertised rate leaves out. The difference is between reporting a rate and selling one.

A funded key on a chain, which this repository deliberately does not hold. Everything else is ready: the contract compiles, its properties pass, four sabotaged builds confirm the suite bites, and `brand.deployments` is the single place an address goes — put one there and every "not built" marker on the site changes by itself. Until then the honest word for a deployed vault is that there is not one.

The scan is one command. The model verification is another, and it fails the build on nineteen claims, five of which are negative controls that must fail for the suite to be doing anything. The re-run panel on /crackle rebuilds the headline statistics in your own browser on a window that did not exist when this page was written.