Demo — not yours · every number is a measured mock, not yours
Points · open mock
100,000 MBRN / month · 33,333 per class · accrues continuously
r=0% · keep $84.10 of $84.10/mo · points 1.00×
r=50% · keep $48.67 of $97.35/mo · points 1.50×Transmuter synced at 25%. The new 50% applies at your next deposit or exit there
no weight yet. run a policy on Defend to earn here
these numbers are estimates and move as others accrue weight · revenue has no distribution window, so sacrifice boosts land the instant revenue arrives and your pending points flush at your next touch · forgone fees boost your own pool
Collateral mock
1.000 BTC
$95,000 · never deployed, never lent
Loan to value 28%
borrow cap 60% · liquidated past 73% · a breach repays to 60%
Headroom
38 h
at the worst 24h fall of the last 90 days (−14%, measured)
Debt
$26,980
of $38,000 borrowed · 71% remains
01 / Throughput mock
The machine is getting faster.
What arrives is already net. The loan’s interest comes out before the yield reaches you. That subtraction shrinks as the debt shrinks. So each delivery is a little larger than the last, from the same venues at the same rates. The machine speeds up by losing overhead.
$10.17
per day, measured across
the last 14 days of deliveries
▲ from $9.27 in your first fortnight · $12.40 at zero debt, if rates hold
vs just holding: +$402 delivered on top of the same BTC · vs selling in January: 0.19 BTC ahead, net of exit cost and tax, both measured
Belt deliveries mock
Delivered, cumulative mock
Every point is a settled delivery — the line can plateau, and does: the flat week in June is the sUSDe cooldown, when nothing arrived. Measured on-chain; last event 2h ago. Benchmarks: holding alone delivers $0 — the whole green line is the surplus; the sold-in-January path is cash interest on the proceeds, net of exit cost and tax.
02 /
What the yield is doing —
mockThe menu previews the intents you did not pick. Nothing switches unless you switch it.Repay
Running
until the debt is gone. The yield pays it off, so you keep the deployed stack as well as the bitcoin. A range, because the harvest rate is venue rates — the same rates every other projection on this page depends on.
Spread
+8.9%
Venues pay 11.9%, the loan costs 3.0%. While that spread is positive, compounding out-earns repaying — but repaying is the leg that cannot lose, and it cuts LTV strictly faster. That is the trade, and it is yours to make. Nothing here picks for you.
Drill this scenario in Builder →03 /
What you have survived
mockNamed after the fact, from measured price history — not from how it felt.Aug 12, 2026
Survived the Aug 12 wick
Bitcoin fell 14.2% in 90 minutes. Your headroom bottomed at 3.1 hours before recovering, and no venue was ever asked for capital. You were drawn at 40% when the same position at 52% would have been recalled — your sizing is why.
price data measured 2026-08-12 · block 21,904,118
Jul 29, 2026
sUSDe cooldown coincided with a 9% slide
The one venue that could not have answered was locked for four days. It never mattered, because Aave and Morpho Vault V2 covered the whole requirement on their own. Two liquid venues did the work of three.
price data measured 2026-07-29 · block 21,712,455
Jun 04, 2026
You paid for a lesson here
A 6% drift crossed the line because the position was drawn at 57% that week. $418 of bitcoin was sold to bring it back to the cap. You reduced the draw four days later and it has not happened since.
price data measured 2026-06-04 · block 21,204,901
04 /
Is it luck?
mockScored only on calls that have actually resolved.Better than a coin flip
+31%
across 46 resolved calls
Brier 0.172
Calls by confidence band
Gold tick is where a perfectly calibrated person lands. Bar is how often you were actually right.
05 /
This week
mockDitto writes one of these every Sunday. It is the card you can hand to someone else.Week 31 · Aug 3–10
Twenty-three deliveries, and you held through Thursday at 41% LTV.
What this page does not show
This page does not show a portfolio value. It does not show a percentage return. It does not show an unrealized profit. These numbers change when the market changes. They do not show what you did.
This page shows only events that occurred. It shows the yield that arrived, the debt that you paid, the shocks that the position survived, and the calls that resolved.
The route data has a measurement date. We measured 1,245 positions and 25 routes in August 2026. We measure them again on a schedule. Each event on this page shows the date of its data. Do not trust a number that has no measurement date.