Model portfolio value
$104,714.99
+$4,715 (+4.71%) since Aug 3, 2025, starting from $100,000 of virtual capital.
The value, compared with doing nothing
The solid line is the strategy, net of execution costs. Each dashed line is the same capital placed once on day one and never touched again: Aave v3 USDC, An even split, Gauntlet USDC Prime. Together they bracket the strategy rather than flatter it.
How to read this chart
The vertical axis does not start at zero. The two trajectories are close, and a zero-based axis would flatten the very gap the chart is meant to show. What matters is the relative position of the curves, not their height.
Gas is already deducted. Every rebalance costs transactions: 89 moves have been charged since the start, $0.93 in total, or 0.001% of yield per year. On Base that cost is nearly invisible; it is precisely why the strategy runs there.
The benchmarks are not neutral, and not equally fair. All of them hold back the same 10% cash buffer as the strategy, so no gap below is the price of staying liquid. What they keep is one head start: none of them pays gas, and the gas the strategy paid is shown just above. Each answers a different objection:
- Aave v3 USDC. The market's reference rate, and the lowest on the list. It holds back the same cash buffer as the strategy, so the gap is about picking, not about liquidity. Beating it proves little; failing to beat it would settle the question.
- An even split. The only benchmark actually pickable on day one: it needs no knowledge of what came next. Every day it puts exactly the same share of capital to work as the strategy did, so the gap is not the price of holding cash — it is what picking is worth on its own. This is the one the strategy has to beat.
- Gauntlet USDC Prime. The largest Morpho vault on the list — as of today. It holds back the same cash buffer as the strategy, so what is left of the gap is hindsight and concentration: nobody could have named this vault two years ago, and everything sits in it. A ceiling, not a fair opponent.
The chart starts on Aug 3, 2025, and not earlier: that is the first day the vaults clearing the admission rules were numerous enough for the engine to place the capital it targets. Before that, a $100,000 deposit was a large share of the vaults on offer and would have moved the very rate it is meant to observe.
How to read the gap with Gauntlet USDC Prime. Every benchmark here holds back the same 10% cash buffer as the strategy, so nothing below is the price of staying liquid. The vaults the engine held paid 5.28% on average against 5.48% for Gauntlet USDC Prime: selection is worth -0.19% a year. That is a thin edge, and it should be — the strategy spreads across the whole list and caps every position, so it earns close to the best vaults rather than beating them. What it buys instead is the absence of a single vault holding everything, and a name that did not have to be guessed right in advance.
- Strategy, net of gas
- Left on Aave v3 USDC
- Split evenly, same capital deployed
- Left on Gauntlet USDC Prime
Hover the curve to read the amounts, click a date to see what explains the gap at that point.
Where the money has actually been
Each band is a curator, each tick below the axis is a rebalance. A yield is easy to promise; a full allocation trajectory can be verified.
Why group by curator
On Base, nearly every sizeable vault runs on Morpho. Coloring by protocol would give one uniform band and say nothing.
What actually distinguishes these vaults is the team that picks the markets and the accepted collateral — Gauntlet, Steakhouse, Moonwell. That is where risk concentration plays out, so that is what the river shows. The grey band at the top is the share never invested.
- Aave
- Gauntlet
- Fluid
- Steakhouse
- Others
- Buffer
- Rebalance
Current positions
Every vault has a cap it cannot exceed. That is the only real protection against an isolated accident: no single position can take the portfolio down on its own.
What the statuses mean
Healthy: nothing to report in the slow signals. Watch: the engine halves the room it gives the vault without exiting. Exit: liquidated at the next decision, without waiting for the usual delay.
Three things trigger an alert: TVL collapsing over seven days (other depositors leaving), a yield falling away from its own thirty-day average, and an APY far above its peers — treated as a risk, not a windfall.
| Vault | Share / cap | Amount | APY | TVL 7d | Status |
|---|---|---|---|---|---|
Fluid USDCFluid · managed by Fluid | 19.5% / 25% | $20,454 | 5.15%7d: 5.05% | −2%$9M | Healthy |
Moonwell Flagship USDCMorpho · managed by Moonwell | 15.0% / 15% | $15,752 | 5.53%7d: 5.18% | +20%$9M | Healthy |
Steakhouse Prime USDCMorpho · managed by Steakhouse Financial | 12.3% / 20% | $12,849 | 4.36%7d: 4.15% | −1%$229M | Healthy |
Gauntlet USDC PrimeMorpho · managed by Gauntlet | 12.3% / 25% | $12,843 | 4.59%7d: 4.37% | +0%$427M | Healthy |
Steakhouse USDCMorpho · managed by Steakhouse Financial | 12.0% / 20% | $12,529 | 3.42%7d: 3.33% | −5%$184M | Healthy |
Spark USDC VaultMorpho · managed by Spark | 10.2% / 12% | $10,721 | 4.13%7d: 3.93% | +0%$7M | Healthy |
Aave v3 USDCAave · managed by Aave DAO | 8.8% / 30% | $9,181 | 3.38%7d: 3.01% | −29%$23M | Watch
|
Liquidity bufferNot invested, available for withdrawals | 9.9% / target 10% | $10,387 | 0.00% |
Bot journal
18 decisions since Aug 3, 2025. Each mark is a day the engine moved; pick one to read what triggered it.
Why so few marks
The engine waits at least fourteen days between two rebalances, and only moves when the allocation has drifted far enough from its target to be worth the gas. Empty months are the normal state of a disciplined portfolio.
A forced exit ignores that delay: when a vault degrades badly enough, it is liquidated at the next decision. Those are the marks that come in clusters.
What this bot cannot do
It does not dodge exploits. A hack, an oracle manipulation or a depeg happens in one block, in seconds. No bot gets out before: it gets out after, at an already collapsed value. Entire risk teams fail at this too.
It does not chase the highest yield. An APY far above its peers is treated as a risk signal and capped, not as an opportunity. Chasing the top yield works nine times and wipes everything out the tenth.
It does not guarantee you can exit. In a panic, everyone withdraws at once and vaults lock up. That is why part of the capital stays uninvested at all times.
It does not know its cost to the cent. The gas shown is an estimate: 200,000 units per transaction, 1.5 transactions per move, ETH at $3,500 and Base at 0.01 gwei. The backtest also ignores slippage and real withdrawal limits. The order of magnitude holds, not the decimal.
What it can do: react to slow degradations — a yield eroding, a TVL bleeding, volatility rising — and hold a disciplined allocation month after month, which nobody maintains by hand.