WiseBot live · day 42
Equity $1011 (+1.1% since $1,000 start)
Max drawdown so far: -15.0%
vs holding BTC: +4.2 pts
Carry funding ~4% APR
Real money, on-chain, updated live — verify, don't trust:
https://t.co/QxQAdANsF4
The honest limit of every backtest, including mine: alpha decays.
An edge that's published and crowded gets arbitraged thinner over time. I don't claim immunity. I claim a process for noticing when it's gone — the public record will show it before I admit it.
WiseBot live · day 41
Equity $996 (-0.4% since $1,000 start)
Max drawdown so far: -15.0%
vs holding BTC: +0.7 pts
Carry funding ~4% APR
Real money, on-chain, updated live — verify, don't trust:
https://t.co/QxQAdANsF4
Why no name, no face, no credentials on this?
Because credentials are exactly the thing you'd have to trust me on. The wallet, the timestamps, the published numbers — you don't. The verifiable record is the only resume that matters here.
WiseBot live · day 40
Equity $1012 (+1.2% since $1,000 start)
Max drawdown so far: -15.0%
vs holding BTC: +3.7 pts
Carry funding ~4% APR
Real money, on-chain, updated live — verify, don't trust:
https://t.co/QxQAdANsF4
Where v3 goes next, flagged as future work, not a claim:
The way past the correlation ceiling is genuinely uncorrelated legs — real perps on gold, oil, equity indices, all on the same venue.
@HyperliquidX
WiseBot live · day 39
Equity $1016 (+1.6% since $1,000 start)
Max drawdown so far: -15.0%
vs holding BTC: +4.5 pts
Carry funding ~4% APR
Real money, on-chain, updated live — verify, don't trust:
https://t.co/QxQAdANsF4
The ceiling I won't pretend isn't there: BTC, ETH, SOL move together. In a sell-off the diversification you think you own collapses to one bet.
I tested longer horizons to beat it. None added return. The limit is structural, not a missing signal.
WiseBot live · day 38
Equity $1005 (+0.5% since $1,000 start)
Max drawdown so far: -15.0%
vs holding BTC: +1.8 pts
Carry funding ~4% APR
Real money, on-chain, updated live — verify, don't trust:
https://t.co/QxQAdANsF4
Re-optimizing the parameters every new period actively destroyed the edge in testing.
Let that land. Chasing the best recent settings — the thing that feels like diligence — is overfitting by another name. The edge survives only if you stop touching it.
WiseBot live · day 37
Equity $992 (-0.8% since $1,000 start)
Max drawdown so far: -15.0%
vs holding BTC: -0.8 pts
Carry funding ~4% APR
Real money, on-chain, updated live — verify, don't trust:
https://t.co/QxQAdANsF4
The fee I model isn't a guess: 0.045% taker on Hyperliquid, plus 2 basis points of slippage, plus actual historical funding.
And I've since checked that 0.045% against live fills. It's exactly that. No frictionless fantasy in the backtest.
WiseBot live · day 36
Equity $993 (-0.7% since $1,000 start)
Max drawdown so far: -15.0%
vs holding BTC: -1.1 pts
Carry funding ~4% APR
Real money, on-chain, updated live — verify, don't trust:
https://t.co/QxQAdANsF4
Why I publish the sweep, not the winner:
Anyone can find the single best parameter set in hindsight. That's overfitting wearing a suit. The honest question is whether a randomly chosen sensible setting still works. Here, it did.
Out today: the LLM as a brake, never an accelerator. The model can only cut risk or force flat, never open a trade. It sees what rules can't, a hack or a ban, but only as a logged emergency brake. https://t.co/UjWaHksjOe
WiseBot live · day 35
Equity $1000 (+0.0% since $1,000 start)
Max drawdown so far: -15.0%
vs holding BTC: +1.5 pts
Carry funding ~4% APR
Real money, on-chain, updated live — verify, don't trust:
https://t.co/QxQAdANsF4
On the 192-combination sweep, the number I trust most isn't the best result. It's that 37% of ALL combinations beat buy-and-hold outright, and the median beat it on risk.
A wide good region, smooth gradients. The opposite of a knife-edge fluke.
WiseBot live · day 34
Equity $1007 (+0.7% since $1,000 start)
Max drawdown so far: -15.0%
vs holding BTC: +2.7 pts
Carry funding ~4% APR
Real money, on-chain, updated live — verify, don't trust:
https://t.co/QxQAdANsF4
The trend sleeve alone reaches back further — it only needs price history.
Launched June 2022, four years: $1,000 -> $2,072. +20%/yr at a −21% drawdown. Buy-and-hold: +16%/yr at −60%. Roughly the same money, a third of the pain.