hood-bot

Read-only view. It cannot trade, and it opens the database in read-only mode so it cannot corrupt the run.

Use the wallet that actually holds the tokens, not the address fomo shows. fomo reports a user's login wallet, but their Robinhood Chain tokens sit in a separate smart-account contract. For 0xdetweiler the login wallet had zero transfers on this chain. To find the real one: open a token they recently bought on the explorer, go to Holders, and find the wallet whose balance matches their buy amount. The bot reads this list every minute, so a new wallet is followed without a restart. Buys under the floor are logged but not copied: small bets are the trader scouting, and on the account reviewed they lost money.

For the signal the copy path could not take. A leader's buy is only actionable for 30 minutes, over his floor and under the chase ceiling; one missed because the wallet was out of gas is gone for good. This takes it anyway. The bot then attaches whichever followed leader holds the most of that token, so the position mirrors his sells afterwards exactly like an automatic copy. If no followed leader holds it, the buy still happens but there is no mirror exit, and only the ladder and the Close button can end it. The conviction floor, the chase ladder and the market cap band are skipped on purpose, because you have already decided. Everything that protects the position afterwards still applies: it must have a sell route, it must be under the price-impact ceiling, gas must clear the floor, and a halted or full book still refuses. It is queued here and executed by the bot on its next pass, never by this page.

The top row is where the money is. The bottom row is what to do about it. Entry vs his is how far above the leader's own fill we got in, averaged: MOO cost 29 points on that alone, he was -9% and we were -38% on the same token in the same window. Cost drag is fees, slippage and gas against what we deployed; the backtest assumed a 10-15% round trip, so above that the edge is being eaten before the strategy gets a say. Caught is how many of his buys that were ours to take we actually took, ignoring the ones we skip on purpose.

His trades we have seen and not acted on yet. A signal is not dropped the moment it cannot be filled: it stays here and is re-checked every poll until it is either bought or runs out of time. Most often it is waiting for the price to come back towards his entry, because copying him 80% above his fill is taking his trade without his price.

Measured over the window we have watched, counting open positions at today's price as well as closed ones. Scoring only closed trades measures the opposite of a trader who cuts losers and holds winners: the losses finish and count, the winners stay open and do not. On this wallet that read as twelve trades, twelve losses, while he was up on the window overall. Tokens he already held when we started following are excluded from the win rate, because his entry price is unknowable from our data. That count is shown as "partial".

Every token a followed wallet has traded since we started watching. Still holds is read from the token contract, so it is exact. His P&L is realised plus unrealised against his average buy in our window, and shows partial when he was already holding before we started, because then his real entry price is unknowable. A ~ means at least one leg could not be priced at the time of the trade and fell back to today's price.

Every buy and sell as seen on-chain, newest first, and what the bot did about it. "history" is from before we started following, "dust" is an airdrop, "copied" is ours.

The same trader, his Solana wallet. No money is committed here and there is no Solana executor: these rows are recorded so we can measure whether copying him on Solana would be worth funding. He quotes in USDC, so every dollar figure is the amount that actually moved, not an estimate.