starling
the lab
Replay it.
Seven hundred real launches, a random sample of the 8,135 we replayed from 4 days of Pons on Robinhood Chain. Set the size of the flock, the creator tax, and how late a normal buyer would have been. Nothing here is made up: the engine is the one the contract tests are pinned to.
Settings
This launch
0%
0%
Across the sample:
| mean | median | lose >20% | lose >50% | worst | best |
|---|
The flock member did better than the late buyer on of these launches.
flock memberlate buyershare of launches by return, in %
How to read thisReturns are what a buyer of the given kind would have made on each launch by selling on the curve at the exit time, with the curve fee and the creator tax paid both ways. A flock exits as a group, so it also pays its own price impact on the way out. The launches are a random sample of a four-day snapshot, ETH pair, default launch configuration. Some launches never trade: they are included, and they cost a flock member the fees. Graduated tokens (about 0.3%) leave the curve and are not modelled beyond that point. This is a replay, not a forecast.