Graveyard / Listings and unlocks
Shorting ETH on validator exit queue spikes
Does a long Ethereum validator exit queue predict a falling ETH price?
The idea
When many validators want to unstake, a large amount of ETH is about to become sellable. A long exit queue is a scheduled supply of coins, so shorting ETH while the queue is high may pay.
How we tested it
| Data | Daily validator exit queue history. Variant A: queue at or above the 90th percentile of the prior 365 days, with a floor of 50,000 ETH. Variant B: 7-day change in the queue with a z-score of 2 or more. Short ETH, long beta times BTC, hold 7 days. |
|---|---|
| Period | 21 May 2023 to 25 October 2026; holdout from January 2025. |
| Costs | 5 bp per side plus actual funding is the cost used for the weekly tests in this batch. |
| Holdout | Development and holdout, at least 30 events wanted. |
| Control | A random twin; a mirror test (long ETH after a surge) was also run. |
Result
| A: development events / t | 13 / −0.85 |
|---|---|
| A: holdout events / t | 22 / −0.82 |
| A: holdout mean per event | −0.99% |
| A: twin p / deflated Sharpe | 0.66 / 0.04 |
| B: development events / t | 8 / −0.91 |
| B: holdout events / t | 13 / −0.12 |
| Mirror trade | no edge (t about +0.8) |
Why it failed
Shorting after queue surges lost money on average, and the mirror trade shows no edge either. This fits the token unlock result: supply that is announced is already in the price. There are also too few events, since the queue rarely spikes.
What would make it worth retesting
A queue above 1 million ETH, with a multi-week wait. That would be a rare regime and fewer than 30 events.
Past results of a backtest say nothing certain about the future.