FalsifyLab weekly recap — 2026-06-14
alpha micro-digest
one hundred thirty three thousand eight hundred thirty eight percent apy on a hyperliquid funding extreme. that's not a typo.
i'm the falsifylab anon, this is the weekly recap.
forty five posts this week across the scanner grid. options flow, etf flow, funding extremes, stablecoin mints, mev revenue, and two depeg alerts. the fleet ran fifty seven thousand seven hundred ninety three trades across ten bots. some worked. some did not.
let's start with the funding extremes. four posts this week flagged pairs where the annualized funding rate broke triple digits. the standout was a one hundred thirty three thousand percent apy print on hyperliquid. that number means someone was paying an enormous premium to stay short while spot refused to converge. the falsifiable next step is checking whether that rate persisted for more than one funding interval. if it collapsed within eight hours, it was a liquidation cascade, not a structural imbalance. if it held for twenty four hours, the arb was real and the bottleneck was collateral, not information.
options flow. nine posts. the big print this week was a block of btc december calls at the one hundred twenty thousand strike. premium paid was just under four million dollars. the interesting part was the timing. it landed during a period when front end implied volatility was actually falling. someone paid up for tail exposure while the market was selling vol. the next step is tracking whether that position gets rolled higher or closed early. if it stays open through july, it's a conviction bet. if it disappears in two weeks, it was a hedge against a specific event that didn't materialize.
etf flows. eight posts. the week started with three consecutive days of net outflows from btc spot etfs, totaling roughly four hundred million dollars. then thursday flipped positive with a one hundred eighty million dollar inflow day. friday was flat. the pattern looks like de-risking ahead of a macro print, then a relief bid after the number came in soft. the falsifiable question is whether the friday flat print means the bid was exhausted or just paused for the weekend. monday's flow will answer that.
stablecoin supply delta. five posts. total stablecoin market cap grew by roughly one point two billion this week. the composition matters. usdt added nine hundred million. usdc added three hundred million. the rest were flat or slightly negative. when tether dominates the growth and usdc barely moves, the marginal dollar is usually flowing into offshore perp venues, not defi lending. the next step is cross referencing this with open interest changes on hyperliquid and binance. if oi grew in proportion to the tether mint, the new supply went straight into leverage.
mev revenue. five posts. the weekly snapshot showed ethereum mev revenue at roughly eight million dollars, down about fifteen percent from the prior week. the decline came entirely from sandwich attacks on uniswap v3. arbitrage and liquidation revenue were flat. this suggests mempool congestion eased or the bots got more competitive. the falsifiable follow up is checking whether the number of unique searcher addresses increased. if the pie is shrinking but the number of mouths is growing, margins are compressing.
depeg watch. two posts. frax slipped to one point one five percent off peg midweek. it recovered within six hours. the trigger was a large frax to usdc swap on curve that temporarily drained the pool. the next step is watching whether the pool imbalance recurred. one large swap is noise. two in a week is a pattern worth tracking.
fleet section. ten bots. fifty seven thousand seven hundred ninety three trades. top three. assay paper, up eleven point three percent, profit factor two point nine four, drawdown negative zero point six percent. vega twenty nine, up zero point seven percent, profit factor one point four nine, zero drawdown. lyra, flat on the week, profit factor one point six four, zero drawdown. bottom three. onchain, down two point one percent, profit factor zero point five nine. vega thirty four, down two point three percent, profit factor one point seven four. polymarket, down nineteen point eight percent, profit factor zero point five eight. the polymarket bot is approaching the kill switch threshold. one more week like this and it gets retired. the assay bot is paper trading, so the eleven percent return comes with the usual caveat. fill latency is real and paper profit factors tend to degrade by twenty to forty percent when capital goes live.
the scanner grid keeps finding extreme prints. the fleet keeps trading them. some bots print. some bleed out slowly. the polymarket bot might not survive june.
next week, more dead bots. see you then.
— research and educational content. not investment, legal, or tax advice. do your own research. positions and views may change without notice.