Shorts paid for the rally
Over 25 days BTC ran +15.66% into a local high at $79,382. The median funding rate was −0.19%. Short positions financed the move the entire way up.
Between March 30 and April 23, BTC ran from $67,388 to a $79,382 print, the highest price of the window. That's +15.66% over 25 days. The median funding rate on Binance perpetuals across the same window was −0.19%.
Every 8 hours, for 25 days, short positions paid longs. BTC spent 62.3% of the period in negative funding. ETH: 48.7%. SOL: 51.3%.
Price up, funding down. The greed side of the book didn't lead the move. It followed.
The split by symbol
| Symbol | Period PnL | Median funding | Time in neg. funding | Long/short ratio |
|---|---|---|---|---|
| BTC | +15.66% | −0.1921% | 62.3% | 1.09 |
| ETH | +12.85% | +0.0257% | 48.7% | 1.56 |
| SOL | +2.06% | −0.0073% | 51.3% | 2.48 |
| BNB* | +0.29% | 0.0000% | n/a* | 2.14 |
| XRP* | −0.69% | +0.2449% | 34.4% | 2.38 |
| DOGE* | +1.19% | +0.6031% | 12.6% | 2.59 |
The symbols that ran were the ones with the most negative funding. The alts, where funding stayed positive, didn't participate. DOGE's funding median of +0.60% and long/short ratio of 2.59 describe a book where retail is permanently long and permanently paying to be so. Over 2.5 days the coin added 1.19%.
Thirteen of fifteen days
From April 9 to April 23, 13 of 15 days closed with a negative median funding rate. April 17 was the only exception. Over that window BTC moved from 71,769 to 77,943 (+8.6%) and printed the $79,382 local high.
| Date | Median funding | BTC close | Note |
|---|---|---|---|
| 2026-04-09 | −0.0147% | 71,769 | Flip to negative |
| 2026-04-10 | +0.0411% | 72,869 | |
| 2026-04-11 | −0.6755% | 72,967 | Start of negative streak |
| 2026-04-12 | −0.3690% | 70,697 | |
| 2026-04-13 | −0.0234% | 74,541 | |
| 2026-04-14 | −0.5234% | 74,032 | |
| 2026-04-15 | −0.7151% | 74,772 | |
| 2026-04-16 | −0.2841% | 75,017 | |
| 2026-04-17 | +0.2781% | 77,050 | Only positive day |
| 2026-04-18 | −0.7082% | 75,671 | |
| 2026-04-19 | −0.8395% | 73,884 | Deepest negative day |
| 2026-04-20 | −0.4615% | 75,820 | |
| 2026-04-21 | −0.5824% | 76,306 | |
| 2026-04-22 | −0.0749% | 78,164 | High $79,382 printed |
| 2026-04-23 | −0.6696% | 77,943 | Still negative |
The period's peak positive funding day was April 6, +0.55% median, when BTC was at ~$69K. BTC added 15% after that and funding never returned to those levels. The market priced the trend in early and then sat on a persistent short bias the entire remaining push.
Anatomy of one cycle refresh: ETH, April 19, 16:00 UTC
Binance refreshes funding every 8 hours. One refresh on April 19:
| UTC | Price | Funding | OI z-score | Long/short ratio |
|---|---|---|---|---|
| 14:00 | 2336.82 | −1.1438% | +1.38 | 1.99 |
| 15:40 | 2323.00 | −1.1438% | −0.60 | 1.99 |
| 16:00 | 2329.17 | −2.2976% | −1.39 | 1.97 |
| 16:40 | 2321.76 | −2.2976% | −1.70 | 1.98 |
| 18:00 | 2294.99 | −2.2976% | −1.13 | 2.02 |
| 19:40 | 2297.15 | −2.2976% | +0.39 | 2.06 |
Three things happened at 16:00:
- Funding stepped from −1.14% to −2.30% in a single print. A 2x multiplier on the cost of holding a short.
- Open interest z-score flipped from +1.38 to −1.70. A wave of closing activity hit the book.
- Long/short ratio rose from 1.99 to 2.07 while price fell from 2337 to 2295.
Retail kept adding longs into the drop. Price moved against them. Funding penalized the short side that had been correct on direction. The imbalance got worse as more longs loaded the losing side.
BTC took the same event on the same day. At 16:00 UTC, funding printed −1.2276%, the deepest single snapshot over the entire 25-day window. Price dropped from 75,874 to 74,711 over the next 4.5 hours. It was a coordinated squeeze setup across all three majors, and the funding rate promptly reversed on it for the next four days.
The alts were running the opposite playbook
On the majors, negative funding matched rising prices. On the alts, funding stayed positive and prices went nowhere.
Share of 30-minute windows with one-sided liquidations:
| Symbol | Long-only | Short-only | Skew | Period PnL |
|---|---|---|---|---|
| BTC | 8% | 12% | 0.7× (!) | +15.66% |
| ETH | 7% | 8% | ≈1× | +12.85% |
| SOL | 27% | 15% | 1.8× | +2.06% |
| BNB | 29% | 22% | 1.3× | +0.29% |
| DOGE | 31% | 20% | 1.6× | +1.19% |
| XRP | 35% | 14% | 2.5× | −0.69% |
On BTC, short positions were liquidated more often than longs (12% vs 8%), which is what you'd expect on an asset in a strong uptrend. On XRP, the ratio inverts hard: 35% of 30-minute windows were long-only wipeouts against 14% short-only.
XRP was the only symbol that spent more time in the bear regime (38%) than in bull (35%) over a period where BTC ran 15%. Same exchange, same 25 days, a completely different book.
A rally without greed
Fear & Greed index over the period:
| Date | FNG | BTC close |
|---|---|---|
| 2026-03-30 | 8 (extreme fear) | 66,699 |
| 2026-04-06 | 13 | 68,834 |
| 2026-04-15 | 23 | 74,772 |
| 2026-04-20 | 29 | 75,820 |
| 2026-04-22 | 32 | 78,164 (high day) |
| 2026-04-23 | 46 | 77,943 |
FNG never crossed 50. At the local-high print, it read 32, still closer to fear than neutral. If the standard cycle-top reading is 80–95, this move unfolded nowhere near the greed band.
Two independent signals tell the same story: funding shows positioning against the move, and sentiment shows disbelief in the move. Both are the opposite of a distribution top.
Methodology
Every data point in this post comes from a private live monitor that polls Binance perpetuals every ~5 minutes. For each of six symbols it stores price, funding rate, open interest change (30m + 2h), OI 24h z-score, long/short ratio, 30-minute liquidation shares by side, Fear & Greed, and a regime classifier.
For this article: 7,022 snapshots per symbol for BTC/ETH/SOL across the full 25 days, ~720 snapshots for BNB/XRP/DOGE from April 21. Roughly 253,000 data points, ~200 MB of raw JSONL.
One note on the ETH figure. The −2.30% 8-hour rate naively annualizes to roughly −2,518% (×3 payments/day × 365). That's not a real return profile; funding rates don't persist at those levels. But it illustrates the instantaneous cost of holding a short at the peak of the April 19 squeeze.