How to Read Options Skew and the Put/Call Ratio

Implied volatility is one number. But an option chain has dozens of strikes, and each one has its own implied volatility. Line them up and they don't sit flat. The out-of-the-money puts usually price at one level and the out-of-the-money calls at another. The difference is the skew, and it's the closest thing the options market has to a mood.

The way it's usually read: puts dearer than calls means people are paying up for protection, so they're scared; calls dearer means they're paying up to chase, so they're greedy. Stock index options have been put-skewed on almost every day since the 1987 crash. This article asks what bitcoin's skew has done, and whether the mood it reports told you anything about what came next.

What the number actually measures

Take a put and a call that are each about 25% likely to finish in the money — a 25-delta put below the price and a 25-delta call above it. Both are roughly the same distance from where bitcoin is trading. If the market thought a fall and a rise of that size were equally likely, they'd carry the same implied volatility. The skew is the put's implied volatility minus the call's, in points. At +1.8 today, a 25-delta put is priced at 36 and a 25-delta call at 34: the market is paying a little more to insure the downside. On 17 June 2022 the gap was 38 points — puts at 126, calls at 88.

Nobody publishes this number's history, so this article rebuilds it. Every strike of every monthly and quarterly bitcoin and ether option on Deribit since 2021 — 5,576 bitcoin contracts and 3,674 ether contracts — is solved for its implied volatility on each day it traded, the 25-delta point is read off each side, and the two expiries either side of thirty days are blended into one reading. The method section has the details. The put/call ratio is simpler: the day's put contracts traded divided by call contracts traded.

Finding 1: bitcoin has no permanent put skew

The S&P 500's skew has been positive on nearly every trading day for almost forty years: index puts are always dearer than index calls, because the people who own stocks buy protection and the people who sell it want paying. Bitcoin's is a coin flip. Across 2,085 days its median skew is −0.9 points — calls dearer — and it was on the put side 45% of the time. Ether's median is +1.0, put side 56% of the time. Neither has a resting shape.

What they have instead is regimes. Bitcoin was put-skewed on 95% of days in 2022 and on 8% of days in 2024. Smooth the series and the sign holds for months at a time: call-skewed from 16 February 2023 to 20 September 2025, 948 days, and put-skewed since 21 September 2025 — 361 days and counting. The last flip came three weeks before the October 2025 liquidation cascade.

Read that last column carefully. From 2022 on, the sign matched the trend it sat in: puts dearer through the 2022 bear (−31%), calls dearer through the 2023–2025 bull (+390%), puts dearer through the fall since (−32%). In 2021 it did the opposite: calls dearer while bitcoin fell 29% into the summer, puts dearer while it rallied 22% out of it. The skew describes who's positioned which way. It doesn't say who's right.

Finding 2: the most lopsided days marked turns, not continuations

If a put skew is fear, the most fearful day in the record should have been followed by the fall it feared. It wasn't. The record is +38 points on 17 June 2022, with bitcoin at $19,213.5. The low of the 2022 bear came the next day. A month later bitcoin was +16%.

The FTX week is next, at +37 points on 11 November 2022, and a month later bitcoin was +1%. Take every day with a skew above +10 — 174 of them, 68% in 2022 — and the next thirty days were up 63% of the time, a mean +2.9%. That is a better month than the average day in the sample, not a worse one. Ether's says the same: skew above +10 was followed by an up month 54% of the time, mean +4.5%. The one big exception is the Luna week — ether's record +47 on 11 May 2022 was followed by another 12% down, because that crash had a second leg.

The other side is messier. The record call skew, −64 points on 13 January 2021, came halfway up the 2021 mania and the next month was +23%: greed was right. The −38 of 11 April 2021 was the top, and the −23 of 9 May 2021 was three days before the crash: greed was wrong. Days below −10 were followed by an up month 55% of the time — a coin flip with a big mean (+6.4%) that is mostly 2021. An extreme on either side is a crowded trade. What it is not is a forecast.

Finding 3: the skew did not warn

If protection gets bid before a fall, the skew should have been on the put side the evening before the big down days. Take the twelve biggest one-day falls since 2021 and look at the skew the day before each.

Seven of the twelve began from a call-skewed market. The 10 January 2021 fall of 16% started with the skew at −21 — calls 21 points dearer than puts. The 4 August 2024 fall of 13%, the biggest since 2022, started from a skew of exactly zero. The same holds over a month: of the eight worst thirty-day windows in the record, four began from a call skew, including the worst of all — 43% from 8 May 2021, which began at −16.

Put a number on it. After a reading below −10 the worst single day in the next thirty averaged −8.3%. After a reading above +10 it averaged −6.4%. The downside was biggest exactly when the market was paying least for it. That's not a paradox once you remember what the skew is: the price of protection is set by the people who want it, and they want it after the fall, not before.

Finding 4: it moves with price, every single year

Week 88 found that DVOL changed sides — it rose on falls in 2021–22, rose on rallies in 2023–24, and rose on falls again since. The skew never did. Rank each day's change in skew against the day's return: the correlation is −0.29 over the whole sample and between −0.37 and −0.22 in every year. Bitcoin down, puts bid; bitcoin up, calls bid. Six years, no exceptions.

That is why the skew lags rather than leads. It is the market's reaction to the last move, priced into the next month. The ten biggest one-day jumps in put skew include 16 June 2022 (+19 points on a −3% day, the day before the record) — but also 17 January 2021 and 7 January 2021, both jumps of about twenty points on days bitcoin ROSE 4% and 6%, because a jump of twenty points that lands at −29 is just the mania unwinding by a day. Read the change against the level, not on its own.

Finding 5: the put/call ratio is not about fear either

The put/call ratio is the other number sold as a fear gauge: more puts trading than calls, the story goes, means the crowd is bracing. On bitcoin the crowd has never braced. Calls out-traded puts in every year of the sample. The median day's ratio is 0.66, and puts were 40% of all contracts traded. Put share peaked at 48% in 2022 and bottomed at 32% in 2023, which is the regime table again from the other side.

The version worth reading is the open-interest version, on a live chain, because it tells you where the positions are rather than what traded today. Which brings us to the chain itself.

How to read the chain today

On 17 September 2026 Deribit carried 900 bitcoin option contracts with 281,693 BTC of open interest in calls and 156,057 in puts — a put/call ratio of 0.55 — against a price of $76,452. Ether: 1,300,743 ETH of calls, 647,780 of puts, ratio 0.50. Here is where bitcoin's positions sit.

• Start with the biggest expiry. 25 September 2026 holds 187,241 BTC of open interest, 43% of the whole chain. What happens to price into that Friday matters more than any other date on the board.
• Find the walls. The strike with the most open interest is $70,000 — 14,933 calls and 13,967 puts, split almost evenly, which marks it as the level the market is arguing about. Above the price the calls stack up at $80,000; below it the puts stack up at $60,000. Those are where dealers' hedging is thickest, and where price tends to slow.
• Read the ratio by expiry, not just in total. The 18 September 2026 weekly carries a put/call of 0.79 against 0.53 on the big quarterly: short-dated protection is being bought, long-dated upside is being held.
• Know the max-pain strike, and know what it isn't. For 25 September 2026 it is $72,000: the price at which option holders on that expiry would collect the least. It's a fact about where the positions are, not a magnet. Price pinned near it on some expiries and ignored it on others.
• Then check the skew. A chain with puts stacked below and calls stacked above is normal. A skew of +1.8 on that chain says the puts are being paid for slightly more than the calls — the 64%th percentile of the record. That's a mildly nervous market, not a frightened one.

How to read it yourself

• Read the sign as a regime, not a signal. A month of put skew after two years of call skew is the market changing its mind about which side needs insuring — the last flip was 21 September 2025. It tells you who's positioned where; it hasn't told you which way price goes next.
• Treat an extreme as a crowded trade. The six most put-skewed days all came in the middle of a selloff, four of them were followed by an up month, and days above +10 were followed by an up month 63% of the time. The most call-skewed days were mania, and two of them came within days of a top. Fade the crowd's price for insurance, not its direction.
• Don't wait for the skew to warn you. seven of the twelve worst days began from a call-skewed market, and the worst single day after a call-skewed extreme averaged −8.3%. Protection is cheapest before it's needed. That's the whole reason to buy it then.
• Read the put/call ratio on open interest, by expiry, on the chain — not the daily volume figure. The volume ratio has never been above 1 for a year on bitcoin and tracks the same regime the skew does.
• Put the two numbers together. A put skew with puts stacked below the price is a hedged market. A call skew with calls stacked above it — January 2021 — is a market that has stopped hedging.

Method, and what this cannot tell you

Each day's skew is reconstructed from Deribit's daily option closes rather than read from a published index, because no exchange publishes a skew history. For every monthly and quarterly expiry with between seven and sixty days left, every strike that traded that day is solved for its implied volatility against the perpetual's 08:00 UTC price, then for its delta. The 25-delta put and the 25-delta call are read off the out-of-the-money strikes on each side by interpolation, and the two expiries either side of thirty days are blended to a thirty-day figure. Skew is the put's implied volatility minus the call's, in points. The put/call ratio is put contracts traded over call contracts traded on the same expiries. Forward returns use the perpetual's close thirty days later; a regime is a run of days on which the trailing sixty-day median skew held one sign for at least forty-five days.

• The skew is reconstructed from last-trade prices, not from the exchange's mark. A strike that last traded hours before the daily close carries a stale price, and on thin days in 2021 the 25-delta point is read from a handful of strikes. Deribit's own risk-reversal, if it published one, would be smoother; the direction and the regimes would be the same.
• Monthly and quarterly expiries only. Weekly and daily options are not in the sample, so the put/call ratio here is the ratio on the expiries that carry most of the open interest, not the whole exchange. The live chain snapshot does include every expiry.
• Thirty-day tenor, one delta. The skew at a week's tenor moves faster and further, and the skew at 10-delta is the tail's own price; both are separate numbers this study does not read.
• One venue. Deribit carries most of the crypto options market but not all of it; the CME, OKX and Binance chains have their own skews.
• Regimes are read off a trailing median, so a regime's start date is known only in hindsight — about a month after the sign actually changed. Nothing here says the sign will keep tracking the trend; in 2021 it did not.
• Ether's series starts on 15 March 2021, when its listed strikes first bracketed the 25-delta point on both sides in the data; bitcoin's starts on 1 January 2021.