BTC Options: Dissecting Volatility Trends

BTC OPTIONS
Dissecting Volatility Trends

Finding Edge in Four years
of Volatility Regimes
collectively written by
Greg Fabio Tony Euan Samneet

Magadini Bassani Stewart Sinclair Chepal
Special Thanks to
ABOUT THE AUTHORS
Greg Magadini, CFA
Greg Magadini is the Director of Derivatives at Amberdata. Previously, he co-founded
Genesis Volatility (later acquired by Amberdata). Greg Magadini started his career as a
proprietary trader for DRW and Chopper Trading in Chicago IL. Greg has nearly 15-years
of options trading experience and has been active in the crypto currency space for nearly
10-years.
Fabio Bassani
Fabio Bassani is an option trader with several years of experience in the crypto volatility
space. Bitcoin enthusiast since 2013, he left his corporate job in 2019 to devote himself
full-time to the sector. He has a degree in Economics-Statistics and has always been
passionate about finance and data science. He currently works at Amberdata developing
the tools which help customers in navigating cryptocurrency derivatives trading.
Tony Stewart
Tony Stewart is a professional derivatives trader with over 25-years of experience,
specializing in equity index options and managing buy-side and sell-side institutional
desks. The last 5 years have been focused on the digital asset space, primarily trading
options on centralized and decentralized platforms. He is an engaged advisor to funds,
protocols and data analytics and regularly writes for Deribit Insights.
Euan Sinclair
Euan Sinclair is an option trader with over 25-years of professional trading experience.
He has traded options on indices, stocks, commodities and interest rate products.
He currently trades equity options at Bluefin Capital Management. He holds a PhD in
theoretical physics from the University of Bristol and has written three books: “Volatility
Trading”, “Option Trading” and “Positional Option Trading”, all published by Wiley. He is a
member of the editorial board of the Journal of Investment Strategies - a publication of
Risk Journals, and has written several papers on options, volatility and bet sizing.
Samneet Chepal, CFA

Samneet Chepal is a quantitative researcher/trader with LedgerPrime, a multi-strategy
crypto hedge-fund. He spends a lot of time researching the crypto options market and
enjoys diving deep into various niche topics ranging from DeFi to MEV. He has a keen
interest in advising crypto startups and actively works alongside founders to help them
scale. Samneet received his Master’s of Data Science from the University of British
Columbia and is a CFA Charterholder.
TABLE OF CONTENTS
Forward. . ........................................................................................................................04
Introduction. . ................................................................................................................05
ATM Term Structure . . ................................................................................................06
Spot/Vol. Dynamics.. ..................................................................................................09
Risk-Reversal Skew. . .......................................................................................................10
Variance Risk-Premium...................................................................................................13
Futures Basis. . .....................................................................................................................16
Volume and Open Interest............................................................................................17
Backtesting Strategies:
Harvesting Risk-Reversal Premium. . ................................................................19
Backtesting Strategies:
Systematic Volatility Trading. . .....................................................................30
amberdata.io/contact-us | PAGE 3
Forward
The nascent crypto asset class is only 15-years old
and has presented investors with unique challenges
and opportunities with respect to valuation.
Deribit traded its first BTC options in November 2016, making the crypto space even richer with
opportunities by layering-in tradable volatility.
We explore big questions that remain in crypto volatility by dissecting the volatility surface and its
associated regimes.
Although the crypto volatility space being new brings a unique set of challenges, it also enables astute
investors to become skilled and competitive in the space. No one has a substantial head-start, so to speak.
We are committed to being pioneers in crypto volatility analytics and to helping our customers find money-
making opportunities. As such, Amberdata Derivatives will make all the following analysis available to our
customers.
Thank you for your continued trust and confidence in Amberdata.

INTRODUCTION
Bitcoin Spot Prices
bitcoin spot prices
70k
60k
50k
40k
$
30k
20k
10k
Jul 2019 Jan 2020 Jul 2020 Jan 2021 Jul 2021 Jan 2022 Jul 2022 Jan 2023
When Deribit first launched its exchange there wasn’t 2021: That year started with a bullish-breakout. In
an ERC-20 standard built on Ethereum yet, nor was January of 2021, we saw extreme positioning in the
there a reliable stablecoin market. options market as BTC quickly slid through new all-
time highs day after day. Extreme positioning then
This led to the creation of an “inverse” options and fu- leaked into the futures basis during April 2021, only to
tures exchange, meaning that BTC options and futures completely collapse in May 2021 after a nasty market
are collateralized by BTC itself and cash-settlement correction.
also occurs in BTC (AKA “coin”).
2022: That year can be categorized into a bear market
From 2019 through 2022, we’ve seen a lot of different year as spot prices retraced from their November 2021
regimes and investment cycles around BTC and the peak. The fundamental picture in crypto became bleak
associated options trading on Deribit. and quickly escalated as LUNA imploded, 3AC ran into
trouble and darling exchange FTX turned out to be
2019: This was the tail-end of the 2018 bear market.
fraudulent, broke and actively traded against its own
The spot market was mostly quiet and a lot of enthu-
clients (Think 3AC’s book).
siasm for BTC investing had faded away. BTC hit its
bear market low in December 2018, then prices con- All taken together, we have a subject dataset with
solidated for most of the first half of 2019. Only later in a multitude of spot cycles and associated volatility
the year did BTC spot prices rally to $10k, a major level regimes.
of spot traders.
Our analysis starts on April 1st 2019 and concludes on
2020: In 2020 we had two significant events. First December 31st, 2022.
March 2020 had the Covid meltdown. This was a
scary sell-off in spot prices that caused a drop of 50%,
accompanied by volatility which exploded higher.
Prices then consolidated back around the $10k level
for much of the year. Volatility hit extreme lows in the
summer of 2020 only to end the year with a massive
bull-market breakout.
ATM Term Structure
ATM Constant Volatility
ATM Constant Volatility
200
150
iv
100
50
atm7 atm30 atm60 atm90 atm180
The term structure is one of the most significant mea-

Term Structure State % Time surements of the volatility surface. The term structure
is essentially the “spine” of the volatility surface and
has significant importance because it deals with ATM
Contango 77.5% (at-the-money) options, which have the most extrinsic
value and “optionality”, so to speak.
Backwardation 22.5% Using our subject data set, we can see that the term
structure is in Contango almost 77.5% of the time and
rarely in Backwardation.
Backwardation occurs in reaction to volatility

“shocks” such as March 2020.
PAGE 6 | BTC OPTIONS: DISSECTING VOLATILITY TRENDS

Consecutive hours of backwardated term structure
Consecutive hours of backwardated term structure
400
300
hours
200
100
Since Backwardation is the abnormal state of affairs,

we have calculated the time spent in Backwardation. Mean 85.7371
Volatility traders can use the “time-to-live” (TTL)
Std 93.4913
measurements to estimate when vol bets should
revert into Contango. Min 1.0000
25% 12.0000
The mean TTL of Backwardation in the Amberdata
50% 48.0000
dataset is around 85 hours.
75% 135.0000
Max 417.0000
(Hours in Backwardation statistics)
Richness (+premium/-discount) of term structure
Richeness (+premium/-discount) of term structure (+backwardation/-contango)
(+ backwardation/-Contango)
1.6
1.4
1.2
0.8
Another important measurement of the term Using the term structure levels enables us to quantify
structure is the relative “level” of the Contango or how extended the term structure pricing currently is,
Backwardation shape. at any point in time.
A reading of 1.00 would be a perfectly flat term Notice that March 2020, May 2021, June 2022 and
structure - as measured by our method - while readings November 2022, all represent “vol shocks” that blind-
below/above represent Contango/Backwardation sided the market but peaked around the 1.45 level.
respectively.
Richeness of term structure box plot

Richness of term structure box plot
contango backwardation
1.6
1.4
1.2
0.8
term_structure_richness term_structure_richness
We can further investigate the term structure We can see that the mean Contango reading is
“levels” by separating the readings between two about 0.93, while the mean Backwardation reading is
subsets (Contango/Backwardation) and then about 1.03.
applying percentile distribution analysis.

Spot/Vol. Dynamics
Rolling 7 days linear regression of log returns Spot/Atm30
rolling 7 days linear regression of log returns Spot/atm30
1
Beta
−1
−2
Taking the previous analysis one step further, we can can easily pass different parameters for analysis) both
now investigate how spot markets affect ATM volatility. on a 7-day rolling window above and as a scatter plot
We compare spot and 30-day ATM volatility (customers using hourly data, below.
Correlations log returns Spot/Atm30

correlations log returns Spot/atm30
2019 2020
0.05
ret_atm30
−0.05
2021 2022
0.05
ret_atm30
−0.05
−0.05 0 0.05 −0.05 0 0.05
ret_spot ret_spot
2022 stands out as a regime-changing year. We can see One of the big questions in crypto vol is whether this shift
a much more negative scatter plot slope and a “tighter” is cyclical (temporary) or structural (persistent).
range of readings.
In traditional markets, equity spot/vol regimes will be
On the rolling plot, we can also notice deeply negative negative (like we see in 2022 for BTC) but commodities
and persistent readings of the beta values. often display a positive relationship, as production & demand
shocks affect the commodity volatility space in a different way.
RR-Skew
While we described the ATM term structure as the “spine” of

the volatility surface, the wings would be the OTM (out-of-the-
money) options. Specifically, we analyze the Δ25 OTM options.
The relationship between the Δ25-Call implied volatility (minus)
the Δ25-Put implied volatility would be the RR-Skew (Risk-
Reversal Skew).
Skew30 Box plot

skew 30 box plot
2019 2020 2021 2022
60
40
20
Skew30
−20
−40
We can see two big trends when the RR-Skew is 2021 stands out as a clear outlier in terms of overall
grouped and analyzed by different year buckets. RR-Skew range.
First, although 2020 and 2021 had similar variance in Secondly, the total range of RR-Skew for 2019, 2020,
the term structure analysis, the RR-Skew was much and 2021 is loosely symmetrical. A clear outlier to
more muted in 2020 than in 2021. this trend is 2022, where the RR-Skew is decidedly
negative and nearly never positive.

Atm30/Skew30 scatter plot per year and quarter
ATM30/Skew30 scatter plot per year and quarter
2019 2020 2021 2022
60 quarter
1
2
40 3
4
20
Skew30
−20
−40
50 100 150 200 50 100 150 200 50 100 150 200 50 100 150 200
atm30 atm30 atm30 atm30
Layering the relationship between 30-day ATM implied 2022 saw the opposite, especially pronounced in Q2
volatility and the 30-day RR-Skew, we can see that 2021 and Q4.
and 2022 have two completely inverse relationships.
2021 is marked by upside momentum - especially true As implied volatility came into the market, the RR-
for Q1 2021. As implied volatility increased the RR-Skew Skew became more negative.
became more and more positive.
25∆ vs ATM Ratio Normalized Constant 30 Days
25Δ vs ATM Ratio Normalized constant 30 days
30
20
10
%
−10
call put
25∆ vs ATM Ratio Normalized Constant 30 Days

25Δ vs ATM Ratio Normalized constant 30 days
30
20
10
%
−10
30
20
10
%
−10
call put
If we isolate the Δ25 Call/Put wings and ratio them Puts have never traded at a discount to ATM volatility,
by ATM volatility, we can see that call wings began to but in the bullish market of Jan 2021, we can see Δ25
persistently trade at a discount to ATM IV (by nearly Put IV traded on par with ATM IV. This type of volatility
10%) - a phenomenon similar to equity index volatility pricing is enticing for 1x2 and put ratio spreads, which
- starting in July 2021. allow you to buy 2 OTM options with nearly no IV
premium to the ATM option you’re selling.

VRP
Atm-Rv 30 Day (Atm Shifted)
atm-rv 30 day (ATM SHIFTED)
150
volatility
100
50
50
−50
rv30 atm30 var premium
Are BTC options fairly priced in general? We start by analyzing 30-day implied volatility versus
30-day realized volatility, and by shifting back IV 30-
In general, does implied volatility overestimate the days, since IV prices future realized volatility.
future realized volatility?
Further, we break out this analysis into two buckets:
Here we look at the VRP (variance risk-premium) in
order to assess this statement. 1) When the term structure is in Contango.
2) When the term structure is in Backwardation.
VRP 30 days Termstructure Box Plot (Atm Shifted)
VRP 30 days termstructure box plot (atm shifted)
contango
backwardation
−100 −50 0 50 100
When the term structure is in Contango, the VRP has a mean of about +15pts. Meaning that 30-day ATM
options are overpricing RV by +15 points (estimated $ value = 15 * Vega)
VRP 30 days Skew Box Plot (Atm Shifted)

VRP 30 days skew box plot (atm shifted)
positive
negative
−100 −50 0 50 100
Furthermore, we analyze the VRP environment by different RR-Skew regimes. The negative RR-Skew regime
has a lot more VRP noise than the positive RR-Skew regime, meaning that selling options in a positive RR-Skew
environment is more certain.

VRP 30 days +Skew Termstructure Box Plot (Atm Shifted)
VRP 30 days + skew termstructure box plot (atm shifted)
contango
backwardation
−100 −50 0 50
VRP 30 days -Skew Termstructure Box Plot (Atm Shifted)

VRP 30 days - skew termstructre box plot (atm shifted)
contango
backwardation
−100 −50 0 50 100
Isolating the positive RR-Skew environment and range, meaning options purchases generally are valued
layering the term structure regime will help us find when spot prices explode higher with FOMO, but it’s a
the truly juicy VRP. We can see that there’s a slight noisy trade.
tradeoff between Contango/Backwardation.
The Contango regime provides a positive VRP but the
Backwardation in a positive RR-Skew environment has variance of the trade truly depends on the RR-Skew
a negative VRP mean but a large potential distribution regime.
Futures Basis
Deribit BTC options are actually BTC future options.
This distinction brings an additional complexity when trading BTC options, in the form of futures basis (aka the
implicit interest rate).
Futures prices can move even further than spot prices as the basis expands/contracts, essentially extending
delta movements a bit further.
Constant Basis Annualized Box Plot

CONSTANT BASIS annualized box plot
2019 2020 2021 2022
50
−50
−100
−150
b30 b60 b90 b120 b30 b60 b90 b120 b30 b60 b90 b120 b30 b60 b90 b120
We can see a few interesting trends when analyzing Secondly, we can see that 2022 was an outlier in
the futures basis percentile distributions by bucket- terms of the basis “compression”.
ed years.
Another peculiar insight is that RR-Skew was most
First, the basis is nearly always positive, except for volatile in 2021 and quiet during 2020. This relation-
outlier readings. ship doesn’t hold for futures basis ranges, where 2020
was the most volatile and 2022 clearly compressed.

Volume and Open Interest
Deribit BTC monthly volumes
deribit BTC monthly volumes
800k
700k
600k
500k
contracts
400k
300k
200k
100k
0
Jan 2019 Jul 2019 Jan 2020 Jul 2020 Jan 2021 Jul 2021 Jan 2022 Jul 2022
On-Screen Block
Deribit ETH monthly volumes

deribit ETH monthly volumes
8M
6M
contracts
4M
2M
0
Jul 2019 Jan 2020 Jul 2020 Jan 2021 Jul 2021 Jan 2022 Jul 2022
On-Screen Block
Looking at contract volumes for both BTC and ETH, Another interesting insight is the share of block-
we can see that monthly volumes have been growing trading growth for both BTC and ETH, although this
significantly for ETH while BTC hasn’t grown much is especially pronounced in ETH volumes.
beyond the highs seen in Jan 2021.
Open interest BTC vs ETH notional $ with PutCall ratio
open interest btc vs eth notional $ with put/call ratio
10B
OI notional $
5B
0
Jan 2020 Jul 2020 Jan 2021 Jul 2021 Jan 2022 Jul 2022 Jan 2023
1
PC ratio
0.8
0.6
0.4
0.2
Jan 2020 Jul 2020 Jan 2021 Jul 2021 Jan 2022 Jul 2022 Jan 2023
BTC ETH
Analyzing the open interest profile, we can also see When looking at the PC Ratio, we can see that
the enthusiasm for ETH options. The “flippening” there’s significantly more interest in ETH calls as
occurred in July 2022 as ETH OI has outstripped well. This divergence is fascinating, given that the PC
BTC OI, essentially making ETH an equally important ratio of BTC and ETH were rather similar in the past;
crypto volatility market. only around May 2022 did this divergence begin to
occur and persist.

Beyond Analysis: Strategy Backtesting #1
Harvesting BTC’s
Risk-Reversal Premium
By Samneet Chepal
For those in the crypto space, 2022 has not been short In this research piece, we will explore one of these com-
of surprises. Whether it be an entire L1 blockchain im- monly traded option structures: the risk-reversal (RR).
ploding overnight or a prominent exchange commingling To aid our analysis, we will reference an academic paper
user funds, the past few months have been interesting. written by veteran option traders which outlines the
Despite the carnage in this industry, the crypto options profitability of trading RRs in traditional markets. For the
market is still chugging along, with traders structuring po- purpose of this research, we will focus exclusively on BTC
sitions to speculate and hedge their risk. Unlike trading options trading on Deribit and look for opportunities to
spot or futures, option structures can give traders the make money across both bear and bull markets.
ability to collect income amidst choppy or bearish market This research piece will be organized as follows:
conditions.
• Overview of the Risk-Reversal Structure
• Analyze Backtested Returns
• Summary and Next Steps
SHORT PUT lon
call s
put strike
Overview of the Risk-Reversal Structure

Before we delve into the details of the academic paper, it’s important to understand the mechanics of trading RRs. As
shown below, we can define buying a RR as selling an OTM put and buying an OTM call (the delta of the call and put
should be nearly the same).
long PUT s
SHORT PUT SHORT PUT long call long call
long risk-reversal
+ =
put strike
call strike call strike
ca
strike put strike
Put strike call strike
Conversely, selling a RR will involve selling an OTM call and buying an OTM put.
PUT long PUT short call shortSHORT

callrisk-reversal
long risk-reversal long risk-reversal
rike put strike
+ call strike
= call strike

Put strike call strike Put strike call strike
SHORT risk-reversal SHORT risk-reversal

Depending on the strikes chosen, buying a RR is Short
At this pointPut Vegaturn our attention to the
we can long call Ve
similar to being long the underlying, whereas selling option’s vega which will be critical in understanding
the RR is similar to being short the underlying. the dynamics of how RRs are traded. As a refresher,
VEGA EXPOSURE
VEGA EXPOSURE
However, as volatility traders, our primary goal is an option’s vega exposure represents how much the
to trade volatility rather than the direction of BTC option price will change for a 1% change in implied
put strike
(also known as delta). For this reason, it’s common volatility. An option’s vega is at its peak when the
CALL strike
to see option traders hedge their delta exposure option strike is near the current price of the asset.
with futures contracts to eliminate any impact of Below, we can visually analyze the net vega exposure
price movements - which keeps their P&L clean and of being long and short a RR.
focused on trading volatility. Specifically, the hedged
RR is a bet on the relative value of puts and calls.
Short Put VegaShort Put Vega long call Vegalong call Vega long risk reversal
+ =
Short call Vega net vega long put Veg
VEGA EXPOSURE
VEGA EXPOSURE
VEGA EXPOSURE
VEGA EXPOSURE
VEGA EXPOSURE
VEGA EXPOSURE
VEGA EXPOSURE
put strike put strike put strike
CALL strike call strike
CALL strike call strike
put strike
long risk reversal

long risk reversal
net
Short call VegaShort call Vega vega net vega
long put Vega long put Vega short risk reversal
+ =
net vega
VEGA EXPOSURE
VEGA EXPOSURE
VEGA EXPOSURE
VEGA EXPOSURE
VEGA EXPOSURE
VEGA EXPOSURE
VEGA EXPOSURE
put strike put strike

call strike call strike call strike
put strike put strike put strike
short risk reversal

short risk reversal
net vega net vega
VEGA EXPOSURE
VEGA EXPOSURE

put strike put strike
Recall that with the long RR position we are long the call Conversely, we can also look at this perspective from a
and short the put. Therefore, our vega exposure to the short RR position which benefits from negative spot-vol
call will reach a maximum positive value at the call strike. correlation:
Conversely, given we are short the put, the vega will • A decrease in BTC price followed by an increase
reach its lowest negative value at the put strike. In other in volatility: in a short RR trade, as the price of
words, we can say that being long the RR will benefit BTC falls, the net vega of the long put position will
from positive spot-vol correlation. Let’s consider two increase. Therefore, it is favorable for volatility to
scenarios which encompass positive spot-vol correlation: increase as prices fall, given the trade’s positive vega
• An increase in BTC price is followed by an increase exposure. (ie: +vega × +vol = +P&L).
in volatility: as the price of BTC increases, it will • An increase in BTC price followed by a decrease in
approach the call strike, which will increase the vega volatility: this would benefit the short call position
exposure of the trade. At the same time, if implied because as price increases the net vega exposure to
volatility is increasing while vega is high, this will the call is reduced. At the same time, if volatility is
result in more profit compared to if vega was lower decreasing then a net negative vega exposure will
(ie: +vega × +vol = +P&L). benefit from this (ie: -vega × -vol = +P&L).
• A decrease in BTC price followed by a decrease in
volatility: in a long RR position we’re short the put *(Note: while this framework does not perfectly describe
option; therefore, for any downside price action the the spot/vol relationship in all scenarios, it can help us
net vega exposure is negative. Consequently, a de- identify regimes where trading certain RR structures can
crease in volatility coupled with negative vega would be more favorable)
result in profit (ie: -vega × -vol = +P&L).
Change in Spot/Vol Long Risk-Reversal Short Risk-Reversal

(Long Call/Short Put) (Long Put/Short Call)
Spot +/Vol -
Negative Spot- -PnL +PnL
vol correlation
Spot -/Vol +
Negative Spot- -PnL +PnL
vol correlation
Spot +/Vol +
Positive Spot- +PnL -PnL
vol correlation
Spot -/Vol -
Positive Spot- +PnL -PnL
vol correlation

At this point, we can now discuss some of for downside protection causes put options to be
the details of the academic paper used as the overpriced relative to calls. To prove this hypothesis,
foundation for this research note. The aim of several backtests were performed to showcase
“The Risk-Reversal Premium” is to demonstrate that buying the RR on index ETF options (buy the
how investors can monetize the supply/demand call and sell the put) outperformed the index with
imbalances in traditional option markets by trading strong risk-adjusted returns. For this reason, the
RRs. Furthermore, the authors share backtested return generated by this strategy is referred to as
returns of each strategy and showcase how including the risk-reversal premium. Using the original paper
RRs in a broader portfolio can improve the overall as a foundation, we can extend this research to
risk-adjusted return. The authors suggest that the BTC crypto options and investigate whether these
risk-averse nature of investors and constant demand findings remain consistent.
Analyze Backtested Returns

Below are some key factors to keep in mind when
interpreting the backtest results:
• The Deribit BTC options data used for this backtest • Throughout the trade, the risk-reversal will be delta-
ranges from 2019-04 to 2022-12. Given that Deribit hedged daily based on the end of day net exposure.
BTC options are coin-denominated, all P&L will be This is the naive baseline approach, as in practice,
denominated in BTC. traders will use dynamic thresholds to execute their
• The mark option price will be used to calculate P&L. hedges accordingly.
This is not entirely realistic in practice; however,
the objective here is to showcase whether any
From here, we can begin to analyze the first two
phenomena or inefficiencies first exist. With this
approaches:
knowledge, the trader can then decide on how to
position accordingly. • Long Risk Reversal: systematically sell puts and buy
• The target maturity for all options will be 30 days - calls
this maturity was chosen discretionarily but generally • Short Risk Reversal: systematically buy puts and sell
has one of the better liquidity profiles on Deribit. calls
• The model will enter into the trade at the beginning
of the week and select strikes closest to the user’s
specified delta target. The model will then close this
position after one week and roll into new strikes
closer to the target delta threshold.
Short puts <> long calls | 30 day maturity | 2019-04-01 to 2022-12-04
Short calls <> long puts | 30 day maturity | 2019-04-01 to 2022-12-01
As can be seen, it’s evident that selling the RR is the to capitalize from this black-swan event; however, to
better of the two approaches. The main driver of P&L keep things more fair, we can remove the impact of
stems from the model’s positioning around Black this single trade and look at the rest of the P&L. It’s
Thursday during March 2020, when the price of BTC interesting to note that despite removing this outlier
nearly halved overnight. Even though the RR trade trade, being short the RR still produces relatively
was hedged, the put option exploded in value and strong returns relative to buying the RR.
outweighed the P&L from the delta-hedge. Being
short the RR would have positioned a trader very well

ex-black thursday:
short puts <> long calls | 30 day maturity | 2019-04-01 to 2022-12-01
Short calls <> long puts | 30 day maturity | 2019-04-01 to 2022-12-01
As we discussed earlier above, the performance of the the 60 day rolling correlation between the change in
RR will depend heavily on the spot-vol regime of the daily BTC closing price and change in 1 month BTC
market. For illustration purposes, below we can see ATM implied volatility over the past 3 years.
rolling 60 day correlation: BTC close % change <> 1 month

BTC ATM IV % change
Overall, the correlation between changes in spot • Each position will be held-on for a maximum of 30
and changes in vol is negative 65% of the time, which days, after which any open position will be closed to
makes sense when we look at the performance of prevent stale positions
the short RR trading strategy. Recall that a short RR
position performs well during periods of negative This process above helps normalize the skew time-
spot-vol correlation. Therefore, the backtest results series into a format which can be used to identify
from above align well with these empirical findings. when skew is out of line, relative to historical data
points. When trading a 25 delta RR if the skew
At this point, we can build upon the previous two is excessively positive this implies 25 delta calls
backtests and include a timing indicator, which can are bid higher than 25 delta puts; therefore, the
help filter for higher quality trades. This approach model would sell the risk-reversal (sell rich calls
would involve using a market indicator to determine and buy cheap puts). Conversely, when the skew
whether to buy or sell the risk-reversal. In this case, is negative, then the 25 delta puts have a higher
our indicator will be the z-score of the 1 month BTC implied volatility than their respective 25 delta calls.
skew time-series. Therefore, the model would buy the risk-reversal
(sell rich puts and buy cheap calls). Below we can see
• The skew data is defined as: 1 month call IV - 1
the backtested returns of running this strategy with
month put IV (the delta of the skew will depend on
a z-score threshold of 1.00.
the user-specified delta target - ie: for 15 delta we
will use 15 delta skew)
• The z-score is calculated as: (current observation
- rolling 30 day average) / (rolling 30 day standard
deviation)

Risk Reversal backtest using skew entry indicator
30 days maturity | z-score = 1.00 | 2019-04-01 to 2022-12-01
Furthermore, on a comparative basis, we can analyze the performance of the metrics across all strategies.
Among the different combinations of strategies, simply selling the risk-reversal offers the best risk-adjusted
returns. It should be noted that the statistics with simply selling the risk-reversal appear stable across differ-
ent deltas, which may suggest more robustness compared to the skew indicator model.
Long Risk Reversal Short Risk Reversal
15 Delta 25 Delta 35 Delta 15 Delta 25 Delta 35 Delta
CAGR (%) -13.600 -12.500 -13.300 CAGR (%) 10.200 9.500 11.900
Max Drawdown (%) -49.300 -47.300 -45.500 Max Drawdown (%) -11.100 -12.500 -13.200
Sharpe -0.449 -0.443 -0.611 Sharpe 0.449 0.443 0.611
Skew Indicator Risk Reversal (zscore=1.00) Skew Indicator Risk Reversal (zscore=2.00)
15 Delta 25 Delta 35 Delta 15 Delta 25 Delta 35 Delta
CAGR (%) 5.300 7.100 0.600 CAGR (%) -4.400 -4.100 -1.700
Max Drawdown (%) -29.400 -25.000 -32.000 Max Drawdown (%) -41.700 -39.400 -35.900
Sharpe 0.291 0.353 0.136 Sharpe -0.027 -0.047 0.062
One final study that we can perform is analyzing The goal here is to determine the optimal weighting
the performance of combining the RR strategy with for the trading strategy in the overall portfolio. Below
a long BTC position. Given that selling the RR had is a diagram outlining the optimal allocation to the
consistent returns across all delta thresholds, we will RR trading strategy. In the cases below, an allocation
use this strategy for our portfolio analysis. anywhere between 75% to 80% dedicated to the trading
strategy and the remainder to spot BTC offered the best
risk-adjusted returns using the Sharpe ratio
optimal portfolio allocation: BTC + Short call and long put

optimal risk-adjusted portfolio cumulative returns: BTC + short call & Long put
Conclusion and Next Steps

Overall, it’s interesting that unlike traditional mar- the crypto options market matures with new institu-
kets, systematically selling risk-reversals across time tional participants, it’s reasonable to expect a different
offer strong risk-adjusted returns. This can likely be regime for trading risk-reversals given the change in
explained by BTC’s strong call skew during the early supply/demand dynamics. As a result, these strategies
2021 bull-market, in addition to several market shocks will likely become less profitable and new approaches
which greatly rewarded buyers of puts. However, as will be required to squeeze out marginal alpha.
Acknowledgments
Greg Magadini, Euan Sinclair, and Ben Moussa for their insightful comments and suggestions on this research piece.
Beyond Analysis: Strategy Backtesting #2
Systematic Volatility Trading

With BTC Options
By Samneet Chepal
One of the benefits of systematic trading is that it pro- traders can take note and use these insights to further
vides investors with a clear plan on how to execute their refine their trading approach.
trading strategy. Having a good understanding of a strat-
egy’s historical performance and expected future returns This piece will be structured as follows:
across various market conditions can provide greater
ease to a trader, especially during abnormal volatility. • Systematically Selling Straddles
In this research piece, we will analyze the backtested • Systematically Selling Butterfly Spreads
performance of several systematic volatility trading strat- • Event-Driven Long-Volatility Bets
egies tailored to the BTC options market. While all of • Overview and Conclusion
these approaches will be purely systematic, discretionary
Systematic Straddle Trading

A straddle is a popular options structure which allows other words, the volatility implied by the options market
traders to make bets on the future volatility of the un- is typically higher than the actual realized volatility over
derlying asset. A trader buying the straddle will purchase the long-run. As shown below, since April 2019, the im-
a call and put option closest to the current price of the plied volatility for 30-day BTC options has been greater
underlying, with the expectation of realized volatility than 30-day realized volatility for nearly 70% of the time.
being higher, relative to the current implied volatility. Given this behavior, it would make more sense for us to
It’s a well-known observation that the market anticipates sell straddles to profit from the market’s tendency to
future volatility to be higher than actual volatility. In overprice implied volatility.
BTC IV/RV 30 day maturity vol spread: 2019-04-01 to 2022-12-07

PUT short PUT short call SHORT
short callSTRADDLE
e put strike
+ call strike
= call strike
The results below showcase the backtested performance the delta of the position to fluctuate, and thereby reduc-
of selling straddles across different maturities. Given that es transaction costs associated with constant hedging
this strategy aims to be delta-neutral, the model will (in this case, we’ll set our threshold to hedging the delta
hedge only after the index price moves past a certain after every 2.5% change in BTC price).
threshold. This is common industry practice as it allows
SHORT STRADDLE SHORT STRADDLE
systematic BTC straddle selling (2.50% hedge threshold)

2019-05-01 to 2022-12-04
Despite several crashes over the past years, this strategy performs well over the backtested period with strong
risk-adjusted performance. Empirically, this makes sense, as selling options allows traders to collect premium and
compound gains over time. As can be seen below, the performance of the weekly straddle strategy offers good
risk-adjusted returns compared to simply holding BTC.
7 Day Straddle 15 Day Straddle 30 Day Straddle Spot BTC
CAGR (%) 53.400 22.000 14.600 39.90

Max Drawdown (%) -24.300 -34.100 -53.800 -76.70
Sharpe 1.382 0.698 0.495 0.69
It should be noted that naively selling volatility like this The maturity selected has a significant impact on how
has led to disastrous outcomes when the risk is not much we should allocate to the portfolio for optimal
properly managed. Selling straddles can in theory lead risk-adjusted returns. In this case, the weekly straddle
to unlimited losses, given that volatility can skyrocket to selling strategy comprises nearly 90% of the portfolio
levels which were previously not observed. The primary paired with 10% spot BTC. Conversely, for longer-dated
reason for the blow-up of these vol-selling strategies monthly straddles, the strategy only needs to have a
revolves around their excessive use of leverage. For ex- 30% allocation before the Sharpe ratio begins to decline.
ample, selling straddles with the goal of collecting “stable Approaching volatility harvesting with this framework is
income” can work in many cases; however, extreme mar- likely more sustainable and can help investors tolerate
ket volatility (which occurs more often than implied by future drawdowns with greater ease.
the options market) can completely erode all profits. As a
result, it would be advisable to allocate only a fraction of
the portfolio to volatility selling. Below, we can observe
the optimal weight for the straddle selling strategy in a
broader portfolio of spot BTC.
optimal portfolio allocation BTC + Short straddle

Systematic Butterfly Trading
Although selling straddles can earn decent returns over price of BTC and buying a call slightly below and slightly
a long enough time period, the amount of risk inherent above the current price (ideally they are equidistant to
with this strategy may not be tolerable to specific inves- make the payoff equal). For the purpose of our backtest,
tors. In this case, butterfly structures can offer a similar the OTM and ITM strikes will be 10% away from the cur-
payout structure to the straddle, except that the infinite rent price of BTC. Note: this butterfly structure can also
long 1x
downside riskotm call at the expense of lower short
is capped profits.2X ATM
beCALLS
replicated using put options. long 1x itm call
This structure involves selling two calls near the current
OTM strike ITM strike

long
long1x
long otm
1x1x otm
otm call
call
call short 2X
short
short ATM
2X2X CALLS
ATM
ATM CALLS
CALLS long 1x
long
long itm
1x1x call
itmitm call
call
ATM strike
OTM
OTMstrike
OTM strike
strike
+ ATM strike
ATMATM strike
strike
+ ITM
ITMstrike
ITM strike
strike
long call Butterfly
= long
longcall
call
long Butterfly
Butterfly
call
OTMstrike
Butterfly
ATM strike
ATM strike
ATMATM
strike
ITM strike
strike
OTMstrike
OTMstrike
OTMstrike ITM strike
ITMITM
strike
strike
Following the same delta-hedging logic as the straddles the straddle strategy makes intuitive sense because, in
above, we can see that the P&L associated with this this case, we have to buy two options which significantly
approach offers decent returns over the backtest period, erodes P&L due to theta decay. Although downside risk
however, the results across different maturities have is limited with this approach, investors are likely better
significant variance. After adjusting for a key outlier off selling straddles but adjusting their position size
trade, we can see that even the best performing strategy accordingly to a point where the portfolio risk is in line
fails to beat the performance of the worst performing with their desired thresholds.
straddle strategy. The lower overall returns relative to
BTC Butterfly +/- 10% strikes | 2.50% hedge threshold

2019-05-01 to 2022-12-14

Event Driven Volatility Trading
As discussed above, it is usually more profitable to be volatility in the underlying. We would want to close this
short options given that implied volatility tends to be position a day before the actual event to avoid the post-
higher than realized volatility. However, there are unique event volatility sell-off. In both crypto and traditional
situations where being long volatility before key events option markets, it’s common to see a “vol crush” after
can lead to profitable trading opportunities. In this case, key economic events have passed, as the market now has
we’ll explore the P&L of systematically buying volatility a less uncertainty to digest.
week before key economic data releases such as CPI and
Fed decisions. For example, recall earlier in the summer of 2022 when
ETH was moving to Proof-of-Stake. In the days leading
To execute this strategy, we will buy an ATM call option up to the event, the options market was pricing higher
one week before a key economic event. The option will volatility, given the uncertainty associated with whether
be hedged based on a 2.50% threshold approach, as the Merge was actually going to happen. Shortly after
discussed in the straddle strategy. The logic here is that the event, ETH IV plummeted once the market was
in the days leading up to the event, the market may have convinced the Merge went through successfully.
uncertainty on the outcome - leading to greater realized
ETH 30 day constant maturity ATM IV: 2022-09-01 to 2022-09-30
Ever since the FOMC started raising interest rates back in November 2021, crypto markets have been highly reactive
to key economic releases (ironically, for an asset class that prides itself in decentralization, trading crypto over the past
year has been highly dependant on the outcome of a single institution – the Fed!). Below we can see the performance
of this strategy individually on CPI and FOMC events.
buying volatility 7 days before FOMC release date 2021-11-01 to 2022-12-13
buying volatility 7 days before CPI release date 2021-11-01 to 2022-12-13

While the performance of this trading strategy is strategy. However, observing the overall positive returns
nowhere close to the straddle selling approach, it should across both approaches may suggest there is some
be noted that there have been less than 15 trades for benefit to being positioned in a long-volatility structure
each individual backtest. Therefore, the small sample size before key macro events. If anything, these long-volatility
makes it too early to conclude whether this strategy will bets may enhance or serve as a hedge against existing
continue to work in practice – we need more data to short-volatility strategies in a broader portfolio.
reach statistical confidence on the effectiveness of this
Conclusion
Despite running different backtests and scenario for the risks and returns of a given strategy can provide
analysis, ultimately, a strategy’s success comes down further validation of a strategy’s effectiveness and give
to whether an investor decides to proceed forward. the trader more confidence to increase their position
Although these strategies above are oriented towards size accordingly. Overall, as the derivatives market in this
systematic investors, discretionary traders can also space continues to grow with new underlyings, greater
find benefit in using a quantitative approach towards liquidity, and new products (ie: exotic options, variance
framing their investment framework. For example, while swaps etc…), we can expect further opportunities to
a discretionary trader may have strong intuition on the quantify new systematic strategies which investors can
nature of a particular market, having a baseline figure add to their crypto portfolios.
Disclaimer
This report provides data and analysis to the user All information from Amberdata is obtained from sources
to facilitate the user’s own investment decisions. It believed to be accurate and reliable. However, errors or
does not attempt to provide investment advice or omissions are possible due to human and/or mechanical
recommendations. Any interpretation of the data error. All information is provided “as is” without warranty
presented that leads to an investment is at the user’s own of any kind.
risk and Amberdata can not be held responsible for any
losses that occur from such investments. Amberdata and its information partners make no
representations as to accuracy, completeness, or
Important notice! Crypto currencies and options markets timeliness of the information on this site.
have large potential rewards, but also large potential risks.
You must be aware of the risks and be willing to accept Amberdata does not represent or endorse the accuracy
them in order to invest in crypto currencies and options or reliability of any of the information, content or
markets. advertisements (collectively, the “Materials”) contained on,
distributed through, or linked, downloaded or accessed
Don’t trade with money you can’t afford to lose. This from any of the services contained on this website (the
is neither a solicitation nor an offer to buy/sell crypto “Service”), nor the quality of any products, information or
currencies or options. other materials displayed, purchased, or obtained by you
as a result of an advertisement or any other information or
offer in or in connection with the Service (the “Products”).
Analysis and pre-built Python notebooks
available to customers
Amberdata.io
[email protected]
amberdata.io/ad-derivatives

BTC Options: Dissecting Volatility Trends

Uploaded by

Copyright:

Available Formats

BTC Options: Dissecting Volatility Trends

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

BTC Options: Dissecting Volatility Trends

Uploaded by

Copyright:

Available Formats

BTC OPTIONS

Dissecting Volatility Trends

Greg Fabio Tony Euan Samneet

Samneet Chepal, CFA

ATM Term Structure . . ................................................................................................06

Spot/Vol. Dynamics.. ..................................................................................................09

Risk-Reversal Skew. . .......................................................................................................10

Futures Basis. . .....................................................................................................................16

Volume and Open Interest............................................................................................17

Harvesting Risk-Reversal Premium. . ................................................................19

Systematic Volatility Trading. . .....................................................................30

Thank you for your continued trust and confidence in Amberdata.

atm7 atm30 atm60 atm90 atm180

The term structure is one of the most significant mea-

Backwardation occurs in reaction to volatility

PAGE 6 | BTC OPTIONS: DISSECTING VOLATILITY TRENDS

Since Backwardation is the abnormal state of affairs,

Richeness of term structure box plot

PAGE 8 | BTC OPTIONS: DISSECTING VOLATILITY TRENDS

Correlations log returns Spot/Atm30

−0.05 0 0.05 −0.05 0 0.05

While we described the ATM term structure as the “spine” of

Skew30 Box plot

PAGE 10 | BTC OPTIONS: DISSECTING VOLATILITY TRENDS

atm30 atm30 atm30 atm30

25∆ vs ATM Ratio Normalized Constant 30 Days

PAGE 12 | BTC OPTIONS: DISSECTING VOLATILITY TRENDS

rv30 atm30 var premium

VRP 30 days Skew Box Plot (Atm Shifted)

−100 −50 0 50 100

PAGE 14 | BTC OPTIONS: DISSECTING VOLATILITY TRENDS

VRP 30 days -Skew Termstructure Box Plot (Atm Shifted)

−100 −50 0 50 100

Constant Basis Annualized Box Plot

PAGE 16 | BTC OPTIONS: DISSECTING VOLATILITY TRENDS

Deribit ETH monthly volumes

PAGE 18 | BTC OPTIONS: DISSECTING VOLATILITY TRENDS

Overview of the Risk-Reversal Structure

Put strike call strike

PUT long PUT short call shortSHORT

rike put strike

Put strike call strike

SHORT risk-reversal SHORT risk-reversal

Put strike call strike Put strike call strike

PAGE 20 | BTC OPTIONS: DISSECTING VOLATILITY TRENDS

long risk reversal

put strike put strike

short risk reversal

call strike call strike

Change in Spot/Vol Long Risk-Reversal Short Risk-Reversal

PAGE 22 | BTC OPTIONS: DISSECTING VOLATILITY TRENDS

Analyze Backtested Returns

Short calls <> long puts | 30 day maturity | 2019-04-01 to 2022-12-01

PAGE 24 | BTC OPTIONS: DISSECTING VOLATILITY TRENDS

Short calls <> long puts | 30 day maturity | 2019-04-01 to 2022-12-01

rolling 60 day correlation: BTC close % change <> 1 month

PAGE 26 | BTC OPTIONS: DISSECTING VOLATILITY TRENDS

Long Risk Reversal Short Risk Reversal

15 Delta 25 Delta 35 Delta 15 Delta 25 Delta 35 Delta