A review of May 6, 2010 flash crash and the far OTM put options

As part of my project of the study of Super Trader Karen, I worked on the analysis of surviving the flash crash using Karen's trade style in the last couple of week and would like to share the initial part of this analysis here. Any comments are welcome.

 On Thursday May 6 of 2010, the US stock market had a famous flash crash when Dow Jones Industrial Average crashed 1,000 points (9%) and S&P 500 Index crashed 100 points (9%). The VIX shot up 18 points (74%) on the day as market volatility exploded. However, the US stock market recovered most of the losses within a couple of days. The 5 minute chart is shown below as a reference to the flash crash market action.

Looking at the big picture, the US market was in a big and long bull market since the historical bear market bottom at March 2009. The flash crash occurred in the beginning of a market correction. The correction lasted for about 4 months ending in late August of 2010 and the fall from top to bottom was about 17%. Although the market crashed about 30 minutes only, the volatility remained extremely high for over one day, as the fear of continuing market crash stayed rampant in those two days.

From this experience of a 100 point crash of SPX, it should be easier for us to understand why Karen chooses a crash point of 100 or 200 SPX fall in her money management rules. In the following sub-sections, we’ll examine the available option chains first. It should help us to understand the possible trade adjustment options. Then, we’ll try to make up a hypothetical trade that is opened before the crash and required for trade adjustment.  It should help us to gain some insights on how our trade position reacts to market crashes and what kind of impacts the position will experience after the trade adjustments.

The Option Chains at the End of Flash Crash Day

As the market started to fall significantly on this day, the put contracts with the original 5% ITM probability would have its probability increased to over 30% rather quickly. A Karen style option trader would be likely to begin the normal trade adjustment strategy by rolling down or rolling out the part of the put options in the portfolio. Considering the trader would not know how deep the drop would be for the day when the probability reached 30% initially, it would be logical for the trader to roll down or roll out to the next month at first.

After reviewing the put option trading volumes and open interests of June, June Quarterly and July cycles, it looked like Karen’s team might have rolled down the June puts to a strike of $800. As shown in Figure 9 SPX June Option Chain on May 6, 2010, the strike had an ITM probability of 4.2% and trading volume of 53K on the day. The trading volume of put strikes of about 5% ITM probability for June Quarterly and July had so much less contracts that they did not appear to be traded by professional traders. For example, the June Quarterly puts at $800 with 4.79% ITM probability had 55 days to expiration which was close to the 56 day norm as Karen specified. But it had only 7 contracts traded for the day.

Note Karen stated she had rolled out to October time frame. But there were no October and November options available on the day of May 6, 2010 according to TOS. I assume Karen did not remember the exact months for her trade adjustments during the flash crash a couple of years ago before the interview.

As the market continued to fall deeper in the afternoon, the Karen style option trader would be likely to see more and more put option strikes in existing portfolio reaching the points of 30% ITM probability. At this time, it would be logic to assume that the trader would like to further roll out the rest of the put options in the portfolio. A review of the option chains of July (70 DTE), August (105 DTE), September (133 DTE), September Quarterly (147 DTE) could give us some clues as well. The September put options of around 5% ITM probability were heavily traded as shown in Figure 10: SPX September Option Chain on May 6, 2010. The September Quarterly puts were thinly traded with the open interests under 100. So it did not look like Karen used this option cycle on that day.

Therefore, let’s take a deeper look at the September option chain to understand what kinds of trade adjustments could be done in this type market crashes...

A study of market drop and volatility rise for the last 4 years

I've been curious about the market volatility changes in the last 4 years (2010 to 2014) since Super trader Karen started her naked option selling strategy around 2010. Today, I finally got a chance to chart the SPY and its implied volatility and performed a brief analysis. In particular, I'm interested to see if there is some kind of relationship between SPY fall percentage and its IV rise percentage.

So I picked a few times in the last 4 years that IV rose quickly as shown in the chart below and created a table to investigate the percentage changes of IV and SPY price. In 2010, the market had the steepest drop (about 17%) within one month time, causing SPY IV (VIX) to rise 220% (more than tripled) in the same period. We have not seen this type of change 3 years since that.

After 2010, we had seen SPY IV rising close to 100% in a couple of months occasionally except in the last Jan in which month the IV doubled within 1 month. But the absolute values of SPY IV ranged from 11 to 15 and 20 mostly since 2013. This is the challenging period for the Karen style option selling. She used a set of different rules in the period, including using Weekly call options of 14 days to expiration, etc.

Selling calls of small Delta in up-trending market

In analyzing my losing bear call spreads in the last couple of months, I updated my TOS option profit and loss charting script so that it can track daily P&L for 6 rounds of trades. Then I did a back-test to see if I could keep selling calls with increased size and fixed Delta around 0.32 to end up with profitable month in current low volatility environment (VIX around 13 to 17).

As shown in the picture above, I made the trades using a combination of technical analysis and high probability method. On the technical side, I thought SPY broke a major support level of $177.5 initially and faced a resistance level of $185 later. All of these were invalidated by market moves. On the high probability side, I used Delta around 0.3 for short call options which corresponded to 70% probability of success approximately. After one adjustment, I decided to close all positions on 2-28 for a loss.

In my back test today using the ThinkScript, I wanted to see if I could adjust losing trades by increasing position size while keep Delta around 0.3 for up trending  market since I believe market does not "crash-up". So I started the bear call spread at 1 contract on 2-5. In the 1st adjustment on 2-12, I increased the contract size to 2. In the 2nd adjustment on 2-28, I increased the contract size to 3. When looking for Delta around 0.3, I found the strike prices got a lot more closer to the current SPY price as we approached expiration day. With 28 DTE, the short call price could be just $2.00 above current SPY price.

Thus, this test leads me to believe that we would still face challenges if we continue selling of small Delta (or small ITM probability) call options of the same month as the adjustment strategy in our efforts to maintain profitability for the option cycle in high probability trades even if the contract sizes are increased to bring in more potential profits to offset trade loses of adjusted trades. The starting point of the trade is also very important. If the trade did not start at the bottom of the Bollinger Band, the adjustment should work much better.

Due to the lack of any option data on 2-28 and in that weekend in TOS software (both ThinkBack and ThinkOnDemand), I could not sell option on that day while SPY price was higher in this back test.

A summary of Karen, the super trader's option selling strategy in 2013

More than one year ago, I was intrigued by the Interviews on Karen, the supertrader and posted an outline of her strategy discussed in the video. Since then, there were many reader comments. So I studied the portfolio margin that she used to achieve high return on margin, and wrote a few other posts related to the strategy. Besides, I also started working on back-testing Karen's strategy and set up a Karen Study Group.

On Feb 11, 2014, Karen was interviewed by Tasty Trade again to provide an update of her trading in 2013 and some of her general trading guidelines. Here, I'd like to share my understanding from the interview on this naked option selling strategy. Please note Karen will adapt her strategy according to market conditions such as the volatility VIX. She highly recommended traders to come up with strategies that work for themselves.

Karen, the Super Trader 2013

2013 Fund Performances & Goal

• Two funds totaling 196 Million USD, yearly return after expense
• 27%, 24% (over 50M USD profits)
• No losing month on 2013 (13 straight months, incl. Jan 2014)
• Close out every month. No rolling out to the next month, closed current month positions.
• Goal
• At least to beat SPX gains yearly (but there were no specifics discussed on what rules were used to ensure beating SPX)

Strategy of Selling Naked Options

• A strategy to profit on Theta (time decay) and volatility
• Rules may be changed depending on market trend and volatility
• Sell calls or puts independently
• Never sell options as spreads (i.e. strangles)
• Sell more puts depending on market trends
• In 2013, maxed out on the put side to sell twice as many puts as calls
• Sell calls with greater risks as market does not crash up
• In 2013, stayed light on call side to sell calls with shorter period of time
• The variable ratio of put to call contract sizes allows the strategy to adapt to market trend
• In 2013, the 2 to 1 put to call ratio kept the portfolio Delta from negative
• Most premium selling option portfolio have negative Delta (but not Karen's)

Trading Vehicles

• Traded SPX options mainly, used options of SPX e-mini futures /ES sometimes
• Main reason not trading SPY is commission costs, plus tax advantage 1256.
• No problems with SPX option liquidity
• Can potentially go to RUT if need other positions to put on
• Disciplined to avoid trading individual stocks even though they have high volatility

Option Strike Price Selection

Sell far out-of-the-money, high probability options depending on market volatility

• In 2011 & 2012 when volatility was higher (VIX >= 15), sold puts with 2 SD (5% ITM Probability)
• In 2013 when volatility was low (VIX around 11 to 13), sold puts with 1 SD (15% ITM Probability) to 1.5 SD (7% ITM Probability)
• Sell calls with 10% ITM probability

Option Cycle Selection

• 14 days to expiration for calls in general for 2013
• 56 days to expiration for calls if price can be projected and good premium can be obtained from far above the projected price
• Use trend analysis including 2 SD Bollinger bands, resistance, and FIB retracements
• 56 days to expiration for puts

Trade Entry (Always legging in trades)

• Sell calls when market rises
• Sell puts when market falls
• In 2013, sold more puts when VIX jumped
• Trades are entered in multiple days around predetermined option expiration days

Trade Exits

• Actively take profits and risks off the table
• For options (puts) starting around 56 DTE
• Close positions if profits reach 50% within 1^st 16 days
• Leave positions on after the 40 DTE neighborhood if they are not closed yet so that they can expire worthless
• For options (calls) starting around 14 days
• Close positions if profits reach 20% to 30% or even better 50% in a few days
• If price continues to rise with 1 week left for call options, then take the calls away and move them to next week while keeping puts on
• After some point of time, may stop pulling back profits since a built-in profit is realized already.
• Leave remaining positions of current month alone and start working on next month.

Trade Adjustments

• When ITM probability is less than 30%, no adjustment is required.
• When price drops and ITM prob becomes 30% (an arbitrary number) or greater for puts with plenty of days left (> 30 day)
• Adjust the position by taking it off and sell other far OTM puts about 20% ITM prob cautiously
• Some puts may be sold 10 points lower in ITM proba and others sold 15 points lower in ITM prob
• It may be pushed out to another week for puts (Not another month)
• Try to fix it in current week if possible
• It may lose one or a few dollars
• Try to make up the loss either through selling more put or call contracts
• Will keep existing calls as their value decrease and let them expire worthless
• Do not use existing calls as adjustments to the losses on the puts
• It avoids risks due to sudden market direction reversals
• Don’t have any option ITM!

Money and Risk Management

• The most important part of the trading game
• Use portfolio margin and further stress test potential loses
• Unlimited risk in theory
• Use TOS Analyzer Tab for risk evaluations
• The daily portfolio P&L numbers derived at 10% up and 12% down of the market price must not exceed net liquidation value in normal cases
• Portfolio P&L losses at the crash price must not exceed net liq
• In 2013, crash price = (current price – 100) x 88%
• The extra cushion is added due to low volatility (11, 12 to 14)
• In 2011, crash price = (current price – 200) x 88%
• The less cushion due to higher volatility
• Managing risk at this crash point assuming the portfolio had a major loss already
• If net liq drops due to adverse market conditions, the unrealized positions are losing money but they are considered safe if they have not reached adjustment points
• In worst cases, may buy some low valued puts or calls to offset the push from market move
• But this is not very often
• Don’t care or fear as long as the positions are safe
• Do not over-manage positions

Psychology and Mindset

• Control the emotions of both fear and greed
• Trust statistics and high probability (Use 2 SD for 95% odds of success)
• Use structured method and manage positions well
• Require a high level of confidence and repeatable successes for a long time.

Team Members

• 7 members: 6 traders and 1 controller
• Encourage other traders to set up of trading team for additional ideas

FXI Position Adjusted

In about 1 week after I sold FXI Feb$36.5 puts, FXI seemed to start a down trend which lasted over 1 month to now. Based on the Chart of FXI, it appears to have another leg of rolling down again. So I decided to reduce my assigned FXI position by selling 25% shares to reduce the risk. For the remaining long FXI ETF positions, I sold Apri$37 calls to collect $0.61 credit per contract. If FXI does not fall too much in the next couple of months or so, my position will be OK. Otherwise, I will continue to roll down my short calls when 80% of profit is reached for this long term position.

---

Created with ProphetCharts®

Initiated MarQ$180/183 bear call spreads on SPY

Market has dropped significantly in the last couple of weeks with a large number of distribution days. So my near term outlook is bearish for the market. The SPY is around $175.20 right now. I felt it was not a good time to buy straight puts since SPY already dropped below major support level around $177.5. As VIX reached 19.60 which was close to the high of last 6 months, I decided to sell a few bear call spreads.

I played with the March Quarterly $180/182, $180/183 & $180/184 in the TOS analyzer and found the 1st two spreads had similar return on margin around 30% while the last spread had a ROM of 27%. So I sold MarchQ $180/183 bear call spreads with a credit of $0.87. The Delta of short strike $180 was around 0.3.

According to the analyzer, the ITM probability of the short strike one month later is 26% and at March 31 is about 33%. I'll treat this trade as a directional trade with some wiggle rooms. If 80% of profit is reached, I'll close the trade. If the trade lasts 30 days, I'll review it and try to close it. In case the market starts to rebound, I may close the trade when SPY reaches $181 area.

Estimating In-The-Money Probability using IV and Standard Deviation

I've studied stock (and VIX)'s implied volatility changes with time before. It was based on the book by Dan Passarelli: Trading Option Greeks. This is a great book about Option Greeks and volatility. My post on the Effect of Prices, Time, Volatility on Option Greeks was also based on this book. It's these types of studies that lead me to the understanding of estimation of ITM probability of stocks.


Recently, I also researched for methods to calculate the ITM probability of options since the supertrader Karen uses it in her naked option selling strategy. If we find a formula to estimate the ITM probability, we can implement it in ThinkScript in our backtests of the supertrader's strategy.

ITM Probability Estimation Using Standard Deviation

Due to lack of ITM probability data in ThinkScript, the ITM probability can be estimated using the standard deviation of a normal distribution as shown below, according to http://en.wikipedia.org/wiki/Standard_deviation.

ITM Probability	SD Multiples
0.27%	3
4.55%	2
5%	1.960
10%	1.645
20%	1.282
30%	1.050
31.73%	1

Stock Price Strikes/Points for Corresponding ITM Probability

In a recent TastyTrade video, the relationship between stock price changes and SD, IV, expiration days is presented in a formula:

Stock Price Change Percentage = SD Multiple x IV / Square Root of (365/DTE).

Putting all the pieces of the above information together, we should be able to estimate the stock prices for any ITM probability:

Stock Price of a certain ITM Probability = Current Price x ( 1 + Stock Price Change Percentage)

It means we can calculate the call/put selling strikes and adjustment points using the above formula for Karen's strategy. Since all of these parameters are available in TOS software, it's possible for us to implement in TS. I'll update it once I get a chance to code it. We should be able to present these prices in stock price charts to provide good visual display as well.

Update

Based on Vince's comment below, I have updated the above table which included probability on both sides. For the strategy of selling naked options, the single sided probability should be used. Therefore, the following table is the right one.

ITM Probability	SD Multiples
0.13%	3
2.28%	2
2.50%	1.960
5%	1.645   (Sold put strike)
10%	1.282   (Sold call strike)
15%	1.050
15.87%	1
30%	0.52     (Adjustment point)

Vince also kindly shared the following web site that gives us the SD multiple (Z) in a graphical representation: http://www.mathsisfun.com/data/standard-normal-distribution-table.html. In this dynamic graph, you can select "Up to Z" button at top left side first, and move curve to a probability value of interests, then read the SD multiple (Z).

Super Trader Karen Study Group

The successful story of super trader Karen has captured attention and imagination of many option traders. Some readers are looking for ways to communicate with each other about the strategy. They left great comments or contacted me via email in the last year or so since my original post. Encouraged by the enthusiasm, I finally created a Yahoo Group to help option traders who are interested in studying this strategy. I certainly expect to learn a lot of insights from fellow option traders as well, as I have seen some members stating great success in the last few years selling options or managing million dollar account.

This is a restricted user group focusing on the study of selling naked options, the Karen way. Main topics that I can think of right now include the following:

• Discussions on super-trader Karen's strategy
• Back-testing the option premium selling strategy
• Trading (paper or real) the strategy
• Other related topics

The strategy requires intermediate and advanced understanding of option trades, such as implied volatility, probability, option Greeks. So it's only good for people with the proper level of knowledge and experience. I'd like to make the posted messages available to the public even though only the members can post. But for some reason, the Yahoo group web site grey'ed out the option of Non Members can view messages section. If I figure out how to enable this option in the future, then the public could read the group messages.

In the past 3 weeks since the inception of the group, we have over 30 members joined which is beyond my expectation. Members seem to be from different countries, include USA, Canada, Italy, India, Singapore, etc. If you think you have the right skill and like to contribute, welcome to join this dynamic community.

Some thoughts on back testing Super trader Karen's portfolio entries

In the last couple of weeks, our study group started some back-tests on Super-trader Karen's strategy. I worked on the entry rules as a beginning. Although major rules are described in my original post, there are some details or flexibility that need further explorations regarding the opening of the monthly portfolio.

Since all positions are not opened on day 56 to expiration, we need to have a range of opening days first. Then we should have some rules on when (conditions) to open options within the opening day ranges. Next we will need to figure out how many more puts to sell if necessary.

Personally, I prefer to test the opening of positions around 56 DTE +/- 7 days, because some of my recent studies showed far OTM option decay with a good rate starting around 2 to 2.5 month to expiration.

So, what are possible conditions for entering new trades in these 2 weeks? I tested the following rule briefly today, using an up-trending period from late October to November last year.

o   Before half positions are entered around 56 days to expiration (56 DTE + 7 days)

§  Sell put if market drops for 1 day and sell call if market rises for 1 day

o   After half positions are entered around 56 days to expiration (56 DTE - 7 days)

§  Sell put if market drops for 1 day and sell call if market rises for 3 days

DTE	Date	Option	Strike	Price	SumPremium
58	10/23/2013	S Put	151	0.265
57	10/24/2013	S Call	184	0.4	0.665
52	10/29/2013	S Call	186	0.35	1.015
51	10/30/2013	S Put	154	0.245	1.26

The December expiration monthly portfolio P&L chart is shown below. In the rising market, the short call strike at $184 got approached but not violated near the end of the option expiration. The 50% profit target worked well for this month as shown by the green bubble. The yellow bubbles named C1Back, P1Back and C2Back were located approximately 1 month after their corresponding entries.

In this example, holding positions longer than 30 days caused maximum profits for the portfolio. It was a surprise that this cycle would end in profits as the original thoughts were trying to find a tough period for the back test in the up-trending days (I'll provide an update if my script is found to have a bug). Due to the multiple entry dates and the fact that SPY never ended above the lowest upper short strike, the P&L of this cycle was not challenged at all.

Portfolio entry tests are a beginning of the back tests. We need to test various entry rules on different types of market conditions. After that, the adjustment rules will need to be back-tested. This may be the most complicated part of the strategy. Later on, exit rules will need some tests as well. Finally, we will put the rules together for comprehensive back-tests and finalize all the rules of the strategy.

The back-tests use end of day data only, as explained in the prior post. We will have to perform and analyze some paper trading which will give us real time trading experiences for this strategy before going on live accounts in the future.

Framework for Backtesting Super Trader Karen's strategy

During the holiday season, our super trader study group made some progresses. We started using TOS software to back test Karen's strategy. It will take quite some time for us to complete it. Here're some of my initial thoughts that I can share.

The ThinkOrSwim software provides two features that can be used for back testing: ThinkOnDemand and ThinkBack. The TOD supposedly offer a market DVR that records all market data at intraday as stated by the vendor's announcements. In reality, we found it not suitable for any option strategies that use high probabilities. These type of strategies require usages of far out of money strikes which are missing in TOD. TOD offers most option Greeks up to 10 strikes. Delta for far OTM options can be retrieved in TOD, but other important Greaks and ITM probability numbers are not available. The implied volatility of far OTM options are also available, but not consistent with actual IV in trading window as pointed out in a previous post. Additionally, the important market crash data happened in 2008 are not available in TOD.

The ThinkBack (TB) feature supports end-of-day data for a lot more option strikes, including the Greeks and ITM probabilities. So we have to use the ThinkBack feature in our back tests in order to find the option's ITM probability, etc. NOTE: The Greeks presented in ThinkBack were reported to have some discrepancies in my post before.

Going through TB manually for any serious back tests are not practical, as it requires users to click on each date. The ThinkScript (TS) becomes handy to help the back tests when used together with the TB. The TS can be developed to show all entry, exit, adjustment points on a chart, along with other important trading portfolio information. However, there are also limitations by TS: the ITM probability is not accessible and option Greeks can not be cross- referenced in a portfolio that includes multiple options.

Back-testing under the above mentioned constraints is challenging. But I believe it will still provide some insights about the strategy. I have created a step-by-step guide for how to use ThinkBack and ThinkScript to perform back tests on the super trader's strategy in a document. Using an initial version of our script, it gives a couple of charts that illustrate a simple strangle trade and a short call trade.