Superfly!

Tues hi-lo closing orders, values are just a first guess, will be reassessed



....call side: vert came off for $8 credit, bw fly off for $5.45 debit...call it a $2.50 net credit. Now watching the put side...

The put side came off: vert for $0.60 credit and the bw fly for 10c debit...so, a 50c credit.

...so, a $3 credit off on a $2.15 debit open, is a net 85c credit.

These work best when you can go 2:1 verts:fly; and try to get > $1.25 credit for the fly; and try to wedge in the vert as tight to highest delta strike of the bw fly as possible without over-spending. I have traded these for 2-3 weeks and not lost one yet. Gains are rather modest - $100-$300/unit, but I'm always looking for that easy daily income, so that's ok. If you tighten up and double up the verts, the gains are much better.

One of these days I'll put one on and SPX will print a thin doji and I might get caught in the dead middle and lose some money, but I think the loss would be very small. We'll see when that day comes. There is a theoretical max for the fly, right on the guts, which you'd have to wait til near the close to grab, but holding til close is a dangerous path.

These are usually done well b4 NY lunch, which is breakfast time for me on the west coast. That's cool - if I am just in some other weekly trades, it means I can break off until the last hour, when the next day's hi-lo can open.

10/30 hi-lo puts, net debit $2.35

calls....$2.50 net debit

puts off already for $1000. Best case is you put these on lae morning and you catch a big move on one side and then cash the profit side out still b4 the end of 1dte.

hi-lo for Friday...I also bought a single 6890 call for $6.30.....total net debit is -$815...will kill the bw fly early, see if the others might run a bit.

for Monday

...net debit $260. Here are the closing orders...really just a set-up for real values to come. Usually I am happy with anything net positive on these...avg has been around $150/unit I guess

full loss. Woulda been worse had I included the put side, as there was no movement to the downside.

Will try again for Tues...$3 net debit

Tuesday hi-lo, closed the puts at the open, a little under $1k. Calls will be a total loss but that's only $100. These work well when there's a big move. There will be a big debit for the bw fly but an even bigger credit for the vert. I have netted near $1k on a few of these now, all at 1-lot still. I'll scale up when it seems wise...

hi-lo for Tues, betting put side only for a modest pullback <6800 off today's big run

gotta luv flys. Idk how I got into this 80pt wide fly for $0.60, but I did. If I were smart, I could figure out an automated way to find those mispriced spreads. Anyway, $0.60 could go > $35 easy if SPX will settle just short of 6650 b4 the close. $360 bet could pay almost $24,000....

,,,,took all but 1 off for $8.50, a 14x winner. Looking like SPX will try to stage a bounce today

hi-lo variation for a Monday pop....$5.30 debit...looking for a run to 6800 and a 2x yield or better

this variation, w/ symmetrical, not bw fly, costs more but pays better and I put it on when I have conviction re: direction. Monday has more often been an up day for the index and the up move got going Friday. Overnight, it's looking good

big loser. Whenever I have "strong conviction" and play one side only, pretty much a guarantee market will eat my lunch.

ATR coming up nicely in the last weeks, now about 80 @SPX.....I am looking for some good movement off the NVDA earnings. Either way. We'll see where we stand tomorrow. 100 pts either way could be in play for a pin. I will look at double flys as shown, similar, 100 pts wide.

I need to up my game on the flys. I still recall that 90-wide put fly I bot for 60c a couple weeks ago. I made some money on it, but it didn't even finish well, but the main thing is, that thing was awfully misprised. I need to figure out how to find those mispriced flys.

I just got a new phone and it came with 6 months of Gemini, been having fun with that. Need to get serious...

The Fool: How would it be possible to search through all SPX butterfly trade setups to find ones that are mispriced?

That's a complex, high-level algorithmic trading concept. Finding mispriced SPX butterfly spreads involves a systematic approach combining high-speed data access, a theoretical pricing model, and rapid execution.
The general approach is to implement an automated screening algorithm that constantly compares live market prices to a calculated theoretical value.

The Process for Screening Mispriced Butterflies


1. Data and Infrastructure

 

1. Real-time Options Data: You need a high-speed feed of live bid/ask quotes for all relevant SPX options (different strikes and expiration cycles) to construct the butterfly spreads.
2. Low-Latency System: For arbitrage or finding mispricings that disappear quickly, an extremely low-latency trading and analysis system is essential. This often involves colocation or direct data connections.

2. Theoretical Pricing Model

A butterfly spread consists of four options legs, which are related by a no-arbitrage condition derived from option pricing theory. A mispricing exists when the current market price of the butterfly (the net debit or credit) deviates significantly from its theoretical, arbitrage-free price.
The most robust way to find mispriced butterflies is through two main methods:

1. Model-Based Pricing: Use an option pricing model, like a variation of the Black-Scholes model adjusted for the implied volatility smile/skew, to calculate the fair value of each of the four option legs. The fair value of the butterfly is the net sum of these theoretical prices. The difference between the live market price and the theoretical price suggests a mispricing.
2. Arbitrage Parity Relations (Box Spread): You can construct a butterfly spread as a combination of other, simpler spreads. For a pure, risk-free arbitrage, you would check for violations of established option pricing relations, such as the box spread parity formula, which states that a synthetic risk-free zero-coupon bond must equal the price of a cash-market risk-free bond. Violations here are rare and often fleeting in highly liquid markets like SPX.

3. The Screening Algorithm

Your algorithm would systematically check every possible butterfly combination for all available expirations and strike ranges:

1. Define a Butterfly: A butterfly is defined by three strikes (K1,K2,K3) with equal distance between them (i.e., K2−K1=K3−K2) and the same expiration date (T).
2. Calculate Theoretical Price: Compute the theoretical value (Vtheoretical) for the butterfly using your chosen model.
3. Calculate Market Price: Calculate the actual cost/value of the butterfly in the market using the mid-point of the bid/ask spread for each leg:
   Vmarket=(PriceK1)−2⋅(PriceK2)+(PriceK3)
4. Identify Mispricing: A trade setup is flagged as mispriced if:
   ∣Vmarket−Vtheoretical∣>Threshold
   The Threshold must be greater than the transaction costs (commissions, exchange fees, and bid-ask slippage) to ensure the trade is profitable.

4. Filters and Refinements

To make the search practical and executable, you'd apply filters:

1. Liquidity: Filter out combinations with low volume or wide bid-ask spreads on any of the four option legs, as this makes executing the trade at the calculated mispriced value very difficult.
2. Transaction Costs: Ensure the potential profit from the mispricing is high enough to cover the cost of four separate transactions.
3. Volatility Filters: Many traders look for times when the implied volatility of the strike prices is relatively high, as this often means the short option (the body of the butterfly) is richly priced, leading to a lower net debit for a long butterfly spread.

The Fool: I can think of a simpler approach.

The Fool: First, it would not be necessary to calculate spread values - I would simply define what "mispriced" means. For instance, for a given date, I would look for all SPX butterfly spreads with, let's say, symmetrical wings 50-pts wide, and within a certain distance from the at-the-money options at the market open, and find those butterfly spreads that would cost less than $1.00 to put on. Does that sound possible?

That sounds absolutely possible and is a much more practical, rule-based approach for a retail trader than attempting complex theoretical arbitrage calculations.

Rule-Based Screening for "Cheap" Butterflies

This simplified screening method focuses on finding outliers based on a defined Risk/Reward threshold, using the net debit as the primary filter.
The core idea is to automate the search for spreads that offer the largest potential profit (defined by the strike width) for the smallest possible cost (defined by your debit threshold).

1. Defining the Search Parameters

You need a systematic way to check all combinations that meet your specific criteria:
ParameterYour DefinitionTechnical Term
AssetS&P 500 Index Options$SPX$
Spread TypeSymmetrical Butterfly Spread (Long)Buy $K_1$, Sell $2 \cdot K_2$, Buy $K_3$
ExpirationA specific date (e.g., 0DTE, 1DTE)$T$
Wing Width50 points wide (e.g., $K_2 - K_1 = 50$)$\Delta K$
LocationWithin a certain distance from ATMOut-of-the-Money (OTM)
"Mispriced" FilterNet Debit Cost $\le \$1.00$Max Loss

2. Execution: The Screening Script

To execute this, you would need an API connection to a broker or data feed that provides real-time $SPX$ option quotes (bid/ask). The script would perform the following steps for a chosen expiration $T$:

1. Identify ATM Strike ($K_{ATM}$): Determine the current At-The-Money strike based on the $SPX$ market price at the open.
2. Iterate Through Strikes: Loop through a range of strike prices that are within your desired distance from $K_{ATM}$.
3. Construct the Butterfly: For each strike $K_2$ in the loop, define the butterfly's wings: $K_1 = K_2 - 50$ and $K_3 = K_2 + 50$.
4. Calculate Market Debit: Use the current mid-price of the bid/ask for each of the four legs to calculate the net debit (Max Loss):
   
   $$Debit = \left( \text{Price}_{K1} \right) - 2 \cdot \left( \text{Price}_{K2} \right) + \left( \text{Price}_{K3} \right)$$
   
   (Note: Using the mid-price is a common proxy, but for real trading, you'd use the current bid/ask spread for potential fill price.)
5. Apply the Filter: If the calculated $Debit \le \$1.00$ (or $\$0.60$ as you saw in the web page context), the setup is flagged as a potential mispricing.

3. The Implied Mispricing

The reason this $Debit \le \$1.00$ screen works as a proxy for mispricing is because:

1. For a 50-point wide butterfly, the Maximum Potential Profit is $50.00 - Debit$.
2. If the $Debit$ is only $\$1.00$, you are risking $\$100$ per contract for a potential maximum profit of $(\$50.00 - \$1.00) \times 100 = \$4,900$. This 49:1 maximum profit-to-risk ratio is an extreme outlier and suggests one of the following:
   1. The market is genuinely mispricing one or more legs.
   2. One or more legs are extremely illiquid, and the mid-price calculation is misleading (which is why you must filter for liquidity).

   
   

This simplified screen is an excellent way to systematically find these high-reward outliers that the trader in the linked post mentioned finding but couldn't find a way to automate.
Would you like to discuss the liquidity constraints (bid/ask spread) that you would need to add to that script to make the results truly actionable?

Ok Gemini, thanks for that example. Right now I am registering as an API Developer with my broker, so I can fine tune the script that will allow me to find cheap (and possibly very lucrative) butterfly spreads.

Normally I can help with things like this, but I don't seem to have access to that content. You can try again or ask me for something else.

I'm looking forward to asking you for more help on how to use their API and fine tune my scripts.

I would be happy to help you with the Python or other code aspects once you have your credentials.

Before I get too far down that rabbit hole, though, do you think there are any simplified API programs already out there that I could use to search for "mispriced" butterfly spreads?


That's a great question, and yes, you're on the right track! There are several ways to simplify the process, ranging from pre-built, options-focused automation platforms to open-source Python libraries that handle the API complexity for you.
You'll likely find the most success with Python given the E-Trade API wrappers available.

Simplified API Programs for E-Trade

Since you are registering with the E-Trade Developer Platform, the easiest path is to leverage existing community-built Python wrappers that abstract the complex OAuth authentication and data fetching.

1. Python Libraries (Recommended)

Python is the most common language for algorithmic trading, and developers have created specific libraries to make the E-Trade API easier to use:

1. pyetrade: This is a widely used, community-maintained Python wrapper for the E-Trade API.1 It simplifies the process of making requests to endpoints like get_option_chains and get_quote, which are exactly what you need to search and price your butterfly spreads. You can find the source code on GitHub.
2. ETradeBot: This is a Python-based automated trading software that is strategy-agnostic.2 While it's designed for full automation, you could adapt its functions for fetching real-time market data to execute your specific screening script (i.e., the Debit 3$\le \$1.00$ filter).4 You can learn more about it on PyPI.
3. wetrade: Another newer Python library created because its author felt existing open-source libraries lacked certain functionality. It supports features like headless login and access to quote data, which is crucial for building a fast scanner.

These libraries handle the low-level API calls, allowing you to focus on writing the core logic for your butterfly spread screener: defining the strike combinations, calculating the debit, and applying your mispricing filter.

2. No-Code/Low-Code Platforms (Alternative)

If you find the coding process too time-consuming, you could use a third-party platform that connects to your E-Trade account:

1. TradersPost: This platform enables automated trading bots for various asset classes, including options, and has an official integration with E*TRADE for auto-submitting orders based on your defined strategy or signal. 5This would handle the execution part of your plan.
2. Option Alpha or SpeedBot: These platforms are built specifically for options trading automation and often use a no-code or visual editor approach.6 They may offer pre-built bot templates and advanced scanning features that you can adapt to look for high-reward, low-debit setups, though they may have fees.

Starting with one of the Python libraries will give you the most flexibility and control over your specific "mispricing" definition.
Would you like to start outlining the Python code structure for connecting to the E-Trade API and fetching an option chain for $SPX$?

Let's do that when I come back. I need to read up in the resources you mentioned and do some background learning so I can attack this problem in a rational way.

That sounds like a perfect plan. Taking the time to understand the E-Trade API and the Python wrappers will make your scripting efforts much more efficient and successful.

Recommended Resources for Background Learning

Based on your goal to automate the search for mispriced $SPX$ butterfly spreads using E-Trade, here are the key areas to focus your reading:

1. E-Trade API Documentation: Review the official developer documentation, specifically focusing on the endpoints for Authentication (OAuth), and Market Data (like get_option_chains and get_quote).
2. Python API Wrappers: Read the documentation or review the code examples for a library like pyetrade to see how it handles the complex parts of the API interaction.
3. Options Pricing and Greeks: Focus on understanding how Implied Volatility (IV) affects the price of a butterfly, and how the Delta of the central strike ($K_2$) impacts the potential profit.

I look forward to helping you outline the Python code structure when you're ready! Good luck with your background research.

I am basically a cave man when it comes to this type of stuff. But, I fully expect to be able to master the API approach to finding msipriced SPX spreads. Gemini will help me.