Michael Stigall is an engineer who’s currently “in training for his retirement as a trader.” He began using neural networks when one of his engineering jobs led him to model plant processes from a plastics company with one of Ward Systems early products. In the 1980s he began thinking about using the same technology on the stock market. He learned about the markets from a fundamental analysis trading club that he joined, but his engineering background made him more comfortable with technical analysis.

For the past 15 years, he funds his "trading experiments" with $5,000 at the start of each year. Using the NeuroShell Trader, he builds a new model each year, He tests the model during the first month of the year and obtains his expected monthly % rate of return as a goal. At the end of the year, he gives all of his profits to his wife for long term investments. He retains $5000 to fund the next year’s experiment. During those 15 years, he's only lost money in 2007.

This year his goal is to earn 10% return per month. In February and March the actual average returns were 15% monthly return. Over his 15 years of trading he said he averages 240% annualized return, which amounts to about $15,000 per year. He said if he can ever make $60,000 profit in one year’s experiment, he’ll retire.

His early models worked on trading computer stocks. He moved on to indexes, which he regarded as an average of stocks and easier to predict. In 1994 he saw that the market was choppy and flat, so he decided to take a look at stock options to leverage his money. He’s written a program that checks for available options on QQQQ that meet his criteria.

“I use QCharts/ESignal as my data provider and they provide “option montage” reports for any particular stock. I trade the QQQQ exclusively. Using QChart’s interface, I selected filters for In-the-money options, Puts and Call both included, and no LEAPS. “

He’s created an acronym for his option selection criteria which he calls DIET, as in “Do Options on a DIET”. He categorizes this as a very paranoid, conservative approach to options trading. The letters stand for:

Delta (how much an option moves relative to the underlying instrument) – must be greater than 0.85

In the money – limit to in-the-money options only

Expanse – this is a term that relates to how much interest a particular option has. If a particular option closed yesterday at $2.00 per share ($200 per contract) and 100 contracts where traded yesterday (option contract volume) then the total dollars traded yesterday would be $200 per contract x 100 contracts = $20,000 traded. If my little portfolio was $5,000 and I wanted to invest it all then my trade would represent 25% of yesterday’s volume. I don’t want situations where the liquidity of the option could limit my ability to exit if necessary. Since the amount of my portfolio changes quite a bit, I just calculate how many dollars traded yesterday and do the percentage calculation in my head.

Two months or more until expiration – my typical elapsed time in an option trade is 3-5 days. The Theta value (a measure of time decay before the option expires) doesn’t impact my trade much IF I purchase an option with some significant time to expiration. I’ve traded options near their expiration dates, and find my stomach treats me much more kindly if I build in a significant safety buffer between my trade and that final expiration date.

For this year’s model, Stigall used the NeuroShell Trader to build a neural network prediction of a combination of change, volume, RSI, and Bollinger Band indicators which he then fed into a cluster indicator. He used this cluster indicator to build trading rules. He monitors the contribution factors of each indicator in the neural network. He notes that if the contribution factors change significantly then he believes the underlying market trend is changing. He optimizes his model built on daily data every day.

He looks at a trade as being composed of 1) timing, 2) direction and 3) magnitude. He expects to only be able to model 2 out of the 3 constraints. He said picking direction and timing is easy; “It’s the magnitude that’s hard.” He’s now working on the best method to determine stop loss and trailing stops. Stigall uses his DIET analogy in another way to keep his trading true to his models. “When one goes on a weight-loss program, they shouldn't keep the fact to themselves. Tell all of your friends about your new effort. When they see you reaching for a piece of cake after a meal, they'll kid you (or remind you... or both) about your diet. This keeps you focused and honest. I try to use a similar methodology on my trading. A small cadre of my friends (ten people currently) is informed the previous night via email of my planned trade, and is invited to watch the trade. If I break one of my stated rules for trading or miss one of my goals, they will kindly remind me and keep me honest.” Stigall characterized trading as his hobby, “like some people rebuild cars. “ He’s been a computer consultant for 30 years, moving around the country, but now he is with a biotech company in California and he’s planning on being there for a while.

Because he works, he doesn’t want to watch a screen all day. He says he has an advantage by living in California, because he can get up at 6:30 am and see how the market opened, and then he can place his trades. When he returns home at night, he can update the data and see if the action is needed for new or existing trades. “I don’t think my boss would be happy if I had stock market screen open on the desktop all day… this way my hobby doesn’t impact my work.”