Steps in Advanced Settings

Hello,
   Below are the steps in advanced mode :

After hitting 'New' you enter the stock symbol and the name of your model. I then keep the historical data set to 366 days, as is the default. I use the EOD data as opposed to EOW. I then pick 'advanced mode' and hit 'next'.

  After this I get to choose the market behavior or characteristics: I tend to choose 'fast normal' as I have yet to experiment with other settings. You can choose from Chaotic, Fast Normal, Slow Normal, Inert, and 'Very Inert'.

Also, it will ask you for a setting of 'noisiness' or the preferable level of data filtration which can range from 'No Filtration' to a setting of 1 through 6. I see this, from my past experience in programming, as a way to set the different layers of processing with the current model. Then I hit 'Next'.


After this I get to choose the type of trading strategy I want to use: Speculative, Short term 'Buy and Sell', or Investment 'Buy and Hold'. I tend to choose 'Buy and Hold'. After that, I get the choice to choose the desired signals frequency from 1 day to 30 days. I choose 1 day. Then I hit 'Next'


Then I am asked about the level of model complexity & genetics implementation : The level of complexity can be Simple, Standard, Complex, or Manual. When you choose 'Manual', then you are asked about the structure of the neural network, or the number of layers of computation which can range from 1 to whatever you want. Just remember, the more complex you set this, the more time will be needed to calculate the results. The program will then ask you for the number of 'neurons' for each layer that you choose. The 'neurons' are  simply the exact same thing as neurons are in the brain... they are the nodes, or data sets' given for each 'layer' of the structure of the network. We could talk about this for hours, but let's go on right now. After choosing, you hit 'next'

Then you will be asked if you want to reserve a percentage of historical data before feeding the model forward. You can say yes or say no..use all of the available data going forward. If you do reserve some days for an "out of sample" period, I suggest no more than 7 days. I've tried 30 days before, but it didn't work out too well. 7 days worked fine for me in later testing.

After this, you are finished. Just try some experiments with different models. I was surprised with some of the results that I got. What works well for one model may not work well for another.


Have a great day,
Roger

Advanced Settings

Hello,
   I was in for a bit of a surprise this evening. I decided to model one of my stocks in 'advanced' mode to see what would happen, and I actually came up with a system that outperformed the one I had already been trading from for the entire year. I started by going to 'new' then choosing my old friend MTG and naming it 'MTG ADVANCED'. I left the history download at 366 days, but changed the model to 'advanced mode'.

 

After hitting 'next' the screen asked me to choose the expected market behavior which offered me chaotic, fast normal, slow normal, inert, and very inert. I chose 'slow normal'. This is really a setting that is based on how you feel the current market is behaving, and I felt that slow normal was appropriate. At the bottom it asked me for 'nosiness'  or the filtration level desired which I set at 2. I am not really sure what this meant, but I am sure it has something to do with the offset of unusual movements in the stock.

 

 

After hitting 'next' another screen asked me for the dominant trading strategy I would be using. I could choose speculative, short term buy and sell, or investment buy and hold. I picked investment buy and hold. I was then asked for the desired frequency of signal which I set at 1 bar.

 

 

 

After hitting 'next' I was asked about the modeling complexity and genetics implementation. I picked 'manual' so I could se the 'genetics' myself which I set at 4 layers with 60 neurons per level.

 

 

 

Next, I was asked about back testing which I said no to, as per my prior experience with back testing, so I set it to use all of the available data.

 

 

 

 

Then I got to the finished screen

 

 

After that a chart popped up, and I just went to the bottom right and hit 'OK'

 

 

Then I was asked if I wanted to start the learning process, and I hit 'yes'

 

 

 

 

When all was finished I ended up with a system that was much better than any of the 5 I had run earlier this year. Much more gross profit, and with a couple more trades, all still with the winning percentage of 100%.

 

 

 

So now I am going to start using advanced mode more. I should have explored it sooner. This endeavor has turned out to be a learning tool for myself also. I hope I can be of more help going forward.

 

 

Have a great day,

Roger

 

 

 

 

 

 

 

Days Reserved For Historical Testing

Hello all,
   I decided to do some testing using the option to reserve days for historical testing which I set at 30, as opposed to using the value of zero days for historical testing. I have found over the years that the default setting are usually the best, but I like to experiment with all the settings. In another post I will get in to using the 'advanced mode', but for right now let's look at the option of reserving a historical period for testing when you are setting up a new model. Of course, you begin with setting your trading strategy. Just for testing purposes, I set the account to $1,000.

 

I am going to choose the stock 'MTG' since I am familiar with it over the past year. I am going to hit 'New' then name the model MTG-1, and set the reserved historical period for testing at 30 days.

 

 

 

Then the program will ask if I want to start learning. I hit 'yes'

 

 

 

 

Then it will ask if I want to set the out if sample period to zero, which I answered 'No'

 

 

 

After the learning process is finished, then I hit the strategy performance tab and I saw this:

 

 

 

So with 30 days set for the historical period for testing, I got the results shown. Gross profit was $799.43 with 18 trades for a winning percentage of 88%, with 16 winning trades and 2 loosing trades.

 

Then I started all over again, setting the reserved historical period for testing at zero days. I started the learning period again, and when it finished I got the following results :

 

 

Gross profit was $840.71 with 14 trades for a winning percentage of 100% with 14 wins and no loses. To me, I think it is better to just leave the reserved days for historical testing at zero. But, others may have different results. This was just my experience and opinion.

 

In my next post I will get into the advanced mode where you can set your own parameters for testing. I will explain the various settings you can choose to customize your own specific model compilation. In general, I have found the default settings to be great. I just run the 5 available model types. But there have been times that manually setting some things can produce tremendous results.

 

 

 

Have a great day,

Roger

 

Choosing Stocks to Model

   In Stock Neuromaster, any stock is tradable. If you go long or short, either is a viable candidate for use with the software. I generally went 'long' when I first started using the software. I look for stocks that met basic requirements: 1) I look for stocks that are trading for less than their book value. 2) I    look for stocks with a positive EPS (earnings per share) 3) I look for stocks with a positive Operating Margin 4) I look for stocks with a positive ROA ( return on assets ) 5) I look for stocks with a positive ROE ( return on equity ). Just these 5 areas allowed me to 'hone in' on stocks that I was comfortable with. You may have other criteria, such as RSI, looking for whether a stock is oversold. But these are short term indicators, and are not really relevant to the long term analysis of a stock as far as modeling is concerned. I am only concerned with the stability of a stock and its financial viability, as far as modeling is concerned.
   I am pretty sure that all of you know what 'shorting' a stock is. For those that do not know, you are essentially 'selling' a stock you do not own, then buying it back at a later date, hopefully at a price that is less than what you 'sold' it at. It is a hope that the stock price will go down as opposed to going up. You are dealing in the number of shares. If you short stock 'ABC' for 100 shares at $10 a share, you are promising to return those 100 shares at a later time. If you 'cover' that purchase a while later at $5 a share, then you have made $500 on that trade. You 'sold' 100 shares for a thousand dollars, then you replaced those 100 shares, but at a price of $5 ($500) and you pocket the difference.
   Stock Neuromaster covers long purchases as well as short purchases. If you are confident enough, you can go to 'settings' and change you signals from 'long positions only'  to both long and short positions both. I have only used long positions myself, but I am looking into doing both, especially with a very wild trending stock that fluctuates on the charts from one extreme to the other.
    That's another thing I look at , are the charts. If I see a stock that looks like a right angle, I don't see much action other than it going up over the year. I tend to look for stocks that fluctuates into a pattern of 'up and down'. I know that a stock that seems to be going 'to the moon' can be tempting, but I prefer a stock with a discernible pattern. I want the software to easily distinguish patterns, and any help I can give to start with is always preferable, at least to me.
    These are just ideas for starting. I only keep about 5 positions open at any time, as more than that makes my head swim. I sit back and let the software do the work. I just need to keep more on top of it, and wake up on time each day, lol. When you are retired, there is a sense of relaxation that comes over you, but when it comes to money, you really can't relax!
    Plus, it helps to have a great sense of humor. You can't live without levity once in a while! For example:

I bought some powdered water yesterday, but I don't know what to mix it with !!!!!

Anyhow, you get what I  mean. Going forward, I will try to be more serious, as money is the focus of all of this. Everyone deserves to have the style of life that they deserve. How you handle your money is one of the most important things, not only for yourself but for your family. I will try my best to help.


Have a great day,
Roger




   

    

Trading Example

Earlier this year in January, I ran a series of 5 models on MTG. I picked the model with the highest winning percentage of 100% and decided that I would follow all of the buy-sell indicators for the entire year, and I am glad I did. My initial investment of $10,560 turned into $13,608 from January 3rd to November 1st for about 129% gain, and that was just on one stock.

I took into account my brokerage fees. Even just a start of $1,000 would have returned $1,288.70.
I tried to follow all of the buy-sell alerts exactly, but there were some days that I forgot, or overslept, LOL, causing me to put my orders in a little late in the day. Sometimes those worked out OK, but other times it didn't. All in all I think this one stock did well.

I posted this not to brag, but someone wanted to see some results of what I have personally done. Plus, all I did was to update my models each day, then follow the buy, hold, or sell signals. Was this the best stock that I have held. No. Not by a long shot. But this is the one stock I have held the longest, following the buy, hold or sell signals exactly. It doesn't take a rocket scientist to do this either. Anyone can do the same thing. I will try my best to help with the use of the software.

Have a great day,
Roger

Comparing Models

Hello,
   I want to go over in more detail comparing models. First, let me tell you a little about creating models when you first start. This is the screen that you start with :

I recommend leaving everything set the way it comes to start with. Leave the simple mode checked. We will get into the advanced mode later. I choose the 'End of Day' field check as opposed to the 'End of Week'. I also leave the start date for downloading history at the preset date, although you could go back more than a years worth of data. The reserved historical period for testing I also leave the way it comes set. The only thing I change is the model type for each of the five models that I am going to run. This is why I rename each model to reflect the model type it represents, thus FORD-1 through FORD-5.

Anyway, when I get all 5 models of the stock run, I then want to do a comparison of all 5 systems. This is easily done by going to the 'Analysis' tab at the top and choosing 'Strategy Performance Comparison'.

 

This will bring up a box of all models that have been run. In this case I want to compare all 5 of the Ford models I have created:

 

 

 

 

I click the 'OK' button and let it run. It then will produce a report comparing all 5 models as such:

 

This way I can compare the performance of each model and select the one I want to go with. You may want to choose more than one model, but I usually choose the best one as far as total net profit, and the winning percentage . But that's just me.

Originally I ran all 5 of these models with a stop-loss of 5 percent. Then I ran all 5 again with no stop-loss just to see the difference in the models. I made myself a little spreadsheet to compare stop-loss vs. no stops. But I could have easily let the software do it for me just by creating models that were differentiated with a different model name, like FORD-1 SL for the type 1 model with a stop-loss or FORD-1 NSL for the type 1 model with no stop-loss. I actually should have done this, but below is the comparison I came up with between the two sets of models :

After adding the lines of net profit up on both sets of models, I saw no clear winner using a stop-loss vs. no stop loss. Although the largest net profit came from FORD-3 without a stop loss, and with the most number of trades (20) , the model FORD-4 with a 5 percent stop showed 100 percent profitability with a total of 14 trades.

 

At this point I could choose both models to run with, renaming them FORD1 and FORD2, and just deleting the others. But the choice would be up to you.

 

You might ask "How do I choose the stocks to model?" In my next post, I will tell you how I pick the stocks that I use in the software. One thing I will let you know is that this year alone I have made a total gain of just over 128% on just one stock I have traded!

 

 

Have a great day,

Roger

Using Neuromaster

   Let's get right into using Stock Neuromaster. As I stated before, any stock is tradable whether you are 'going long' or 'shorting' . I am going to begin by focusing on long positions. I will go over shorting stocks at a later time. Let's begin by choosing the settings you want to use in the software. Just open up the software and choose the 'settings tab'.

Then choose trading strategy

This is where you can choose to model long positions, short positions, or both. I am choosing long positions only, and I am setting the account size to $1,000. Any account size works for me, as I am only using it for comparing models to choose the best one.

                                          Then I am going back to 'settings' and I am going to set my stop-loss settings. I am just going to do a 5 percent stop-loss, but I usually don't use any stops for modeling purposes. I did it this time so I could compare the models using a stop with those not using a stop just so I could personally observe the difference.

 

 

Next, I am going to click on the 'new tab' and set my first model to run. I am going to pick a stock, any stock, and I randomly picked Ford which trades by the symbol of "F".

 

 

You simply put in the symbol, them choose the name of your model, which I chose as FORD-1, as I plan on doing a total of 5 models from Type'1' to Type'5'. When you hit 'finish' the program will immediately tell you how many 'stops' were activated, then it will ask you if you want to start learning. You hit 'yes' of course and the software starts doing the learning process for your model.

 

 

When your model has finished learning, then you can do a new model, which I always do all 5 of the preset models. In this case I named each model FORD-1 through FORD-5.  After all models are run, then I go to the 'analysis tab' at the top and choose 'strategy performance comparison'. This way I can choose the best model I have created for using going forward.

 

   In my next post I will get more in depth. There are so many more tools and things you can use to help. This only scratches the surface of what is possible, and I will do my best to go over everything so you can get the most out of Stock Neuromaster. I have been using this software for over 8 years now, and you frankly would not believe the results I have obtained, so I will not try to give you a bunch of numbers. Let's just say that I am happy with the results.

 

I will try to get better screenshots of every step of using this software. This was just a very quick basic start-up. I would be interested in what all of you would like to know or learn. Much of what you might need to know would be based on your experience with equity trading. I don't want to be too simple, but I also don't want to leave anybody in the dark so to say. Thanks for all of your comments so far.

 

 

Have a great day,

Roger