Recently in Random Category

Happy New Year of the Snake my Asian readers!  I'm so glad that 2012 is over because it was a terrible year for me.  I've been stuck on a soul crushing project, which continues to make me miserable and absolutely demoralized.

I've transferred the blog to moveable type, hopefully minimizing the hacker exploits.  Wordpress, while an easy to use blogging platform, is riddled with security holes. Likewise, the same can be said with all of them (Expression Engine, Text Pattern), but I hope this nuttiness is behind me now.

I use Google Analytics and I probably should use more it in conjuction with Rapidminer to do more data modeling, but it's a question of free time; something that's a luxury right now. For fun I'm posting my top 10 referrers for the period of February 17 through March 17, 2011. Of course Rapid-I is on top of that list and thanks to them and all the other blogs and websites for linking to this site.  You all make me smile and without you I'd probably be just another loser on the Internet.   What's interesting in this referrer snapshot is that I have a relatively low visitor rate (highly niche blog) but my average "time on site" is over 4 minutes, with the last month time period clocking in over 7 minutes.  That's because of my video tutorials, the biggest driver of traffic for this measly blog.

Thanks to all who signed up for their very own Dropbox account using my referral link! I went from 2 gigs of space to just over 5 gigs! I can now move large files easier between computers (without using memory sticks), and be able to collaborate on Rapidminer projects with people.

Good Morning!  I've been using Dropbox as a way to share Rapidminer Experiments and Data with readers.  It's vastly better at sharing large data files than my forums and I might completely abandon my forums (for now) in favor of consulting with Dropbox.

If you haven't used Dropbox before, I suggest you get it.  Trader W0NK0 was the guy that clued me into it last year and since then we've been sharing market data, python files, and silly stuff.  Make sure to signup using my referral link so I can get more free space with every signup.

I felt like a kid in a candy store tonight when I picked up the daily mail.  I got my first shipment of old German money notes in the today, the ones I won on Ebay for about $1.75.  Below is a photograph of an almost 100 year old 1000 Reichsbanknote.  It will be 100 years old on April 21st of this year.  Its in really good condition and beautifully decorated.  Even if its fake, you can't go wrong for $1.75!

(Front - note the date below the 1000)

(Reverse)

Its hard to believe but the German people, just a mere 100 years ago, probably never fathomed that their Reichsnotes would never lose value and be sold on Ebay for a fraction of its original value.

When I was I kid I used to collect stamps and baseball cards, just like nearly every other American kid does. Then, about 3 years ago I started collecting US coin sets, mostly proofs from the US mint for my kids (they make great presents BTW).

As a lover of history and economics, coin/money collecting (numismatics) just made sense to me and now I go to the monthly coin show in my area and spend a few hours a week on Ebay searching for for interesting coins and paper money for my collection.  Just this week I picked up something that reminds me of runaway inflation from the past.  Its a 1923 - 20 Million Marks bank note from the Weimar Republic of Germany.   Inflation was so rampant during the Wiemar Republic that people used these notes as wall paper.  

Since I want to own a piece of history, I bought it for $1.75 from Ebay.  Goes to show you, one man's wall paper is another man's treasure.

A great master at the game of GO once wrote that many advanced players begin to trip themselves up when they forget the basics of ladders and nets.  Ladders and nets are just simple beginner strategies that are grounded in the basic rules of the game.  These simple game formations are much like a stock's price and volume, and its technical chart.

Going back to the basics of understanding candlesticks, prices, volume, and formations associated with an asset can make the difference between success or failure.  Basics are important and now is a perfect time to go back and study a few.

I wanted to take a moment and say thanks to Tibor, one of my 12 readers, for forwarding me some really interesting research papers on modeling NFL and NBA games with neural nets. There are some really good nuggets of information in those papers, especially the discussion on setting the right momentum and learning rates.

I used that information to fiddle around with my neural net model and I'm posting some recent results from my DRAFT NFL neural net point spread model.  As you can see, the model is pretty good at determining if the home or visiting team will win (a negative sign means the home team wins) but the predicted spreads are way off relative to actual spreads.

This leads me back to developing some sort of ranking system to feed the model, which I wrote about in my "Thoughts on Ranking Football Teams" post.  The good news is that the research papers that Tibor sent me allude to a type of football match system where the model learns the results of previous games and then applies its statistical analysis to new match ups.  Despite this good nugget of information, I feel that I have a long way to go to get something solid before the season starts.

In the interest of science, and because I love my 12 readers, I'm uploading my EasyNN Plus data file for this particular model.  However, you'll have to have the full version of EasyNN Plus to use this file because the model uses 980+ example rows and the test version only allows you 100 rows.  If you follow the link above and buy EasyNN Plus from there, I will get a small commission from Steve.

 

I've made some serious headway last week in analyzing NFL football data to model game point spreads.  I was able to determine with great accuracy (84%), using backfitting, what team would win a matchup. I did this by building a backpropogation model from 2007-2008 data with about 20 matchups as the prediction set.  The only bad thing I discovered was that the point spread prediction was way off.

As I wrote before, backfitting is NOT the way to go for building this type of  model and the fact that the magnitude of the actual point spread was off bore that out.  You might be asking yourself right about now, "if this is not the way to go, why did Tom do this?" 

I did this because I'm still in my data discovery phase looking for relationships and tinkering around.  Now that I understand various data relationships and can detemine the winners relatively well, the next step, and perhaps the hardest now, is determining the ranking system.  I suspect that the ranking system will help get the game spreads closer to what they should be for future games.  If all goes well I should be making spread calls on this blog for the next season as a way to determine if my model is worth a sh*t.