Business Intelligence and Handicapping

[traynor]

Business Intelligence (also called business analytics) is a field that goes beyond the spreadsheet/database accumulation stage to developing techniques to analyze and predict events by viewing data in different ways. Much of the information consists of fairly simple concepts that may be usefully applied by handicappers in their own race analysis processes.

Easy-to-read overview of BI, along with some great links to other information sources:

http://blogs.techrepublic.com.com/decisioncentral/?p=142&tag=nl.e106

Good Luck!

[JBmadera]

I've worked for BI companies for the past 10 years (bobj and cogn) and while their reporting tools may be helpful for standard slicing and dicing of data I'm not sure their predictive analytic tools are on par with a SAS for example. I toyed around with the idea of putting CD on top of access/orcl db but in the end I went with HTR and use the robot function - to me it's much more about interpreting the data then gathering the data.


good luck!

jb

[njcurveball]

My own tests using Oracle business intelligence were very fruitful in terms of ROI. But to play that output forward is an exercise in futility. To a computer a 69.8 K rating may be very predictive of a positive ROI in past events. In real life, those conditions wind up more back fitting than anything.

Someone designing a business intelligence model must first teach the computer handicapping concepts. The factors are far from random and far from independant.

The most important thing is value as compared to the crowd. In another thread a long losing streak was laughed at. And if the game were just about picking winners, most of us probably would not lose 10 races in a row.

As said all too often in the past, it is not about picking winners. Most casual players cannot accept that simple fact. They box the 3 favorites and when the trifecta pays $60, they brag to their friends how their $12 box made money! Unfortunately at the end of the day, they have spent $120 to make that $60.

The different layers of a model are handicapping basics, value of handicapping factors, and value of the bet. For me to train a business model on all of these things is a very time consuming effort.

Querying a database on combinations of factors that have proven profitable is much more productive.

As stated above, many users who could not design a single table, do very well using the robot in HTR.

Also, many users who can design complex computer applications also do very well using the robot in HTR as well.

Jim

[LottaKash]

My own tests using Oracle business intelligence were very fruitful in terms of ROI. But to play that output forward is an exercise in futility. To a computer a 69.8 K rating may be very predictive of a positive ROI in past events. In real life, those conditions wind up more back fitting than anything.

Someone designing a business intelligence model must first teach the computer handicapping concepts. The factors are far from random and far from independant.

The most important thing is value as compared to the crowd. In another thread a long losing streak was laughed at. And if the game were just about picking winners, most of us probably would not lose 10 races in a row.

As said all too often in the past, it is not about picking winners. Most casual players cannot accept that simple fact. They box the 3 favorites and when the trifecta pays $60, they brag to their friends how their $12 box made money! Unfortunately at the end of the day, they have spent $120 to make that $60.

The different layers of a model are handicapping basics, value of handicapping factors, and value of the bet. For me to train a business model on all of these things is a very time consuming effort.

Querying a database on combinations of factors that have proven profitable is much more productive.

As stated above, many users who could not design a single table, do very well using the robot in HTR.

Also, many users who can design complex computer applications also do very well using the robot in HTR as well.

Jim


Sank yu velly much, dis is velly intellesting to me.......:ThmbUp:

best,

[Robert Fischer]

If you can apply recent trends and real time data to what has been successful longterm ...