Developed by

 
 
TSSB book on Amazon.com
 
     
 
Why is TSSB Free?
All software is in a constant state of evolution and TSSB is no different. In its current state we believe TSSB is extremely valuable. We are aware of no other product on the market that facilitates the application of machine learning to predictive-model based trading systems that copes with the issue of data-mining-bias in a statistically sound manner.  However, we would like to continue its development and have numerous ideas for its improvement that will make it even more valuable. We think the best way to do this and recoup and possibly profit from our investment is to offer the software for free and charge only if support is needed.
If TSSB is Free What is the Revenue Model?
Our primary source of revenue is from sales of the book Statistically Sound Machine Learning for Algorithmic Trading of Financial Instruments, a step by step tutorial for learning to obtain maximum benefit from TSSB. We also offer support services should users desire additional assistance. We hope that satisfied users will make voluntary contributions. Consulting services are also available on a case by case basis, please contact us for details.
Who funded development of TSSB and at what cost?
Hood River Research Inc. and investors funded its development at a cost of over $100,000 entailing more than 3,000 hours of work. The product reflects over 50 years of combined experience in trading system and filter development work by David Aronson and Dr. Timothy Masters.
What is the cost of support?
It depends on the type of support.  The most economical is via email and we suggest that all support interactions begin in this fashion. Should a higher level of support be required, such as consultation by phone or phone plus screen sharing via Go-To-Meeting that is also offered but at higher cost.  Email support is billed at $150 per hour paid in advance via PayPal.  The minimum is one hour with any unused  balance carried forward.  Phone consultation is $250 per hour, with a one hour minimum paid in advance and any unused balance carried forward.  Phone consultation with screen sharing is $300 per hour, two hour minimum paid in advance with any unused balance carried forward.

Support packages in 10 hour increments at discounted rates are available. Contact us for details.

Who is the intended audience for TSSB?
Anyone involved in applying machine learning to trading system and filter development and interested in obtaining unbiased performance estimates and statistical significance levels adjusted for data mining bias.
How can I decide if TSSB is for me?
Reviewing this website means you've definitely taken the right first step. Visit our Product Info page for more details or visit our Downloads page to view the TSSB Tutorial & Manual. Our tutorial is progressive and designed to bring all readers up to speed on even the most sophisticated quantitative methods. 
Is TSSB compatible with other software?
TSSB can import or export standard ASCII text files or Microsoft Excel spreadsheets, but is designed to be self-contained.
What Some Things Can TSSB Do?

Our tutorial and manual cover all of TSSB's capabilities. Here is just a sample that gives a sense of what it can do.

 

Develop and test predictive-model based trading systems for individual instruments as well as multiple markets and provide unbiased out-of-sample performance and in many instance p-values that are robust to data mining bias. All development is machine-learning based.

Develop and test market neutral (long/short) that trade portfolios of instruments.

Develop and test signal filters for any existing trading system.

Develop complex trading system architectures completely integrated with walk-forward or cross-validation testing
     Committees (aka ensembles)
     Oracles (intelligent committees with conditional differential weighting of components via gate variable)
     Regime Specific Trading Systems via Event-Triggering

Develop optimal portfolios of trading systems based on individual models, committees, oracles based on the unbiased out-of-sample performance of each candidate and then walk-forward the portfolio to get its  unbiased out-of-sample performance.

Create Signal Filters to eliminate just a few of the worst trades. Most signal filters retain just a few of the best trades. TSSB can do this as well but also permits the development of filters that retain most trades except the very worst.