No. of Recommendations: 15
Here's a GTR1 backtest (treated as a long screen) from 19970902 through 20121004 (sorry, I've been too busy working on the user interface to update the database for the last few days):
step0: [Actual closing Price; share_lag=1 days; quote_lag=1 days] > 5
step1: [[Actual Average daily Vol 63d; lag=1 days]*[Actual closing Price; share_lag=1 days; quote_lag=1 days]] > 500000
step2: [Security Type] != 30,31,36
step3: [Fama/French Industry Code] != 48
step4: [SI Inventory Q5] > 0
step5: [1*[SI Inventory Q1] - 1.4*[SI Inventory Q5]] > 0
step6: [SI Accounts Receivable Q5] > 0
step7: [1*[SI Accounts Receivable Q1] - 1.4*[SI Accounts Receivable Q5]] > 0
step8: [[Actual Average daily Vol 10d; lag=1 days]/[Actual Average daily Vol 63d; lag=1 days]] Bottom param0; Long, Cash When Fewer
Holding period = 20 mkt days; Fully rebalance every 1 periods

Here are the results when the 5-stock variant is given a weight of -1 (short) in a blend with 2 parts cash (which I can only guess is what does):

                Avg     Min      Max      SD
CAGR: 9.86 1.11 18.40 4.97
TR: 403.96 17.98 1162.36 327.58
GSD(20): 51.14 45.61 55.74 2.52
DD(20): 32.60 27.64 36.75 2.28
MDD: -67.66 -75.10 -56.66 4.65
UI(20): 32.92 24.31 41.63 5.46
Sharpe(20): 0.37 0.18 0.56 0.11
Beta(20): -1.45 -1.62 -1.28 0.08
TI(20): -10.60 -15.91 -5.18 3.12
AT: 10.99 10.84 11.21 0.11

An average CAGR/GSD(20) of 10/51 vs 24/43 for There seems to be enough of a difference between our short screen backtests that it's worth a serious investigation. I'll soon be allowing VL and SI Pro subscribers to get current screen picks and field values one way or another (if no one offers to help with some of the more sophisticated options for validating subscribers, I'll just go with something simple such as the MD5 hash of the latest installation file as a password), thereby opening the backtester up to more scrutiny.

In the meantime, people with more time than I have can (1) compare the long version of this screen at with my long version and (2) get daily portfolio values for the long version from my blender and verify that the daily portfolio values of the blend make sense. My backtester applies daily margin interest to the value of shorted positions and cash interest on the entire cash portion of the blend. If "Margin Premium" is set to 0, then those two interest rates are equal, and so in the case of a -1:2 blend, the entire portfolio is effectively earning cash interest on roughly its net value. I have no idea what does as far as interest (margin or cash) is concerned.

Robbie Geary
Print the post  


When Life Gives You Lemons
We all have had hardships and made poor decisions. The important thing is how we respond and grow. Read the story of a Fool who started from nothing, and looks to gain everything.
Contact Us
Contact Customer Service and other Fool departments here.
Work for Fools?
Winner of the Washingtonian great places to work, and Glassdoor #1 Company to Work For 2015! Have access to all of TMF's online and email products for FREE, and be paid for your contributions to TMF! Click the link and start your Fool career.