Here is my latest/greatest LargeCap portfolio simulation. The custom universe screens for:
AvgDailyTot(60) > 20000000 (20 Million)
MktCap > 5000
I will be keeping this system private. I’m just letting everyone know that good results are possible with high liquidity.
By the way the ranking system has only two factors 
Steve
LargeCap Summary.pdf (42.6 KB)
LargeCap Stats.pdf (31.1 KB)
LargeCap Performance.pdf (72.6 KB)
LargeCap Universe.pdf (65.3 KB)
Steve:
Those are encouraging numbers for the 5 year test, especially getting over 40%/year with a low draw down of -15%. Well done.
I am curious have you tested this on the tough 18 months of data before November 28, 2002?
Brian
I’ve put together a large-cap portfolio that I have been playing with the last couple of days with the new Benchmark functions added.
Like Steve’s Port, this portfolio also has the following parameters:
MktCap > 5000
AvgDailyVol>20000000
It uses the Stitts Robust Rev 2 ranking system and the Wilshire 5000 Benchmark.
I’ve been having good luck with my newsletter since I started timing the market, usually using data provided by InvestorsIntelligence, but I picked one of the key market-timing indicators I use and added it to this port and the performance picked up very nicely, with a return of 88%+ for the same time period (1 year) and a drawdown of less than 7%.
Chris
Olikea -
I wasn’t testing on 2001 because I am using PrevRank() - causing sim to crash if I start from 3/31/2001. I have now gone back and ran from 4/31/2001 and got results which mimick the S&P500 for 2001 and up 20% over the S&P500 in 2002. I’m not going to get hung up on that though since we may never see those market conditions again (bubble burst). And the system has a naturally slow startup. It is one of those port where you want to carry over the positions from the simulation when creating the port. It has to do with the initial buying conditions resulting in all stocks selected being depressed.
Steve
How does this SIM hold up on a larger number of holdings and with robustness testing?
Charles - I tested at least 15 different time periods at random. The one year performance is better than the S&P500 benchmark for all time periods tested. The system is set up to buy 10 stocks. If I increase to 20 stocks then I still get ~30% annual returns but the trading account is no longer 100% utilized. I would have to change the buy/sell rules to accommodate larger number of stocks.
Steve