| Index | Recent Threads | Who's Online | User List | Search |
|
|
 |
Portfolio123 » List all forums » Forum: Ranking Systems and Factors/Functions » Thread: A Good Small Cap Ranking System |
|
Total posts in this thread: 18
|
[Request a Feature]
[Post new Thread] [Add To My Favorites] [Watch this Thread] |
| Author |
|
|
DennyHalwes
Advanced Member UNITED STATES Joined: Apr 28, 2004 Posts: 1532 Status: Offline |
All, As a result of my discussion on page 2 of this thread I thought I would see how well a similar approach would work for the Russell 2000 stocks. I selected one of the best public ranking systems, the BompusRank + TF 50/50 system, and added a MktCap factor to the system. I set the factor so that the smallest cap stocks had the highest rank value. After playing with the weightings a little I ended up with this ranking system. This system achieved the highest performance for the universe of the Russell 2000 stocks of any system that I have found other than a few pullback systems. See the results below set for the smallest 10 stocks in the Russell 2000. These are stocks with a minimum market cap of about $780 Million: Denny  ---------------------------------------- 
Performance.JPG
(71979 bytes)
(Download Count: 719)
---------------------------------------- "The significant problems we face cannot be solved at the same level of thinking that we were at when we created them". Albert Einstein ---------------------------------------- [Edit 2 times, last edit by DennyHalwes at Feb 11, 2010 2:51:31 PM] |
||
|
|
strader1
Advanced Member
|
Denny: You are just awesome! Thanks for all the wonderful posts. -- Bill |
||
|
|
strategyx
Member
|
Very impressive. Thanks Denny. |
||
|
|
gfagerlin
Advanced Member
|
With 138% top bucket return there has to be a way of getting a good sim from Denny's ranking. It has been a while since there has been a public sim development so I thought I would put a sim together and make it public for others to improve. Here is a start. My goal was 100% AR with -20% DD. I expected the challenge would be to tame the Bompus part of the ranking. Looks like the benchmark from here works pretty well but the last few moths have been a challenge. Glenn ----------------------------------------
Untitled.jpg
(17750 bytes)
(Download Count: 646)
---------------------------------------- [Edit 1 times, last edit by gfagerlin at Feb 12, 2010 4:26:18 PM] |
||
|
|
dwpeters
Advanced Member
|
Hi Denny, thanks for sharing the ranking. It has done remarkably well the past couple years. Glenn, my first effort was pretty comparable to yours, but then I tried combining our timing strategies. Sim is here , and may not stay public for long. 160% AR with less than 24% dd, and that's over 5 years. Returns the past 3 years are much higher. My timing rule is based on a rule Oliver posted a couple years back, only even more aggressive at getting back in to the market. Combining with the EPS rule causes it to be out of the market even less. Interested in thoughts about this timing rule. I think it may be too aggressive and too well fit to the past to expect very much from it in the future. Don ---------------------------------------- [Edit 1 times, last edit by dwpeters at Feb 13, 2010 2:09:35 PM] |
||
|
|
strader1
Advanced Member
|
All, A word of caution: Sometimes the market goes down for reasons other than earnings. Examples: 911; the Cuban missile crisis; the aborted Russian coup. With more nations going nuclear the chance of a non-earnings driven sell off are rising. --Bill |
||
|
|
gfagerlin
Advanced Member
|
Bill, On your word of caution about the market going down for reasons other than earnings, this is certainly true. I recently watched as the market went down in reaction to Greek debt and then went down further when Obama took on the banks on bonuses. In the middle of this my benchmarks started to sell me out. I certainly can't predict the next thing that will happen. But, I think I can react to whatever may happen. Depending on how good my reaction is my losses will be lower. The "art" comes into play on the timing of the exit and the re-entry. Will I lower losses but give up overall return because I miss the reversal? If I look at the benchmark in Don Peters sim above, it uses earnings and two types of market movements to trigger the benchmark. If you look closely, his sim uses S&P earnings and Russell 2000 market movements, which I find interesting. In my testing of a Bompus+TF 9/08 sim, Don's benchmark works better than any others I have tried. I am getting +120% AR and -17% drawdown with Don's benchmark. Without a benchmark its +92% AR and -56% drawdown. The -56% digs a pretty big hole to recover from and I doubt I could ride a port through that kind of storm without giving up. In fact, I know so since I have made money on Bompus and I have also lost big on Bompus declines. With benchmarks I am used to giving up AR for lower drawdown. In this ranking and Don's benchmark I am getting higher AR and lower drawdown, which is unusual. I don't think Don's benchmark will work exactly like it has in the next decline since the next one will probably be different. I would bet it would work reasonably well in the future and a lot better than no benchmark. Will it work better than the others I have tested that didn't do as well over the last 5 years? I don't know but I will bet on history, believing it will do OK. Glenn |
||
|
|
gfagerlin
Advanced Member
|
All: Here is a better sim than my starter sim mentioned above. Over 5 years AR is 120% with -17% drawdown. It uses Denny's ranking, Don's benchmark and pretty standard liquidity. Some comments: Q Stocks. I found a lot of bankrupt companies (stocks with Q as the last letter) in the transactions. Using a list of Q stocks, I tested the effect of bankruptcies. Trading the Q stocks actually was better; AR +11% & DD the same. This was a surprise. I used Denny's list from the forum with some updates. The list is attached. Liquidity. The sim is quite flexible to higher liquidity levels. Using AvgDailyTot > $1 mil. and close(0) > 3, the AR is 73% and DD -18. Trading. This sim uses open pricing and slippage of .25%. I tested trading options with excel / xlq and find that the best way to trade this model over the last 5 years would be to buy at the market open and sell with a limit order set at the port recommendation (prior close). The sim would buy at the open and the sell limit order would be filled 58% of the time at the open, 34% intraday (when the high exceeded the limit), with the remaining 7% sold the next day (2% at a high and 6% at the close). These trading rules would result in slippage of .07%, which equates to AR of 150% and DD of -14%. The ranking has top bucket returns of 108% over five years and the sim provides 120% to 150%, with remarkably low drawdown, depending on how you would actually trade a port. Glenn ----------------------------------------
Q Stocks Text File.txt
(1382 bytes)
(Download Count: 33)
---------------------------------------- [Edit 1 times, last edit by gfagerlin at Feb 17, 2010 5:34:17 PM] |
||
|
|
dwpeters
Advanced Member
|
Hi, Is there a need for liquidity rules? Does the PRussell 2000 universe only include stocks that were in the Russell 2000 at the time, or is it as of a certain date? Don |
||
|
|
jerrodmason
Advanced Member UNITED STATES Joined: Jan 14, 2005 Posts: 599 Status: Offline |
Glenn:
Does filtering out the Q's introduce look-ahead bias? What do you think of a feature that would return the last letter of the ticker? ---------------------------------------- The smart money was on Goliath. "He's not the Messiah. He's a very naughty boy." |
|||
|
|
[Show Thread Printable Version] [Post new Thread] |
A Good Small Cap Ranking System
