| Index | Recent Threads | Who's Online | User List | Search |
|
|
 |
Portfolio123 » List all forums » Forum: Ranking Systems and Factors/Functions » Thread: TOP Bucket |
|
Total posts in this thread: 4 |
[Request a Feature]
[Post new Thread] [Add To My Favorites] [Watch this Thread] |
| Author |
|
|
hunter
Advanced Member  UNITED STATES Joined: Apr 23, 2010 Posts: 21 Status: Offline |
So ive got a pretty good ranking system, and I'm happy with the results. Backtesting 9 years, the top bucket (100 buckets) has outperformed the bottom bucket 84% of the time. Since I'm following a market neutral strategy of going long top bucket and going short bottom bucket I'm happy with these results. My question to others on this forum is about the movement within buckets. This morning I look at the top 30 stocks (top bucket), some of them were in the bucket last week, and some are new. Has anyone ever taken their research to the next level and looked at the performance of the top bucket and found out if the newcomers perform better, or the consist top bucket stocks perform better? Is there a way to write this code into a simulation? Does bucket direction even matter? Am I asking the question clear enough? Thanks ALL!! ---------------------------------------- Emotionless Investing |
||
|
|
gfagerlin
Advanced Member
|
You might try the RankPrev function, if I understand what you are trying to test. RankPrev(weeksAgo): Historical weekly rank based on the selected ranking system. Weekly ranks are updated every Saturday. Running a sim with the buy rule RankPrev(1) < 90 AND Rank > 90 should be the newcomers. RankPrev(1) > 90 AND Rank > 90 should be the old timers. Comparing the sims should give some insights. |
||
|
|
DennyHalwes
Advanced Member UNITED STATES Joined: Apr 28, 2004 Posts: 1530 Status: Offline |
Hunter, Glenn idea is a good one. If you want to test the top 1% bucket use; RankPrev(1) < 99 AND Rank > 99, for the new comers, and use; RankPrev(1) > 99 AND Rank > 99 for the old timers To see the performance in the next bucket use; RankPrev(1) < 98 AND Rank > 98 AND Rank < 99, for the new comers, and use; RankPrev(1) > 98 AND Rank > 98 AND Rank < 99 for the old timers. I am always surprised how much difference a 1% difference in rank makes for the best ranking systems. You also might want to test the top ½% bucket using 200 bucket and Rank > 99.5. The best way to set up the comparison in a Sim is to set transaction cost and slippage to 0, allow sold stocks to be re-bought, use only the above buy rule and an approperate liquidity rule, and use a single sell rule; Rank < 101. This will sell all stocks each week, and re-buy all that meet the buy rule. So you know that only the stocks that meet the buy rule will be held in the Sim as long as they still meet the rule. After you decide which rank values you want to include in your sim you can combine the buy rule with you other buy rules. You might need to add a sell rule that is the inverse of the buy rule you choose. Denny  ---------------------------------------- "The significant problems we face cannot be solved at the same level of thinking that we were at when we created them". Albert Einstein |
||
|
|
dcnelson
Advanced Member
|
Excellent. I just added that rule to one of my existing sims and over the last 5 years it doubled the outperformance to the SPX. Over the last 1 year it it took the outperformance on a 20 stock portfolio from 5 to 15. If i created a new portfolio today it would only buy half the names. |
||
|
| [Show Thread Printable Version] [Post new Thread] |
Re: TOP Bucket
