| Index | Recent Threads | Who's Online | User List | Search |
|
|
![]() |
Portfolio123 » List all forums » Forum: Errors » Thread: Universe in buy rule vs Universe field |
|
Total posts in this thread: 3 |
[Request a Feature]
[Post new Thread] [Add To My Favorites] [Watch this Thread] |
| Author |
|
|
stenci
Advanced Member
|
I'm not sure this is an error, but I can't understand why there is a difference between two simulations that select the same universe in 2 different ways (and don't use the absolute value of the ranking.) These two simulations differ only in the way the universe is selected. One uses universe() in a buy rule and the other selects the universe on the universe field. They don't use the rank value in the buy rules. Test with Universe on Buy Rule Test with Universe on Universe These are the lists of stocks / rankings bought on the first period: KBH 99.0Ranking all the stocks vs ranking only the selected universe explains why the values of the ranking are different, i.e. why VLO is 100.0 in one case and 99.8 in the other case. But it doesn't explain why the 4th highest value is BR in a case and SUN in the other. Whether you discard the stocks that don't belong to the universe before or after ranking them, the resulting order shouldn't change. Why SUN is not the 4th stock on both the simulations? |
||
|
|
marco
|
The ranks and order do change depending on universe. Take for example Growth. A company with a growth rate of 10% might get a very high score when compared with large caps but not when compared to whole universe. You can verify which component of the Ranking System is causing this by entering the tickers in the Ranking->Ranks page ---------------------------------------- Portfolio123 Staff |
||
|
|
Stittsville123
Advanced Member
|
Stenci - I have given this issue some further thought and I think I can answer your question. I just not sure how well I can explain it. During the calculation of a stock ranking, each individual factor is ranked first, then weighted and combined with all of the other factors to create the final ranking. The issue is that the factors are ranked first; if raw data were used for the factor data and the raw data weighted and combined then Stock A would always be ranked higher than Stock b regardless of the universe definition. This is because all of the mathematics is linear. But the factors are ranked first and you have to consider that the ranking function is not a linear mapping to the original data. In this case the universe definition can play a role in the final outcome of the ranking. Steve |
||
|
| [Show Thread Printable Version] [Post new Thread] |