I’m trying to figure out what Rankpos does.
If I have a sell rule like this rankpos > 75, what happens? https://www.portfolio123.com/port summary.jsp?portid=163775
And a purchase rule like this: RankPos > 6 https://www.portfolio123.com/port summary.jsp?portid=1633166
The sell rule is to sell stocks that are not in the top 75?
The reason I’m asking is that when I use this instead of just rank x, I get a lot more turnover, but also much better performance.
And is there a good reason to use this instead og just rank > x?
I’m not sure if this addresses your question, but:
Rankpos would be invariant to the size of your universe. If the universe is 1000 stocks then selling Ranpos>75 would reflect selling stocks when they fall out of the top 7.5% of universe. If the universe is 2000 then that same sell rule would reflect selling stocks when they fall out of the top 3.75%. The universe size is likely changing over time, so it’s different than using rank which just selects the top x% of companies regardless of size of universe.
I’d think RankPos might be valuable in cases where you have an exact # of stocks you want to have in the portfolio rather than a percentile range which could shrink or grow over time depending on the universe. Could make allocation/rebalance actions more straightforward also.
RankPos is just the ordinal rank. If you rank 10 stocks, the best will be ranked 1 (first place) and the worst will be ranked 10 (last place).
When to use Rank v RankPos? In part, I think that it’s personal preference. Last I saw there were about 7000 stocks in the Easy to Trade US universe, so using RankPos<700 is going to be pretty similar to Rank>90.
But the biggest reason is that in testing the universe size can change over time. If you use RankPos then you always want the top 700, even if the universe is now 100,000 stocks or 1000 stocks. Rank will remain constant as a percentage of the universe.