Nah, I think it's +12. So long as we are just posting numbers with no context, I think my is better because it spans a larger signal space than yours; bigger is better, no?
Well, let's see what the market says: SPX 8/29/2011 1:35pm $1,203.51 +26.71(+2.27%) NDX 8/29/2011 1:36pm $2,212.09 +50.12(+2.32%) GOOG 8/29/2011 1:20pm $537.35 +10.4899(+1.99%) (delayed quote) So GOOG is performing about par with both the SPX and the NDX. Heck of a stock that is +12 that can't even beat the averages percent wise on a day when huge sums of money are being distributed to the stock market. My model appears to be, for now, closer to reality. Let's see what happens on a big down day.
Who said +12 means long?? My signals are in cross sectional residual space - so +12 means it's richer than all the other names in the universe, and therefore it's a short candidate.
I think you need to read the first post carefully. There it is explained that these numbers have exact context. How they are derived may not be, but there is no uncertainty as to what a ranking means.
I did read it. I (think) I know your general methodology here. The problem is, you can't backtest it. If you can't backtest this - and this is a quantitative method, then it's somewhat pointless for you to post these since there's no information as to what each signal means (in terms of expected performance). You don't know them because you can't run the analysis. Moreover, if you were actually serious about looking into doing this, you would at least keep your universe constant (rather than running a bunch of different groups) and show the daily PnL. Obviously, you are not obligated to do any of the above; But why post it if not then? There's no meaning in posting a bunch of rankings.