There are analysts that appear to be quite good. I don't know how much of it is luck or not. Still, worth following:
http://www.cnbc.com/id/44846927/
http://www.cnbc.com/id/44846927/
Stock Model
- Thread starter nitro
- Start date
Be weary of bank earnings:
"...Some of the damage is obscured by the fact that banks get to use an accounting gimmick that automatically boosts their bottom line in difficult times.
When times are tough for banks, the value of the bonds they've issued to fund themselves falls. Banks are able to count this fall in the value of the bonds as revenue, since they could theoretically buy the bonds back and reduce their overall indebtedness...."
http://www.cnbc.com/id/44929076
"...Some of the damage is obscured by the fact that banks get to use an accounting gimmick that automatically boosts their bottom line in difficult times.
When times are tough for banks, the value of the bonds they've issued to fund themselves falls. Banks are able to count this fall in the value of the bonds as revenue, since they could theoretically buy the bonds back and reduce their overall indebtedness...."
http://www.cnbc.com/id/44929076
Model is exactly even. It does a decent job of defending itself against higher overall markets, and makes good money on selloffs. That is perfect given that intermediate term, I am macro bear (although short term (week or so) I am macro bull).
It is worth noting that the model trades on longer time frames, not because it is built into the model itself, but in the way I structure the portfolio to lean one way or the other, based on FV of SPX.
It is worth noting that the model trades on longer time frames, not because it is built into the model itself, but in the way I structure the portfolio to lean one way or the other, based on FV of SPX.
What does this mean? Does your model trade itself? I thought it just generated ratings for individual stocks.
A few more questions: Is your model time-frame daily, weekly, intraday or other? How many samples do you need to calculate a rating (e.g., 5 bars; 10 bars; 55 bars). You said something about "congruence with the VIX" but the VIX is meant to be a prediction over 30 days. Is your model output a prediction or a metric? Thanks.
Sorry, I am confusing people. You are correct, the model just gives rankings. I take those rankings and made a "book" of relative value of rankings. That currently is not part of the model because it is not quantitative, but it will be soon when I program an efficient frontier etc. Model being even just means mark-to-market. It doesn't mean that I exited the positions. I guess I should post the book from time to time, otherwise it gets too confusing.
The model is none of those. The moves can play out on any of those time frames. But if you asked me what the information it uses, it is basically using daily to weekly information, so it naturally plays out on an intermediate time frame.
The model mixes many things that an investor/swing trader might consider. "Bars" is a very small universe. The congruence with VIX just means that I have observed greater mispricings on a high VIX, especially rallies from a high to a lower VIX. It is not part of the model, but the way that I choose what to go long/short.
The most important input to me choosing how to lean, other than the stocks themselves, is my macro view, based on FV and this thread (I am just linking to the end of the thread at this moment)
http://www.elitetrader.com/vb/showthread.php?s=&postid=3332294#post3332294
If FV is below SPX I lean short, and vice versa.