Contents of AB-FT afl03 zip

8-10-2006

Proprox.zip       4/4/05    Bill Barnard-As part of my work in AmiBroker I have made least-squares, best-fit proxies for the 12 Profunds we use most (1x, -1x, 2x, -2x for RUT-I, SP-CP, and NDX-X). While these are a better fit to the actual Profunds, the differences from earlier proxies is minimal and probably inconsequential in our work. The main point was to make a complete set, covering the three main indexes of interest.

The attached zip file contains 12 FNUs, which should go in your FT directory, and which should be retained for use each time the proxies are updated. There is also an ini file to be run in Trade to make the current proxies.

This is not necessary for anyone who is presently using an ini file to make proxies, unless they need all twelve of them (as would be the case if one were studying hedging with SP-CP or NDX-X index funds).

Ruttr-Ft bis.afl          4/27/05   Bert Steele- IO Ruttrft.afl example:

Attached is my Ruttrft.afl. I used it to run 5 anchored IO runs. For each, Fitness was Car * 0.09 ^ Trades / Mdd, Current Ticker at DFSCX, Long only with Delay =1. For each, the IS period started on 9/1/1988 thru 3/31/2000. OSS ran to present. I wanted to see what IO provided for April 2000 to present and more particularly from April 2000 to the middle of March 2003. Here are the results from 4/1/2000 to present for the 5 separate runs. Best risk adjusted return first.

IOEQ5: 12.94 CAR and 12.18 Mdd with Every, 6Months.

IOEQ3: 12.17 CAR and 11.98 Mdd with Every, 3Months.

IOEQ2: 12.57 CAR and 17.04 Mdd with Long, Exit.

IOEQ4: 8.06 CAR and 25.85 Mdd with Every, 3 Months.

IOEQ1: 3.70 CAR and 24.17 Mdd with Long, Exit

 Sirken1.afl & Sirken1b.afl   4/28/05  Tom Ttsao-Here are a couple of AB translations of one of Josef P's recent contributions. I *think* I understand the flip and exrem statements, but I don't guarantee it's done correctly.

Also, does anybody know how to do the Trade equivalent of HoldMinDays in AB? Does it have something to do with the "n-bar stop"?

RuttrFt3.afl          5/3/05   Tom Ttsao-I'm slowly working on various translations, starting with yours and Josef P's. Anything in particular you'd like to see?

Fred posted a version of ruttrft a while back. Below is a copy of one version.

Fuseme.afl  & Utilities7.afl  5/4/05  Bert Steele-I only have a few ini's translated to afl. I have a hard time translating Trade's "or" and "and" signals. Here is one you will recognize. The ruttrft one that was just posted is another.

It used Bruce Robinson's Utilities6.afl to prepare fnus and trade signals. I have attached utilities7.afl which is the update. Of course, you can make sure that the lines that use utilities6.afl are either changed to utilities7.afl or comment them out with //

Pat-Fubar5.afl   5/4/05   Bert Steele-And, here is a version of Fubar5 that one of the guy's in the Houston investment group posted to their Web page. This is Pat Finnegan's work.
 

AMI NDXsi1.afl       5/4/05  Bert Steele-And (maybe) finally Dave, here is a version of NDXRSI1 that Bruce posted some time ago.

Noncor1b.afl        5/4/05   Tom Ttsao-I'm not entirely happy with it. The signal dates track Trade's signal dates from 8/5/99 to present, but there are some discrepancies before then. I think it will take some time for me to figure out why, so I thought I'd put it out there for now...

I *think* the way I have it set-up, AND's should be done in state form, while OR's should be done in impulse form... but I haven't totally convinced myself of that... (I have some really weird looking drawings on my pad of paper right now)

BTW, it also depends on how you write the condition. For example, z = x > y gives you something in the state form, while z = cross (x,y) gives you something in impulse form... both are fine as long as you do some exrem/flip at the appropriate time

Ts22a.afl        5/8/05       David Serbin-  posted and modified by Tom Ttsao 5/9/05 -  

Take a look at the attached. It's the same thing you posted above, but I commmented/uncommented a few lines, just to show you how I do one form of debugging.

Essentially what I do is that in each section, I have two lines that read as follows:

test_b = something;
test_s = something_else;

then, at the end of the afl, I have this:

Filter = test_b or test_s;
AddColumn(test_b, "Buy", 1.0, colorDefault, IIf(test_b, colorBrightGreen, colorDefault));
AddColumn(test_s,"Sell", 1.0, colorDefault, IIf(test_s, colorRed, colorDefault));

Then, when I click "explore" in the AA window, I'll get all the Buy/Sell signals in green/red, and then I compare w/ the Trade signals.

In the attached file, I have it set up to print out the pfgcom_b and pfgcom_s signals, which correspond well with your ts22.ini (3Xpf11) generated signal dates.

After that, you can comment/un-comment the appropriate lines in the afl to test out each section. I haven't gone through all your subsections, but if I had to guess based on my experience, I'd say that something funny may be happening when you combine all the signals.

BTW, I took a closer look at Bruce's ft-utilities7, and I see that there is a FTPutSig function (obviously, I haven't used it yet) which will export the signals to a .sig file