Re[2]: Advance-Decline data

[Jim Johnson <jejohn@xxxxxxxxxxxxxxxx>]

Hello david,

thanks, dave.  I thought it was me.

-- 
Best regards,
 Jim Johnson                           mailto:jejohn@xxxxxxxxxxxxxxxx

Friday, March 22, 2002, 7:50:45 PM, you wrote:

dbs> Hmmm. Gene and Chris must be talking apples and oranges or I'm just
dbs> confused.

dbs> Both of you using natural hour bars? Index or futures for data1? How far back
dbs> do the tests go?

dbs> Gene Pope wrote:

>> Ahhh... back from some warm weather... time to crack the knuckles...
>>
>> At any rate... I tested these variations a while back, both raw + and -
>> values, and ROC models using UpVol and DnVol... they tested consistently
>> worse... as if their info was "lagged" somehow. Note that I did not mess
>> with the RL, which might affect it.
>>
>> Your milage may vary.
>>
>> Best regards,
>>
>> Gene Pope
>>
>> ----- Original Message -----
>> From: "Chris Cheatham" <nchrisc@xxxxxxxxxx>
>> To: "Jim Johnson" <jejohn@xxxxxxxxxxxxxxxx>; <omega-list@xxxxxxxxxx>
>> Sent: Thursday, March 21, 2002 5:44 PM
>> Subject: Re: Advance-Decline data
>>
>> > Jim,
>> >
>> > I didn't read the article, but I have found that intraday uvol-dvol is
>> > infinitely more valuable than adv-decl. This is true on NYSE, but esp.
>> true
>> > on Nasdaq, where few stocks account for much of the volume and price
>> > movement.
>> >
>> > Chris
>> >
>> >
>> > ----- Original Message -----
>> > From: "Jim Johnson" <jejohn@xxxxxxxxxxxxxxxx>
>> > To: <omega-list@xxxxxxxxxx>
>> > Sent: Thursday, March 21, 2002 7:27 AM
>> > Subject: Advance-Decline data
>> >
>> >
>> > > Hello omega-list,
>> > >
>> > >   I read McMillan's article in TASC last night about the junk that's
>> > >   included in the Advancing and Delcining issues data stream.  It
>> > >   seems a fairly compelling case (many issues are closed end bond
>> > >   funds that may change price by only 1 cent based on interest rate
>> > >   changes).  In addition, the 1 cent trade increment has cheapened the
>> > >   definition of an advance or a decline.  He does suggest that because
>> > >   these "bastard" issues (my term) have low volume, that advancing/
>> > >   declining volume may be a better alternative.
>> > >
>> > >   As they say, this issue seems non-trivial.
>> > >
>> > >   I appears that Neoticker's NeoBreadth product may attempt to create
>> > >   custom A-D indices.  Is that what it does, has anyone tried it?
>> > >   other thoughts on this data cancer problem amongst us?
>> > >
>> > > --
>> > > Best regards,
>> > >  Jim Johnson                         mailto:jejohn@xxxxxxxxxxxxxxxx
>> > >
>> >