[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re[2]: BondBall beats OddBall in year 2002



PureBytes Links

Trading Reference Links

Hello Heiko,

right.  not to be shill for Ruggerio but Murray also has some code,
don't know if its inthe book, that uses on on-going corellation of
variables as a signal.  In his case I recall he uses the cor betw open
interest and volume as an indicator of commercial behavior.

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

Wednesday, March 13, 2002, 9:28:25 AM, you wrote:

>> I believe it is Mark Cook who describes Bonds and equities as cycling
>> thru phases--dating, married, divroced.  That may explain the 1999

HS> For this you can use correlation.
>>0.X they married
HS> <-0.X they divorced

HS> Then you now in with direction you can use the TBond or better not.

HS> Heiko


HS> ----- Original Message ----- 
HS> From: "Jim Johnson" <jejohn@xxxxxxxxxxxxxxxx>
HS> To: "Beyond OddBall" <beyond_oddball@xxxxxxxxx>
HS> Cc: <omega-list@xxxxxxxxxx>
HS> Sent: Wednesday, March 13, 2002 12:38 AM
HS> Subject: Re: BondBall beats OddBall in year 2002


>> Hello Beyond,
>> 
>> I believe it is Mark Cook who describes Bonds and equities as cycling
>> thru phases--dating, married, divroced.  That may explain the 1999
>> findings for example.  this type of system may be a great candidate
>> for equity curve filtering.  Murray Ruggerio presents some code for
>> this in his book, Cybernetic Trading Strategies, as I recall.
>> 
>> -- 
>> Best regards,
>>  Jim Johnson                           mailto:jejohn@xxxxxxxxxxxxxxxx
>> 
>> Tuesday, March 12, 2002, 2:22:27 PM, you wrote:
>> 
>> 
>> BO> I'm getting ~$60,000 TNP, with a profit factor of 3.55
>> BO> and ~$8,000 MIDD.
>> 
>> BO> Constructing OddBall systems with data offsets running
>> BO> ADV issues earlier than S&P is cheating yourself. But
>> BO> US Bond futures DO open over an hour earlier, so why
>> BO> not explore any possible advantage. Even besides
>> BO> 8:30am news events, Bond traders are known to be some
>> BO> of the first to react.
>> 
>> BO> Too bad it does not perform as well under long term
>> BO> historical testing (needs tweaking ?). But maybe there
>> BO> is some use for the concept - as a filter - or as an
>> BO> addition to . . .
>> 
>> 
>> BO> This code is roughly opposite of OddBall, but with
>> BO> lookback of only 2 bars. This strategy failed during
>> BO> most all of 1999 (when running a normal oddball code
>> BO> on bonds made money), I would still hope a strategy
>> BO> could be devised to characterize the changing cycles
>> BO> of the bond futures so when the bonds started moving
>> BO> in synch with the S&P's you could stop trading this
>> BO> code and monitor it until it returns decent
>> BO> profitability.
>> 
>> BO> {Data1 = S&P 500 futures.
>> BO>   Data2 = CBOT 30 yr. Bond Futures
>> BO> Always natural hour bars}
>> 
>> If time >> 1100 and time <= 1600 then begin
>> 
>> BO> Inputs: RL(2), BZ(0), SZ(0);
>> 
>> BO> If RateOfChange(Close Data2, RL) < BZ Then Buy;
>> BO> If RateOfChange(Close Data2, RL) > SZ Then Sell;
>> 
>> BO> End;
>>