|
PureBytes Links
Trading Reference Links
|
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;
|