I don't quite understand what you mean by 'iterates through a bar'. Given that there is only a single set of values (O/H/L/C) there is nothing to iterate through.
However, it is valid to have conditional logic in the assignment, yes. Though, your syntax usage is incorrect. See below for correction. PositonSize = iif(L<limitprice2, -20, -10); BuyPrice = iif(L<limitprice2, (limitprice1+limitprice2)/2,limitprice1); Mike --- In [email protected], "jmdeacon" <jmdea...@...> wrote: > > Can the position size vary as the program iterates through a bar? For > instance can I write > > iif(L<limitprice2, positionsize = -20, positionsize = -10); > iif(L<limitprice2, buyprice = (limitprice1+limitprice2)/2,limitprice1); > > John > > --- In [email protected], "Mike" <sfclimbers@> wrote: > > > > If you want to do what you are proposing, you don't need CBT. > > > > Whenever you recognize a "scalein" condition, just set the BuyPrice to the > > average of buyprice1 and buyprice2, and the position size to double the > > usual size. The backtester exits all positions (initial and scaled in) at a > > single Sell signal, so that is effectively what is happening anyway. > > > > Mike > > > > --- In [email protected], "jmdeacon" <jmdeacon@> wrote: > > > > > > What I want would not need intraday data. The way I would want ti code > > > with OHLC is something like this: > > > > > > limitprice1 = (ref(c,-1)*0.99); > > > limitprice2 = (ref(c,-1)*0.98); > > > > > > buyprice1 = min(O,limitprice1); > > > buyprice2 = min(O,limitprice2); > > > > > > Buy would occur if L < limitprice1. > > > Scalein would occur if L <limitprice2. > > > > > > If the low is less than the limitprice1, I would enter the position. If > > > the low is less than the limitprice2, then I would scalein. If the open > > > is below both prices then I would enter a double position at the price of > > > the open. > > > > > > I already have this coded to work on seperate bars, but many times, the > > > pullback occurs on the same bar and I would like to backtest that > > > occuring on the day of the initial purchase. > > > > > > So, the question is, is there some kind of CBT coding that would allow me > > > to test what I want to test. > > > > > > Thanks, > > > > > > John > > > --- In [email protected], "Mike" <sfclimbers@> wrote: > > > > > > > > Hi > > > > > > > > You can't do that. All you know is open/high/low/close. You don't know > > > > what happened in between. > > > > > > > > Consider these two scenarios: > > > > > > > > Open, gentle steady decline in price down to -2%. > > > > Open, next trade after the open is a 2% drop. > > > > > > > > Now, would you still want both an initial buy and a scale in for both > > > > scenarios? How would you know if half the position was bought at -1% > > > > and the other half at -2% (first scenario) vs. all bought at -2% > > > > (second scenario). > > > > > > > > If you want intraday trades, you need intraday data. > > > > > > > > Mike > > > > > > > > --- In [email protected], "jmdeacon" <jmdeacon@> wrote: > > > > > > > > > > Is it possible to scale in on the same bar? To make it clear, lets > > > > > say I am using EOD data and I want enter a position as the price > > > > > pulls back. XYZ starts the day at $10 and I want to buy if it loses > > > > > 1% and scalein to double to position if it loses another 1%. Is > > > > > there a way to code for that to occur on the same bar assuming XYZ > > > > > traded below $9.80 on the bar in question? I know how to do it with > > > > > Scalein when it occurs on seperate bars? > > > > > > > > > > Thanks, > > > > > > > > > > John > > > > > > > > > > > > > > >
