Bar Length versus Stops

[Phil Lane <logical@xxxxxxxxxxxxx>]

I came up with this "amazing" system using hourly bars.. But one of the
things that made it work so well was the use of tight breakeven and
trailing stops, so a lot of trades were closed in the same bars as the
entries occured.

Too bad that feeling of "eureka" was wiped out when I tried to see what
would have really happened during those in-and-out hourly bars. I re-wrote
the system to use 5-minute charts, and to "synthesize" hour bars for entry
signals by taking the highest(h,12) etc., and using a modulo function to
look at it once an hour. Here's the kind of thing I tried:

{make 60 minute bars out of 5 minute bars)
input: dayratio(12);   {ratio between 60 and 5 minute bars}
if Mod(TimeToMinutes(time)-30,60)=0
   then begin
   barhigh=highest(h,dayratio);
   barlow=lowest(l,dayratio);
   barclose=c[0];
    
   if (whatever) then buystop=highest(h,3*dayratio)
   end;

On 5 minute charts the entries (on stops) seemed to happen at about the
same points, but the tight stops that had worked so well on the hourly bar
were a complete disaster. The entries would tend to get kicked back out
immediately, and then the system would re-enter on the same or the
following bar..

The question is a general one: Is this what we'd expect? And then are the
built-in stops totally misleading when a majority of the exits (breakeven
or trailing) occur on the same bar as the entries?

Thanks for any thoughts...