Cash, using your example, well said, of the prices all occurring within one bar of 1min, the suggestion is that an execution would catch those desired prices you mentioned, hence your back test would validate your tesing parameters. might I suggest that the OHLC of the 1min bar might fail those testing parameters as a valid conclusion irrespective of the myriad of prices that occurred within that 1min bar. for example, if those testing parameters issued a valid or true, then would you accept that back test and go forward with it and real money at risk, or would you continue to test and average the results over a series of tests as validity? consider, the often touted claims of these "trade caller services", they find a way to back test and pass their trade parameters whether or not one could actually get executed at those prevailing conditions. If you're still with me on this, then their proof statements/process would be overrated and false, hence not worthy of risking money safely. short story long, test wisely and conservatively, which will allow for a greater probability of success...
So, if I have a narrow bar breakout system setup and the next bar breaks out on both sides (long and short = outside bar), can neoticker determine which came first - the upside breakout or the downside breakout? pretzel
I'd never trade a system based on just one test. It would be multiple ones across multiple symbols, but the scenario I mentioned is not an uncommon one, so it would occur on multiple tests.
YES. tick by tick testing is slow. Indicators, systems, formulas etc... are recalculated on every tick.
Check out Investor RT: http://www.linnsoft.com Although I don't backtest tick-by-tick, I believe that is one of IRT's capabilities. Give them a call and they will confirm.