Unless I'm not using it properly I'm disappointed with the so called replay feature. All I want to do is step forward manually one bar at a time over historical data in off-line mode to examine how my studies work from one bar to the next. I have a signal study and several indicator studies applied to a 1-minute chart. I also have other charts at other time intervals displayed in other windows in the same worksheet, each running a couple of indicators. I'm also using global variables to send information from all the indicators to the signal study. I open up the Backtesting tab in Properties for the signal study and check the box next to Use Bar Magnifier in the BackTestting Precision. My only option available under that is Tick, which is to be expected since it's a 1-minute chart. However, I get a runtime error message "Not all of the required data has been obtained. Bar Magnifier mode will be disabled." I presume I'm getting this error because I do not have tick data for all of the bars, but I do know I have them for some.

Before I waste any more time on this I like to know if I can the following. I want to step across the history of the 1-minute chart one bar at a time regardless of whether there is any tick data or not. At the same time I want the other charts to step in sync. So, if I have a 2-minute chart, it will only step forward every 2 minutes to be in sync with the 1-minute chart. At the same time all indicators on the 2-minute chart will execute, perhaps only once if there is no tick data, which is fine since I'm only really interested in the indicator values at the end of a complete bar in real-time anyway. I'm also using the RecalcLastBarAfter command in the signal study attached to the 1-minute chart to recalculate after say 10 seconds. So, I like to have that signal study to run again after 10 seconds even though the other studies have executed only perhaps once since there were no tick data for the corresponding bar.

If this is not all possible then I'm afraid the so called replay feature is useless for me, and isn't taking advantage of the core feature of MultiCharts, namely the ability to display multiple charts in a nice and user friendly manner.

I think I found a workaround. All I need to do is insert all the other 6 charts into the 1-minute chart window and make them invisible to avoid the clutter. The downside with this workaround is now I have lots of studies executed twice as I still need to watch the indicators on the other 6 charts. It's a shame I can't link the other chart windows to the one being replayed and have them all replayed together to avoid this workaround; unless there is a way to do this. Is there a way? If there isn't I will submit a suggestion to have this feature added into a future version of MC, unless such a suggestion has been submitted already.

Update:
Just noticed the Global Mode feature. That allowed all windows to be played back together. So I don't need to use my workaround after all.

I remember reading that some day they are going to try and make the replay an actual replay tick by tick (a real and true replay). I guess you already know the replay only replays when barstatus=2. For my purposes this is not that big a deal but I did have to change code to get it to execute with the barstatus=2 restriction (I set a GV to tell the code it is running replay). I suspect a real and true replay could be a fair challenge to program since they will have to record when each tick hit the chart so it replays exactly they way it did during the day. Even if it is not that hard to program it would add some overhead at least. If this is done users should have the option to turn off the logging of tick arrival if they do not use replay. Maybe when the 64 bit is fully established in the user base the true replay would be more appropriate.

Yes bowlesj3 I've noticed the same problem, and using a similar workaround. A true replay feature to re-enact virtually the same conditions as when the data was collected in real-time won't be possible until sub second timestamps are used (ie, milliseconds). For my current proposes this isn't necessary though. I'm happy the way things are despite the problem you mentioned. At least for now.

Yes I guess you are correct about the Milliseconds. I had not thought it out that much assuming the MC programmers would get right down to the nitty gritty. It occurred to me this morning that the current replay should be left intact if they do decide to put in a true replay for two very good reasons 1/ it will probably run faster than the true replay and 2/ some may find it good enough. I personally would probably start using the true replay if it was available rather than the current one since it would be a bit better for me (unless it bogs down too much). I do not use the global function for that very reason (I find it too slow). However when I some day go to 64 bit I suspect it will not be too slow. I have to upgrade both my desktop and notebook before I go 64 bit (I think). So it may not happen too soon.

I think this should all be put in the bug tracker as a feature request including keeping the old version as a comment so they do not remove it. If it has not been done maybe I will get it started some time soon and encourage a lot of input from others somehow. I still feel that if they get the 3rd party vendors who created the bug tracker to include "My votes" lists there would be a lot more info going into the bug tracker for things like this.

Before I forget, another thing that needs to be thought of regarding a true replay is the request for other methods of executing a study such as change in bid or ask. For me I do not need it but I am sure some will use this feature and would like to replay it. Another item to be put in the request so that if it does come to be it is done in a very well thought out manor.

Getting back to the topic of this thread,

There are two requests for tick by tick replay in the bug tracker (74 and 774). I have added my thoughts regarding "true replay" or "tick by tick replay" to this bug tracker request.
http://www.multicharts.com/pm/viewissue ... _no=MC-774