Ettibebs,
Since you acknowledged that simulation can not replace field time I will let that one go and try to answer your question based on my experience. The key to employing simulation effectively is to clearly define what the training aims are and where it fits into the training plan. Usually this means earlier in a cycle (either individual or collective) to work out the bugs before going to the field and wasting time learning lessons that could have been learned in a more sterile environment before adding the stressors that will naturally be injected during field training such as veh breakdowns, 24/4 ops, sleep deprivation etc., etc. That will come.
IMO one of the problems with simulation in the CF is a lack of resources (money) require us to get one system that does everything okay and nothing really well. JCATS for example was a good tool for allowing leadership at various levels (ex dependant) to practice conventional war-fighting ops i.e. emphasizing reaction to a fluid environment, decision making, passage of orders, use of enablers etc. Did it take into account every variable that one could encounter? Absolutely not but used in the earlier stages of a training cycle it allows teams to work out SOPs or it gives individual students an idea of what to expect thus making them more functional once they get back to their units where the real training begins. What JCATS did not do well is the asymmetrical piece but we try it anyway because JCATS is the tool we have. You can not hold a shura in JCATS (although its been tried), search vehicles or compounds, identify civilian or insurgent etc. But forcing the tool on the situation just because it is the only one we have adds no training value whatsoever.
Another example was the Indirect Fire Trainer (IFT) at the Artillery School. There were two and a few years ago, a movie set designer was hired to come in and convert one to a simulated [destroyed] building and a dismounted OP. This did nothing but decrease the utility of the room as it reduced the number of students that could be easily accommodated and it became a very awkward place to give actual lectures/classes/presentations in. IMO this is even more apparent when we keep in mind that the intent of the IFT was a "procedural trainer" not a simulator. Students need to first understand the basics of how to conduct a fire mission properly. The "doing it while uncomfortable - cold, wet, lying down after humping for 12 hours" comes later in the course. What it did though, was make the school look good because they were on the "cutting edge".
One of my least favourite "simulators" although not digital was the 14.5mm trainer for Artillery. Drills at every level had to be altered somewhat to make it work, so few people other that the person "firing" a fire plan got any real training and even that could have been better achieved in a well run CPX. So in the end, it arguably taught bad lessons and took away real bullets from the units. But I'm sure there is enough debate there to start its own thread.
To sum up, I guess my point is that if you have any input into the types of simulators we develop, make sure you figure out why we need it (and make sure we do) then be identify exactly what you need it to do and design it to do (just) that well. Don't look for one that promises to be an all-singing and all-dancing.