I haven’t served in the armed forces for 35 years now, and when I did it was on a strictly amateur and part-time basis. Mind, a lot of the NCOs in my TA battalion were wargamers, and our RSM was the principal author of the WRG’s “Infantry Action 1925-1975” rules (available at http://www.wrg.me.uk/WRG.net/History/OLDWRG/InfantryAction.pdf and still, for my money, one of the best and most carefully-considered sets of rules at this tactical level). The TA background perhaps explains why, as early as 1972, the rules included night fighting, trip flares, and smoke grenades, and distinguished between belt-fed and mag-fed LMGs. At that, one might perhaps detect a strain of sergeant-majorly disapproval of excessive randomness, as the intro to the rules says “On first sight, you may think we make undue use of dice. This is forced on us because we are dealing with individual figures. Take heart, the dice do not decide how many get hit, only who they are.” On the other hand I do vaguely seem to remember that in the 1970s there was a bit of a fashion for variance reduction in wargames rules, on the grounds that skill should count for more than chance.
Possibly more relevant is my experience of umpiring doctrine development wargames at RMA Sandhurst in recent years. Here, the brief was to reduce chance factors as far as possible. The reason for this was nothing to do with any puritancal dislike of gambling — anyone who has seen “Monopoly” played for blood in the RA mess at Larkhill will know that there’s such thing, and from my time writing fleet operational availability simulations for helicopter squadrons I understand that RAF maintainers are inordinately fond of conract bridge. Nor was it even to do with the preference for “games of skill” over “games of chance”. The reason was that the game was not being played for the sake of the game, it was being played for the sake of the insights the game produced into questions of future force suitability against different enemies in different environments. Insights on future force development should not depend on chance.
We had been warned beforehand that there might be a negative reaction if polygonal dice were produced, as a snarky remark along the lines of “Oh, I didn’t know we were playing dungeons and dragons” might stop people taking the exercise seriously. However I did see one occasion where it was generally agreed that the results from a certain situation would have to be determined randomly, 10-sided dice were produced, and the game proceeded smoothly — but then we were dealing with a bunch of very bright officers, mostly young lieutentant-colonels, who had a strong professional interest in the exercise. One might not get the same reaction in a training game for company officers.
The objectives of a training game will probably be very different from a doctrine development game, but it is still probably more important to draw the right training lessons than to have a fun game. One of the training lessons you might want is “some days you can do everything right and still get killed”, but more likely you will want to teach people how to make tactical plans that are robust under uncertainty, so a healthy measure of randomness might be useful here, and indeed it was incuded in military training games such as SPI’s “Firefight”, the Canadian Army’s “Contact”, or the US Army’s Dunn-Kempf game.
All the best,