sounds like you added a fleet training mechanic (probably because it occurred to you or you've seen it in another 4x, and it sounded cool - a natural reaction), but have yet to determine exactly what its supposed to be modeling. and now you have to come up with something believable, to explain the existence of the game mechanic and why it works the way it does.
pretty much the reverse order of how i design things.
no offense, but sometimes its makes me laugh or at least crack a smile, the way you seem to keep doing this.
ok, so, just what should it be modeling? well, knowledge of mill ops in general will help there.
so it produces a bonus vs certain opponents, but is not based on some other game mechanic such as weapons upgrades, personnel assignments, etc.
a few things come to mind:
1. tailoring tactics/strategy/operations to a specific opponent type
2. code breaking operations vs a specific opponent type
3. (automatic) espionage and/or R&D efforts which translate to the desired bonuses. tweaking sensors, disrupting enemy "automatic R&D bonuses" with espionage (you don't actually have to model enemy R&D bonuses vs two other factions, unless you want to get into that depth).
as for the terms, "maneuvers", or "training exercise", or "running a drill" are commonly used. these would apply to fleet actions that improve tactics , strategy, and operational methods vs a specific foe. its would take time, cost fuel and ammo, and a handful of casualties from accidents and such.
other things that come to mind include: sensor | guidance system | ECM + re-calibration | adjustment | refit
or do it this way:
why should a fleet get a bonus vs two specific factions?
training is the obvious answer. just chalk it all up to continual training programs that concentrate on tactics, strategy, and operational procedures vs specific foes.
in the real world, the USA vs Russia would be a different sort of war than USA vs China, which would be different from USA vs Syria, which would be different from USA vs N Korea. each would require different strategies, tactics, and operational procedures.
just imagine fighting in Finland vs fighting in Saudi Arabia, VERY different battlefield conditions, each providing unique challenges in training, equipment requirements, and logistics operations.
next time you add a rule to the game, you may want to ask yourself why first. that way you don't have to come up with the "why?" later to explain the rule.
and needless to say, if "why" doesn't make sense, the rule shouldn't go in the game at all.
doesn't apply in this case, targeted training makes perfect sense, just make sure the mods are reasonable (5% - 10% max), and not too Monty Haul.
i'm discovering that when you build a game where a lot of stuff mods things (provides bonuses), its best to make the individual bonuses very conservative, perhaps even slightly under powered. keeps things in balance. things have an effect, but nothing affects anything too much. and percentage based vs hard coded bonus points seems to work better for bonuses. that way you can change the underlying points, but the bonuses remain balanced.
i'm dealing with the same thing in Caveman 3.0 now. things like how much skills, tool type, tool and part quality, stats, and god relations should affect the time required and chance of success for actions like gathering resources, researching skills, and making things.
of course, specialized training could be a double edged sword. history shows again and again that such training in peace time is typically geared towards winning the last war, not the next one. as a result, such "training" has often been a penalty rather than a bonus, once hostilities commence. in other cases, the training may have been for a totally different kind of war, again causing a penalty until tactics and methodologies are changed to suit the new battlefield environment - the USA in Vietnam and Beirut and Iran (the failed hostage rescue attempt) come to mind.
