Well, they are accurate, but at range, and with homogeneous types of ships. The unit statistics literally record the results of an encounter between equivalent numbers of those ships when they start at a distance and rush at each other. In those cases, the frigates slaughter the electric shuttles before the shuttles even get near them.
Where the uncertainty comes in is with mixes of ships. If there is something that the frigates are even better at killing than the electric shuttles, then they'll tend to attack that first, giving the shuttles more time to get in range and fire off shots. Or, if the frigates come our of a wormhole right on top of some electric shuttles, that can also be really bad news.
So the statistics are "right," but only in a limited sense -- that's all they can ever be, by nature of the fact that they are just a single percent and time number. (For those taking notes, this is the same sort of over-simplification that led to the real estate market crashing the US). In other words, those single numbers are a rough guide, and tell you the truth under a very specific set of circumstances. As a guide, if you know what the circumstances are (now you do), they are very useful for helping you predict the outcome of engagements, but they are not foolproof. Tactics matter, as does position, as does ship mix. So that's where on-the-fly judgement comes in, and sometimes you have to learn the hard way from failing miserably with one mix (my AAR from last night demonstrates me losing badly against my own AI, because of underestimating the potency of a specific ship mix in the hands of the AI).
I guess it boils down to the fact that you can't boil something so complex down to one or two numbers like this. But, taken for what they are, these one or two numbers are very useful -- they just aren't the end of where you should be looking.
Some general notes:
-With frigates or other long-range ships, you can assume that they will perform much worse at closer range, because their enemies will actually be hitting them.
- For very fast ships, using "kiting" and other similar tactics can make them perform better than their stats would indicate by them staying out of range.
- For ships that do area damage, obviously they are more effective against groups of enemy ships. But, for ships that do a small amount of area damage over a very large area (like electric shuttles), stacking a lot of them in one area will make them much more effective than small batches, since they might well kill all of their foes in one go.
- For ships with explicit stacking abilities, such as autocannons, those can become more powerful than their stats would indicate by grouping the max of them together.
- For ships with stacking penalties (electric bombers and sentinel frigates), those have an ideal number per planet (two or three, usually), and then degrade -- to a point. Past that point, they don't get any more penalty and so they start becoming collectively more powerful again. Often the AI uses them past that point, whereas it is better for humans to keep them below the first degration point.
- For ships with multi-shot abilities, they are certainly better against groups of enemy ships. But, pair those with munitions boosters and they can really do some damage because they are doing more damage to a group of targets, rather than way-overkilling their single target like many other boosted ships will tend to do (of course, boosting helps any ships against hardened targets with a lot of health).
In conclusion
I'm sure there are some other general rules of thumb, as well, but they all boil down to logic and thinking about the mechics of any given ship. Given the number of ships and the number of possible interrelationships, you won't think of everything in advance -- that's expected, and happens to me as well. The main thing to remember here is that if you come up against an unexpected result of an encounter, there is a reason and you can figure out that reason just by looking at the board to see what might be going on in this specific scenario. Then you can assess, decide how to respond, and appropriately handle that situation with your newfound knowledge about that specific setup. That's a big part of the game, honestly, even for expert players -- I'm the designer and programmer, and I still have to do that, because all of those complex interrelationships can lead to some unexpected fruits.
This was a good discussion -- I'm moving it to strategy discussion to make sure others can peruse it if they are inclined.
