One thing we've learned from automatic skull/shield tracking in online play is that it is only a fairly crude measure of luck in the game.
Consider: three 3v3 attacks against squad figures. I roll 2/3 skulls, 2/3 skulls, and 0/3 skulls. You roll 1/3 shields, 1/3 shields, and 3/3 shields. The stats say I rolled slightly below-average and you rolled significantly above-average. However, the reality is that I got an above-mean number of kills. I was the lucky one, not you.
Of course, the person whose percentages are above average will usually be the one whose luck is better, but I've had games where the percentages looks close but one person was getting most of the key rolls, or games where the percentages were pretty skewed but the person with "bad dice" was rolling just enough quite often.
If you really want to break down luck, you have to just keep a table of each attack's dice rolled, and their results. Then you can compare each roll to expected return, and see who came out ahead. Even this doesn't account for a lot of positional and OM-related advantages that a good roll can confer, but it gets as close as anything I've tried. I've only had the patience to actually break down a game this way once or twice.
So, tracking skulls/shields rolled is obviously a lot easier, and it's fun to look at, so go with that.