Mathammer, or the attempt to apply statistical analysis to the game of 40k, is a tool in the box, so to speak.
Understanding the mathematics behind how the game plays gives you a more accurate means of planning your attacks. If you know that [
completely made up numbers] a trio of Autocannons has a 75% chance to inflict 2 or more HP against an AV 11 vehicle, you can be pretty confident that if it's out in the open, and has already taken a HP of damage, you
probably only need to attack with the one unit to deal with it. So you could potentially move a second / third unit away to attack something else, reasonably confident that the single unit will be sufficient.
That could be important if you have a secondary target unit that you need to deal with, but don't have enough without committing the extra resources.
If you follow what I'm getting at, a strong understanding / memory for statistics can help you to make the best possible decisions in the game. It can give you an edge against someone that's going with a "gut feeling" about the situation. Not necessarily, and luck still plays a role... and if you smear goat blood on your dice the night before then the Gods of Hope and Wishing will carry you through, regardless of the stupid decisions you make in game, so such knowledge is
completely worthless in the eyes of some.
But if you aren't into ritual sacrifice, or you just don't want to make a mess of your carpet! You can get an edge in-game with the application of Mathammer. It can also help in the list-building phase of your game, by helping you take units that most efficiently fulfill the role you want filled. This applies not just to how much damage a unit can deal to a target, though that is the most common use. It can be things like, "Which is a more efficient bullet sponge, Ogryns or Conscripts?" You might want to use infantry to screen a tank, and that's their main goal. So you'd be looking at durability of your own unit for the least investment, so that the unit you choose can withstand the greatest amount of... Gauss Blasters... to pick a weapon. The idea being that so long as you have a couple Ogryns, or 5 Conscripts, they can provide a cover save to your Russ.
So you calculate how much damage that 30 shots can do to Ogryns, and then how much damage they can do to Conscripts... figure out how many points you lose per turn on average... and then pick Conscripts because they're dirt cheap and that's a perfect job for them.
If that's the purpose you need a unit for, you can use Mathammer to calculate the better unit for a task.
It doesn't factor movement, or other myriad forces at work. For example, Meltaguns are always preferred to Lascannons at short range, but can you consistently get there? Many times, Lascannons are better because you can attack at range, from the first turn, instead of needing to get into position [Questionable, as a Guardsman!] for a couple turns. So Mathammer is a tool to help you pick units, but you still need to apply the numbers with wisdom.
Like all tools, their worth is dependant upon the hands crafting with them. A master craftsman can be limited by poor tools, but that same master with masterwork tools can craft incredible things. That said, a novice with masterwork tools is still a novice, limited by their skill. So yeah, the
ideal of Mathammer isn't clairvoyance of the future. Expecting it to give you perfect prediction [Especially when relying on averages!] will lead to disappointment. Ideally, you are simply aware of the risks you are taking, and are then able to wisely judge that against your rewards of given actions.
That said, I've never been a strong believer in "points / damage" as it's a cart before the horse thing... but Mattler and I had that discussion a long time ago. Although Explodes are rare, they should still be factored into their potential, which pure HP stripping doesn't account for. So there's another aspect to consider, whether or not you agree to the premises presented in the statistics. In this case, I don't agree that the points / damage system is accurate, since it doesn't account for Explodes results.