"Gentlemen, we are being killed on the beaches. Lets go inland
and be killed."
It's a war game. Troops are going to die. The models you spent countless hours converting and painting are most likely not going to live through the game. How can we make sure their deaths are not in vain.
There are different reasons that troops die, and it is important to know the difference. Troops can die as part of a plan that puts them in danger, or troops can be outright sacrificed, hopefully to fulfil some other objective. It is important to draw this distinction before we discuss sacrifice. If your marines assault your opponent's marines, any that die as a result die because the plan involved some risk. By getting the charge on your opponent, your marines actually have a statistically better chance of winning this fight. You'll probably have some marines left after the combat. This is not a sacrifice.
Sacrifice occurs when you ask your troops to do something that you have no realistic expectation that they will survive. Sacrifices occasionally have to be made in wars, and at least you don't have to write letters home to your model's mothers. But, before planning a sacrifice move, you want to make sure that it will accomplish something. A sacrifice move is always intentional - you are intentionally expecting your troops to die for the cause. Unintentional sacrifice is also called a mistake, and we try to eliminate those.
So what do we expect to accomplish through sacrificing our units? There are a number of possible reasons to sacrifice troops. Some of them are fairly obvious, and some are not so obvious. But, it's good to go over them, so that we know what to expect.
Sacrifice a Unit to Kill a Unit
This is possibly the most obvious sacrificial move. It involves knowingly putting your unit at risk in order to kill a (hopefully more valuable) enemy unit. These sorts of sacrifices are often made against vehicles, as vehicles are the most likely targets to be worth enough points to justify such a move. Still, sometimes, you may sacrifice a unit with a special kind of weapon or attack, in order to kill some opponent models that may be especially vulnerable to such attacks.
Some examples of sacrifice to kill units include deep striking behind enemy lines in order to attack the rear of their vehicles, assaulting heavy weapon teams knowing that you will be either counter-assaulted, or left out in the open, or moving out of cover to take shots at something, knowing the enemy has enough firepower to destroy you as a result. Another really good example is sacrificing any unit in order to destroy an enemy unit that is going to score extra points for holding an objective. The extra victory points granted for holding an objective make such a move a good decision.
It is important when planning such sacrifice moves, to have a sufficient chance of success to warrant the loss of your unit. If you have a 100 point unit that you know you will lose as a result of your maneuver, and you only have a 50% chance of killing your target, your target should be worth at least 200 points to justify the odds. If you have a 75% chance of killing the target, the target should be worth about 133 points, and so on. In Poker, this would be called the Pot Odds, but the concept is the same. How much you are willing to risk is based on the odds that you will achieve the result. Divide the point cost of your unit by the odds that they will accomplish their task successfully, and you have the number of points that the target should be worth to make the sacrifice meaningful.
Sacrifice a Unit to Protect a Unit
If you can sacrifice to kill your opponent's units, it stands to reason that you can also sacrifice to protect your own units from destruction. This can take a number of forms. And, as with sacrifice to kill, you want to make sure that the unit being protected is more valuable than the unit being sacrificed. It makes no sense to protect a unit worth 50 points with a unit worth 300.
Some examples of sacrifice to protect other units include using empty transports or low-cost vehicles to block lines-of-sight, making a less valuable unit the closest target, or placing a less valuable unit in between an enemy assault unit and a unit you want to protect.
Unlike sacrificing to destroy the enemy, there are limits to sacrificing to protect your own troops. If we could sacrifice every unit in our army to kill every unit in the opponent's army, we could ensure a draw every game at the very least. If we sacrificed every unit in our army to protect some other unit, we'd be left with only one unit, and it wouldn't be protected anymore.
As this is the case, you should think very hard before pursuing this course of action. Does the protected unit have any other possible way of being protected? If it can move behind cover, that is better than losing another model to block line-of-sight to it. Does your ability to win the game hinge on the protected unit? If your opponent has several tanks, and the only heavy weapons you have are in the protected squad, the answer may be yes. Finally, will the sacrifice ensure the other unit's survival for more than one turn, or at least allow it to win back the points being sacrificed? If a unit is going to die in two turns regardless, there is no reason to sacrifice another unit to protect for one turn.
Sacrifice as Bait
Sometimes we want the enemy to go somewhere that they really don't want to go. You have an assault squad, but the enemy is just a little too far away for you to assault them without sustaining significant shooting casualties. How can you solve this dilemna? If you place a sacrificial unit where the enemy has to move to kill them, then you may make the opening you need.
As with the other sacrifices, there are questions that you should ask yourself before going through with the plan. Can the enemy kill the bait without moving in the way you want them to? Sometimes your opponent will spring an obvious trap with a different unit than the one you planned on them using. Will the loss of the bait allow you to improve your position, or score more points than the value being sacrificed? If not, why are you doing it?
Sacrifice for Time
Sometimes, you just need a little more time. Sometimes, you need to ensure that a combat will end on your opponent's turn, instead of on your turn. Sometimes, you need to prevent an opponent's troops from moving in a straight line towards an objective. In these cases, throwing more troops at the problem can be the answer.
If your opponent is threatening to end a combat on your turn, allowing them to move and assault another unit, you might as well bring that other unit into the fight, to keep his troops tied up for longer. You're not going to with the battle with your basic troops against his elite combat unit, but you can prolong it long enough for your other troops to get outside his assault range the following turn, allowing you to shoot the assault unit before it progresses through your ranks.
Similarly, sometimes, your opponent will want to get somewhere, and you don't want them there. This may be the case in turn 6 of a mission, where his troops have the opportunity to capture an objective. Rather than allow them to move unobstructed, you can put another unit in his path. They then have to either go around your unit, preventing them from getting to the objective, or they have to engage your unit, locking them in combat, also unable to achive their objective.
Knowing when and how to sacrifice units has been the mark of competent generals as long as humans have fought wars. Sometimes, 1000 more bodies can hold a pass for one more hour, allowing reinforcements to arrive, or buying time for others to evacuate. It's a useful tool at your disposal, as long as you think through the moves before making them. As I said before, unintentional sacrifices are also called mistakes.