It's really the same argument I have for carrying troops over if you fault in manual deployment. To me maintaining the troop count is important.
Imagine playing on the classic map with escalating reinforcements. You get to the point where a few players are receiving 2-5 or so troops for bonuses and a 3-5 for territories. So a range of 5-10 reinforcements per turn for each player. Spoils have escalated to 20 for a set. At this point in the game spoils far out power reinforcements and to maintain any sort of balance you have to pit spoils against spoils. Some use them aggressively and others defensively but what really matters is the math behind it all.
Let's use a single battle as an example:
Player A receives 8 troops for bonuses & territories and another 20 for spoils.
Player A then places them all on a single territory.
Player B receives 4 troops for bonuses & territories and another 22 for spoils.
Player B then attacks Player A. A battle of 26 attacking vs 29 defending.
Based on this calculator it's a pretty fair fight:
http://riskodds.com/ (27 on attacking, 29 on defending)
Now imagine Player A faulted on spoils.
Then it's a battle of 26 vs 9. A guaranteed win.
At that point it doesn't matter how good you're strategy was. Mathematically you can't win an engagement solely because of the odds spoils introduce to the game. You're bonuses and territories don't really matter once spoils get high enough.
Sure it's your fault for screwing up the spoils deployment but if thus far it's been an even game with some good engagements back and forth this pretty much throws the game. It's not a satisfying way to win and pretty disappointing way to lose.
That's a really simple example and dispersed across a board things do look totally different. It's still staking the odds against one player over a silly mistake that really isn't a part of strategy but has the potential to ruin it. To me it's not ideal to win or lose because the odds were suddenly stacked against one player because of a mistake outside of their strategy.