Learn to Protoss! Okay, not exactly, but Havoc asked me for some help the other day and I figured I might as well try to spread the help around as much as possible. I'm just going to write down the things I look for and think are important in the different matchups. Obviously, I spent a LOT of time away from the game, but I was pretty good when I stopped and game knowledge is one thing I'm pretty good at so I feel confident making these points. Note that I don't intend to give you builds here, just what to look for and some things to think about from what you see. I'm leaving the builds completely up to you, mostly to make it easy on me - this stuff is true no matter what.
I'm gonna start with PvT. When scouting, the important things to look for are gas timing, number of refineries, number of rax and if any addons. Now how important each of these is relies on what build you're doing. For instance, I do a DT opening alot, so I really need to know if he's getting two geysers fast, but the question of whether he's rax expanding or going tech lab or reactor aren't as important. If you're doing a fast expand build, you'll need to know exactly what those addons are and when extra raxes go down, etc. Realize what pieces of that list you need to know for the build you're doing, scout for them, and then get out of there, if you can. If you're checking for the timing on his raxes, as you would if you were gate expanding, you're going to lose that probe almost all the time, but if you're trying to do an aggressive rush push, you'll want to get that probe out alive so it can drop a forward pylon. Honestly, scouting in all matchups comes down to determining exactly what you need to know, then finding it out.
Alright so after the probe scout, you know a little bit. If you saw 2 gas quickly, then starport play is definitely on the table and you'll need to adjust accordingly. Likewise, if you see no gas, then obviously a quick starport is out of the question and a fast expo or a marine or marine/scv all-in are on the table. In both cases, this information mostly affects the robo timing in whatever build you're doing. Obviously if they're getting no gas, you don't need a fast robo, but if they're getting two, you do. There are other subtleties, of course, but that's too build dependent for me to want to go into here.
If he's walled and you got there too late to get in, you can still see things like a marine count, whether he's getting an add-on on the rax and what it is, etc. Obviously if there's no addon, the build is not a quick reactor or tech lab fueled push and you can cross those off your list of worries. It does leave a fast expand, a gasless all-in and a 1-1-1 variant on the table though, so using your units to poke and see his unit count and whether he bunkers/how many bunkers he puts down is a very good idea.
Now once the builds are going on either side the question starts to shift to how aggressively can you implement that strategy. The question of unit losses comes into play strongly here, as if you've lost more units than your opponent you either have to be a little less aggressive or wildly more aggressive, depending on how big the difference is and how you want to play it out. Obviously if you have lost less than your opponent, you can afford to be slightly more aggressive with your plan (keep in mind aggressiveness doesn't mean attacking necessarily - it can mean attacking, but it could also mean teching a little faster than usual, expanding a little sooner than usual or getting up more gateways quicker than usual because you don't have to make out of each of them quite as much, all of which speed up the later stages of your plan.)
A quick quantification of being slightly aggressive: in a PvP you pick off a stalker early on without losing anything in return, so you make an extra probe. That's a slight speed up of your overall plan that was granted you by your early fortune. Being wildly more aggressive would be something like taking a super quick hidden expansion after being that person who lost the stalker. It's probably going to lose you the game, but if not you're going to be ahead.
Speeding up or slowing down your gameplan is alot of what the midgame is about so definitely pay attention to unit losses. Resource count of lost units is definitely important, but it's not the end-all be all factor. Trust your gut, and if it turns out it was wrong, watch that replay and find out why.
Most of the late game decisions for Protoss are already made in this matchup. You pretty much know what army composition you want, though you may want a little more of X and a little less of Y depending on their composition, but the actual units will stay the same. Scouting is the key in this part of the game, because stopping expansions before they can turn into planetary/turret/scv strongholds is your goal. Honestly, late game decision making for Protoss is essentially this: scout expansion -> decide to kill expansion. If you have an army composition with no colossi, you have more options, including lots of openings to attack various positions with warp prisms and the like, but if you are making colossi, the decision facing you is attack one position at a time with your army ball of death or lose.
Alright so that's definitely not exhaustive, but that's all I really care to write down right now. Maybe it's a little helpful, though. I'll do another thread for PvP and PvZ later on and maybe expand on this some more later. Feel free to fire away with questions and comments.