....the assumption I should sit back and let them attack since a maxed protoss army should be more powerful than a maxed zerg army.
Yeaaa shame it doesnt actually work like that.

One finds that a protoss deathball is just as susceptable to death by positional weakness as any other race, and as zerg uses more higher tier units and most importantly infestors then I think we'll find that it can be just as balanced as any other race's maxed army.
Ok, about the game!
(quick note: by a saturated base I mean at least 16 probes on a mineral line)
Build Advice
Gas Income:
If you put 3 in the gas as soon as your cyber core starts you'll have enough gas to start warp gate immediately when the core finishes. Because of the late cyber core and very late gas income you were stuck building zealots for a long time. In this situation you are seeking actively to do damage while your zealots still have high utility against your opponent's units and economy (queens in particular), rather than investing those minerals in, say, an expansion of your own. I would suggest that if not expanding early you'd be better off with a gate->gas->core->more gateways build.
Quick question: Was the robotics tech a response to seeing the roach warren?
Probe Production:
You shouldve bad enough probes at your main to just about saturate the main and the expo by the time it finished, so you wouldve instantly had 2 fully functional bases. Remember you get additional income for every probe up to 3 per mineral line. So that's up to 24 probes to maximise your income on an 8 patch mineral line. Generally as you get to 4-5 mining bases you want about 2 on each mineral line because the 3rd probe will give less marginal income than the other 2. But this means that you can keep building probes in your main and
keep getting more income. This is very very important when taking on terrans, because you have to expand ahead of them to stay ahead, and you have to use chronoboost on probes to equalise on income.
Also don't forget that you can have more than one nexus building probes. This is incredibly important for fast expand builds, because you can chronoboost out twice the probe production and match, if not surpass drone production from a zerg, while still being able to build an attacking force. It is also very useful when you have a main and natural that are still unsaturated at 14:20. You ended up in a position where you were slowly building probes one by one from 0 at your third when you could have had a saturated main and 25-30 probes at the natural. Transfer the excess over 16 probes to your third and you have a near fully saturated mineral line. You'd be surprised how fast 3 nexuses worth of chronoboosted probes can make up the difference and get you a fully functional base running.
A note on probes in the gas:
I've never considered rallying 3 probes of production to the gas, although from what I can see there, it may lead to mistakes such as too few or too many probes in a given geyser. As your game improves and you become more focused on other stuff it'll be more difficult to have that focus to re-rally your nexuses like that. I would suggest that if you want to put probes in gas, just box 3 of them and click on the geyser. Keep building probes as normal from your nexus and you never have to worry about going back to reset rally points or anything like that. Just less to worry about over all
Unit producing structures:
I was a bit confused when you added 2 more gateways and a stargate, but continued collossus production. This brought you up to a total of 6 gateways, 1 stargate and 1 robo facility. All this was added and you had 200 minerals and roughly the same amount of gas. And your 4 existing gateways were ready for a production cycle! If you'd taken a third just before then it wouldve made sense, but don't go overboard with unit producing structures if you can't build out of them. Your money was keeping pretty low that game, partly because you needed more probes

, so you were certainly ok on structures. Did you panic and want to mass stalkers and phoenixes? 4 gates + a stargate wouldve been fine. But you also made collossi so you were never going to have the resources to sustain that production.
Unit based advice
About zealots:
At about 7:00 we see how bad zealots are at dealing with roaches. It's ok to keep producing a few more zealots against zerglings but either way you need to get stalker production started a bit earlier. Stalkers scale well against zerg compared to zealots as you approach the midgame and actually do ok against zerglings when not surrounded and supported by a few zealots. As you approach the mid-lategame zealots should generally be a lesser part of your army composition (unless they really are just going pure zergling in which case roflirl and build zealots

. But chances are that they're teching to mutas in secret). This is a problem I have, that I rely too much on being able to build chargelots as a core part of my army and they just get destroyed against a late-game zerg army.
Dealing with roaches:
Don't bother, it's gg as soon as they start massing them.
Dealing with mutas:
Don't bother, it's gg as soon as they start massing them.
...
Seriously though, in both cases, it's all about stalkers. If your opponent starts investing in roach upgrades, get an observer out, a few immortals and get blink! Also keep investing in attack upgrades. Alright let's do this seriously..
Dealing with roaches:
In the early game the most important thing is having something that'll hold a roach all-in, especially while expanding. This can take one of two forms: stalkers or cannons. I like to play a very greedy mass-expand style so I tend to rely on having one or two cannons at the natural at the start to guarantee my early nexus. If you don't want to expand that quickly invest in some stalkers. Any zerg will also tell you that forcefield from sentries is just a pain, because they can't just take their roaches, run in and take advantage of the range 4 and frankly unreasonable damage on roaches to smash you apart. Get the right position and you cut their army in half, then a quarter, then an eighth and suddenly all of their units are vulnerable to stalker fire. A few pew pews later, they have nothing left and you're free to secure another expansion
Dealing with roaches later on:
Roaches are one of zerg's best scaling units, and best of all they're super cheap. They get huge gains from +attack upgrades (don't let the +1 fool you, if you look at the attack values they go up by more than one

). So while your army dies, they can replenish their roaches in greater numbers for a lower cost. Conclusion: Don't let your army die. There are a few ways to go about this in the mid-late game.
1. Collossus deathball - Yea ok, not that interesting but it can get the job done if you have ok positioning. I'm not sure how unbeatable it is as a counter now that zergs are using better units in their composition though.
2. Blink Stalkers - Blink stalkers alone are
alright against roaches. Blink allows them to stay alive, whereas assuming you have an observer they don't have the luxury of a quick retreat. However, if you get behind at all(!!) on attack and armour upgrades they will just run in and kill you. These battles will be very much about positioning, and sentries with
3. Blink Stalker/Immortal - My personal preference. I've been playing around with responding to mass roach with a 2 robo immortal style on 2 bases with mass blink stalkers. It's actually pretty sweet

and you can invest in chargelots and templar pretty easily as you secure a third to help deal with hydras and infestors in the lategame. (Templar will be very important in this, as said I use chargelots too much).
Dealing with Mutalisks:
Unless you're being rushed, in which case cannons and stalkers are your friends, mutas will only tend to pop up in the midgame. There are some pretty brutal transitions which if not correctly scouted can take you by surprise whcn 15 mutas waltz into your base and start killing your dudes.
General guidelines -
1.
2 cannons at
each mineral line to
delay the mutas. They will not destroy them, but they may fend the mutas off for a while and at least give your stalkers time to get there.
2.
Twilight council + blink. This should happen ASAP as soon as you know mutas are coming. What makes them so difficult to deal with isnt so much the damage they do (although it is insane), but the mobility they have. So assuming you havent been massing phoenixes from the early game, stalkers with blink are your best option to chase them down. Get as much vision on the map as you can and react as fast as possible with your stalkers as soon as you see an attack vector from the mutas.
3.
Armour+shield upgrades. Another thing which makes mutas incredibly annoying to deal with is their bouncing attacks. 3 bounces in total, so a single +1 armour upgrade negates 3 total damage per shot, so +3 will negate up to 9 damage. A worthwhile upgrade to be investing in to be sure!!!
4.
HT+Storm. High templar are not what they once were. Once upon a time, as soon as mutas came out, you could drop the templar archives, and as soon as storm and amulet were done you were in pretty good shape. Once mutas were out, you'd warp in 2-3 templar and storm them to hell. Not so any more! Now our core defence is stalkers with blink. But that doesnt mean that storm isnt still useful. It does massive damage, and mutas clump as much as any unit in the game. Of course it's pretty easy to just retreat from a storm, but that's exactly the point. It does some damage and buys even more time for your stalkers to get into a good defensive position. It also delays their harrassment. While your core army continues to get stronger.
Once the HTs are out of storm, you can morph them into Archons. Archons are not very mobile in this context, but in the right spot they'll make mutas their ***** and have them for breakfast.
5.
Go kill them?. Funny thing, but mutas are really expensive. Catch an inadequately prepared zerg at the right time before this investment pays off and it's an easy win.
A note on carriers:
Keep building them, they're awesome.
A Better Build?
I'm hesitant to do this, because ultimately you should feel comfortable and happy with what you're doing, so I dont want to just come along and preach my favourite style if you're happy with what you're doing. Also I think you have more important things to think about in your gameplay. Your macro is your weakest point, so that's what I'd suggest as a main focus point. Keep those probes building, and get a sense for how much you can do production-wise with the resources you have.
But on Tal'darim I find that the best style is an aggressive mass-expanding style. You need to be able to assert map control and keep the xel'naga towers. Otherwise you end up with the problem of your opponent having a large number of potential attack paths, whether that be generally looping around paths to your 3rd/4th or dropping anywhere around your base (or in this case, muta harrass). The best way to deal with it is vision, and vision is obtained by asserting map control and placing units and observers correctly around the map (zerg has this easy as they can put overlords wherever they like).
So what's a good build? Honestly, I think a 2 gate zealot opening is going to be very weak on Tal'darim. Zerg will almost always fast expand and you'll instantly be behind unless you can pull off a miracle harrassment (due to the long distances). You saw the weakness there when having to deal with roaches. You ended up very far behind.
1. Early Nexus:
The ever popular greedy protoss build. It requires scouting to do effectively (or just luck?), but the idea is essentially to wall off your expansion and build a nexus as fast as possible, with 1 cannon (2 or more if under big pressure) to defend.
2. 3 gate expand:
Yea I always mention this, because it's so good! Can be defensive or offensive, but either way you get an easy expansion from it. Idea: get 2 gas, 3 warp gates and warp enough sentries that you end up with 400 minerals for an expansion.
My preference is an early nexus, because it's not easy to pressure in the early game and it's not easy to guarantee enough damage to justify delaying the expansion. But this section whould be taken with a pinch of salt. Until you're simply dying in the early game, do what works for you.
Your main focus of improvement should be on macro and getting into the rhythm of probe and unit production. Once you have that down, you'll already be well into diamond league

so then we can talk about precise builds.