[G] Using ranged phoenix in PvZ - Page 2

On January 01 2013 04:38 ineversmile wrote:
I have been using Gateway expo-->Phoenixes in over 80% of my PvZs in the last few months. From trying to make it work, I have come to a conclusion: high level players don't go Phoenix throughout a game because it's almost impossible to execute decent protoss macro while controlling your Phoenixes. In the early game, it's not hard because you are just building probes, phoenixes, and then you can have a gateway explosion and maybe an upgrade or two. But if you're on a 3-base economy and you're still getting to that 70ish probe count, you have to do all of the following:

-build probes from 3 Nexi
-build production
-decide a tech path
-check upgrades
-position your army
-keep an eye out for runbys
-keep an eye on your opponent's tech and unit composition
-micro your phoenixes
-choose which units you're building
-warp in units regularly
-queue units from startgate
-keep expanding
-make a decision about how you much/how little defense you want at each base
-set up the right wall-offs where they are needed
-keep an eye on your opponent's expansions
-chronoboost

This isn't a list of things you do one at a time; this is a shitload of tasks you have to constantly check, over and over and over again. It's unbelievably difficult to handle that. So, I've been trying to simplify it by just doing Zealot/Immortal/Phoenix so I could just queue from robo and startgate and then just build 3-4 gates per base and warp in almost all zealots. That made it like a modified protoss Muta/Ling, as primarily phoenix/chargelot. I though that would help things--and it did, but my money still skyrocketed so much.

So my next plan was to try to take "macro breaks" every 15 seconds or so. I would only worry about controlling my phoenixes, and building units that could be queued (probes, stargate, robo) without looking. After ~15 seconds of harassment, I would back the phoenixes off somewhere and build structures, warp-in zealots, etc. It was a a good plan, but ultimately it wasn't enough. My brain would want to take a break, too, so after about 10-15 minutes of playing, my mind would just shut down and I would just throw the game away in one form or another.

I have been pondering the reason for my failure. What makes Chargelots and Phoenixes so much harder to play than Speedlings and Mutalisks? Why is it such a reasonable, fluid midgame for zerg and such a daunting task for me, as protoss? Clearly, it's not the units' fault--chargelots and phoenixes are both amazing DPS machines that are great for harassment and strong against a lot of zerg midgame compositions. Here's what I realized:

1. Phoenixes take more work to use than Mutalisks. They may fly faster and auto-attack, but having to G-click-click-click is focus-based micro. Actually, it shouldn't even be the G key, for a number of reasons--but I don't have time to go into those details. Besides having to press buttons and click precisely on the right target(s), there's also the fact that you have to use your decision making about what units to pick up, how many to pick up at a time, and when you need to leave. Mutas also require good control and an understanding of where and when to fight/to flee, but using Phoenixes for a similar job...it taxes my decision-making. This isn't anything to say that Mutas are a better unit or there is an imbalance, but it's something to consider when you choose to use Phoenixes for a long period of time. The unit takes a lot of babysitting. It's a hybrid air harass unit AND spellcaster.

2. Zerg macro is mostly about building one tech structure, the occasional hatchery, regularly hitting injects, and building units without looking at your larva. Protoss macro involves queuing workers, queuing stargate and/or robo units, warping in unit in the right location, adding on production buildings chronoboost, and probably one or two other major things. I guess that the chrono and injects are about the same, and upgrades are basically the same...it's just that the warp-ins take more actual mechanics to use than building stuff from just hotkeyed buildings. For zerg and terran both, they can do that--for a toss, you have to actually look at where you're putting them, and wait for a cooldown.

I almost want to just turn my warpgates back into gateways and queue my units, just so there's one less thing I have to look at when I macro. I think it would seem a lot less stupid to use regular gateways if the cooldowns for a WG weren't just so much better than building from a gateway. It's seriously like 10 seconds more per zealot/stalker from a gateway than from a WG...I kinda wish that the WG tech upgrade reduced the build times to the cooldown length of a gateway. But it doesn't, so I guess that's a moot point.

I mean, there's also something to be said for just overmaking gates and then having big warp-ins at intervals longer than the cooldown for zealots, but then you have the common problem from the gateway explosion in PvZ: you have a billion production facilities from powering, but your supply is just shat on because you have only built gates and you don't have actual units. My goal is to improve my ability to macro while using Phoenixes, so my supply can be close to the zergs due to a combination of harassment and of building those cheaper cost-per-supply zealots.

Like I've said before, I'm totally not complaining about zerg having better macro mechanics than protoss--all I'm saying is that muta/ling is a logical progression for zerg because most of their macro is done without looking at the zerg player's base. That's not really true of protoss. So if we could solve this problem, that would make it way easier to keep an eye on the delicate units while still keeping up with macro. Maybe there are some techniques we can figure out, or something. I mean, it's possible to keep building the same things constantly while controlling phoenixes and an army, but usually that's about as far as my brain can handle...if I try to transition or to otherwise tech, or to think about some kind of big-picture reality of the game, I'm screwed.

--------

OK, so that's an explanation of the issues with phoenixes in the midgame. But what about the lategame? I think it gets a lot easier, at that point. Because there are already 70-75 probes on the map, that means I don't have to keep checking my Nexi and saturation--I just make sure that I take another base as one of them begins to dry up. So that's one less thing to worry about--and a big thing, at that. So in the late game, because there's no worrying about my economy (aside from it getting attacked), I find it a lot easier to do multipronged harassment, and also overall easier to focus on tactics and positioning. Having more time to think about that stuff directly correlates to being able to do it.

So, to me, it's clearly possible to use Phoenixes in the early game and to use Phoenixes in the late game, in PvZ; as far as being able to control them and an army while functionally macroing. The question is how to bridge that gap in the mid game. I know that I like to just send them home shortly after the infestation pit finishes, but there's a timing window before then when I want to do as much damage as possible, and that's right when I'm:

-Taking a third
-Setting up a proper wall for my third and natural because I'm really worried about runbys at that time
-Establishing another tech path properly--not overteching or underteching
-Building more production--the right kind, and the right amounts of each
-Taking my next set of gases at the proper timing

That's a lot of decisions to make, and even the things I can memorize (building placement for walls on each map, gas timing, etc) are usually thought-intensive and require careful deliberation. But, simultaneously, ain't nobody got time for that--so what to do?

I think that there has to be a way to make it work, but there's a fundamental technique that I'm missing. I keep thinking back to the early days of the game.

I'll write more on this, later. Still thinking about making this into a full-blown blog.

On January 04 2013 11:31 ineversmile wrote:
I find that the armor upgrades aren't as big a deal for dealing with corruptors; it's more about infested terrans, queens, and hydras. Corruptors against Phoenix is a lot like Roaches against Zealots...the armor isn't that big a deal because the zerg unit isn't a DPS machine. And for mutas, you should own them hard with range anyways, so armor is irrelevant. I would rather just build another stargate, zealot, and pylon for the 300/150 required for second core and earlier air armor.

EDIT: I played some games tonight. Here's one of me being very tired, and losing my grip on macro when I hit 3 bases. But it shows what I've been doing, when I actually go for the whole "Phoenix all game" plan.

http://drop.sc/290816

I got a lot of pressure in and canceled his third a couple times, giving me an earlier third, a constant worker lead, and a tech advantage. I even had a faster fourth, and a few timings to move in for the kill. My macro totally fell apart, but even then--I was even 170-170 supply just before the 20 minute mark when the greater spire was done.

I think there were a couple of key points in which I could have done crippling damage, or set up checkmate situations.

The first was when I went home with my ground army after a hatch cancel and I was taking my third, and while I did that I went into his natural and main to kill queens. I was busy macroing hard, so I kinda just let phoenixes do their thing and I lost 1-2 to the spore in the main. I panicked and pulled back, but if I had stayed I would have killed 4 overlords minimum, denying infestors for an extra minute. I knew infestors were coming, so trading phoenixes for overlords through a spore or two is perfectly reasonable, and it's simple micro: I could just shiftclick attack them, then shift-right-click home. But I got tunnel vision with macro, and didn't look at the opportunity at hand for my attack.

The second was when I had 4 immortals. One of the obs pointed this out after the game: if I hit then, I could have won. I also could have won right before broods morphed, since the infestors were all just below 75 energy. That's the thing about the phoenixes against infestors: even if the phoenixes all get killed by infestor energy, infestors without energy are basically worth 1 infestor egg each. So you can create timings by trading phoenixes for energy, then going in for the kill when they can't chain-fungal everything and they can't afford to make an army out of energy they don't have.

And lastly, I could have set up the mass stargate play more efficiently, but that was water under the bridge because the pre-brood timing should have ended the game, whichever I chose. I should have scouted in and figured things out.