Halberds on hold fire to find the phantoms and then lance to kill them. success rate is better if you use an iris to cloak your lances.

With enough daggers, reavers will die in a single group barrage - As will the accompanying phantom. A hold-fire halberd is indeed a good idea for soaking up porc and reaver damage whilst your daggers get in close.

Note that you only linked one game, not two. In that game, by the time the first phantom came out you were already way behind on total value (10k vs 15k). So there was probably nothing you could do by that point. You have to win the raiding and expansion game first.

Don't build nuke. Halberd is your friend.

Looking at your second game, where he had 4 phantoms and a lot of reavers and you had similar total value, maybe facswitch to get firewalker would make sense. You can also terraform low walls (height 80 is sufficient, could be less, depends on the height of your units) to prevent phantoms from shooting you from a particular direction.

edit: cloaking your whole army with an iris also stops the bleeding and makes it a lot easier to chase down the phantoms.

Mace is extremely vulnerable to Phantom. If you need to escort your Lances with something other than Halberd you are probably better off doing it with Bolas or Daggers. At least you will be giving away less metal per Phantom shot.

Stream of consciousness time!

First match

B952903 2 on FrostyCove v1.11

• Hovers probably aren't a good factory choice on FrostyCove. Hovercraft are very sensitive to even minor terrain changes, and FrostyCove is quite hilly.
• That initial raid hurt you a lot and put your economy behind. A lotus on the highground to the side is quite a strong deterrent and can easily make cost even later in the game.
• This is the state of your relative economies 10 minutes in:
• Energy stalling is an economic crisis. You should always make sure you have enough energy to match your metal production, and outside of the first minute or so, almost always want to have a buffer in case of reclaim, repair, and to overdrive your Mexes.
• You have an uncaptured Mex on the high ground in your main base, while your opponent has not only taken all of the Mexes in their main base, but have taken all the Mexes on your hill due south of their base. At this point, they have nearly double your income.
• Major overdrive in the sense of fusions does have diminishing returns, but incidental overdrive in the sense of each Mex having one solar does pays for itself only slightly slower than taking Mexes. You opponent has at least one solar by each of their Mexes, and often have more.
• You have 5.3k in an army, which is very nice at this time. Unfortunately, 2k of this is in two Lances, which I feel is a little early in general, and one of which was definitely wasted by having it stall indefinitely trying to shoot a Gauss to the west. I feel that this would have been better spent on mainly scalpel, mixed with some mace, and the occasional hold-fire Halberd for vision/anticloak. Scalpel+Mace is very hard for most factories to deal with. 2k in artillery is a lot of metal for the density you get in a 1v1, and you would usually be better off switching to planes or gunships for a locust or raven deathblow before the lines truly saturate to the point where artillery is warranted. The same comment applies to your opponent making two Snipers this early on.
• 12 minutes in:
• Nice raid on the west! It's very easy to tunnel vision in on single areas, so it's good to see you poking all over the place and forcing your opponent to respond every time.
• Solars into fusion solves all energy problems for the foreseeable future, splendid!
• It's just in time for that forward Caretaker to help repair your army, and it's in range of most of the wreckages to more than pay for itself shortly, too!
• You've completed a Gunship Plant. This is very good! But it really should be building something, especially as...
• Your economic bottleneck right now is expenditure. You are using only 63% of your income, with the rest being permanently excessed into nothingness. 32M/s to 57M/s isn't a great place to be in, but only 20M/s to 57M/s is debilitating. :(
• 13 minutes in:
• The gunship plant is online!
• You're building a caretaker to help your production!
• You're capturing the Mexes on the hill above your base, and you're finally taking your home Mex too!
• Five Locust is okay. I think I would have preferred more Locust over a Nimbus, since you already have heavy artillery fire support in the form of a Lance, and a ball of 9 locust is much scarier than a ball of 5.
• 15 minutes in:
• The locusts didn't quite stay away from the Reavers. I know Fusions are Caretakers are prime targets, but especially with a small ball you really need to be afraid of riots.
• Your Nimbus is sitting in isolation in the middle of the map. That should really be over your army, so it doesn't have to run away from/die to a single Glaive running into the blind spot directly under it, let alone a cloaked army.
• Still bottlenecking on buildpower. 80% usage is excessing a lot. With that said, your opponent is also on 85% usage and excessing at an even faster rate.
• Your factory's buildqueue is 11 dagger, 1 mace. While there's a lot you can do with Dagger if you fully understand potential burst intervals and have extremely good micro (I don't), it's going to be very unlikely that any number of raiders will do anything meaningful to 12 Reavers (or any similar ball of Riots). I would be much happier with a queue that has something like 3 scalpel, 1 mace, on repeat. (Just in case, there's a state button on the command panel which looks like three arrows pointing clockwise in a circle which turns on repeat mode. When set on a factory, that factory will add finished units to the back of the queue, letting you easily maintain compositions without the risk of your factory going idle. You can hold alt while pressing a unit build hotkey or clicking a unit icon to queue up a unit immediately after the current order without having it go into the repeat queue, so an easy way to build a builder without paying much attention or APM is to tap Alt-QQ, where the first Alt-Q select your first factory, and the second Alt-Q has it build a constructor once.)
• I see your commander has a Trinity queued. It's going to take 11240 metal before that would have any impact on the battlefield. At 38m/s, even if you were dedicating all of your production to it, making no army at all, that's 4m55s of nothing for your opponent to build up tempo. Trinity is a very risky move even if you're very far ahead. That could more cheaply be a Strider Hub and three Dantes if you have that sort of breathing room and were only struggling to finish, and adds to your power with finer granularity if it turns out that you don't have that sort of breathing room.
• 17 minutes in: You did queue up a lot of Scalpels, so your intuitions are in the right place, but it's a bit late at this point, and the game ends.

Second match

B952926 2 on Adansonia v4.1

• Hovers are a staple on Adansonia, and the engagement 2 minutes in helps demonstrate why. Very nice.
• I'd normally make sure my first solar is in range of a Mex for early overdrive. Some energy gets spent on overdrive even before your energy bar is full, provided you're net positive on energy, and the first drops of energy into any Mex are extremely efficient. For example, when you're feeding a meagre 0.1E into a Mex, which is a fairly common situation when you've been building Mexes and energy in fairly even proportions, the multiplier is 1+sqrt(0.1)/4, or 108%. On a standard 2.0M/s Mex, that nets you 0.16 metal from 0.1 energy! Your overdrive will never be as efficient as it can be in the first few minutes of a game.
• 4 minutes in:
• Your first constructor seems to be... reclaiming trees instead of making Mexes in the area? While I always stress that early overdrive is very good, it's not "ignore your own mex spots" level good. By the time you're done and take the first Mex, you and your opponent's respective metal incomes are 25 to 15.
• Your factory is idle again. While you probably want to focus on expansion at this point, you might find it helpful to, once your factory's repeat queue is set up, set it to low priority so that all mexes go to expansion projects first. You can also change the default unit states of Mexes in Settings->Unit Behaviour->Default States->Economy so that their nanoframe build priority is always high, and/or control click the metal bar at the top of your screen to set a small amount of metal reserved for high priority to projects so your expansion doesn't miss out on small amounts of metal spent while constructors were busy moving.
• 5:45 in:
• Your incomes seem to have evened out a bit, going to 32M/s to 30M/s. though partly due to your opponent being slow to expand too.
• Unfortunately, your metal usage has dropped to 83%, and you're only making 24E/s to your 30M/s. You'll be glutting metal and stalling energy very soon, but both problems could be completely avoided if you notice now and start telling your idle builders to make energy structures. This could either be one solar by each expansion Mex or adding a bunch of Solars/Windgens at home, depending on how much relative value you put on being greedy with overdrive, the safety of having energy at home, or having cheap 2000hp buffers around the map (if you plan in advance, you can stick a lotus behind three Solars in one of those Mex clusters to make it very obnoxious to raid. You could also consider taking your natural geothermal generator soon.
• 6:30 in:
• Very very nice raid, and good call in not trying to have hovers crawl uphill with a close pursuit force.
• The prophesy has come to pass: You're still on 24E/s, and are heavily excessing metal into the void. One builder is making a solar between two mexes (nice placement!) but that's not going to be enough.
• 7:40 in:
• Very nice raid, again!
• That's a really early Lance again. I don't think I've seen anything that would call for it, everything should be fightable with Scalpel/Mace, and your Daggers are doing an excellent job raiding everywhere.
• You haven't queued any more energy past that one solar, and have stayed on 26E/s. More importantly,
• aaaaaaa you have 38M/s and are spending none of it aaaaaa you are printing beautiful cashmoney and just setting it on fire aaaaaa
• 8 minutes in:
• :( You have cumulatively excessed 1.6k :( your army could be 50% bigger just from the excess alone :(
• 9 minutes in:
• You built a few units, and now your factory is idling again and we're back to excessing :(
• If you wanted to rush the Fusion, you can change its construction priority in the commands tab so that it has high priority, and will consume available metal before anything else does. You could also set the hovercraft factory's build priority to low, so that it only consumes metal after everything else has had a chance to use it. Repeat-low-priority is preferable to idle!
• Generally speaking, if you're not producing enough energy, you want to place energy buildings that are very fast to build like Solars or Windgens, rather than fatter ones like Fusions, to avoid any e-stall based metal excess.
• 10:30 in:
• Metal usage is at 87%. You have now cumulatively excessed enough to build 11 scalpels (or pretty much any other kind of skirmisher) which would be enough to completely destroy those Reavers. Your opponent destroying your Mexes with scythes, which is normally one of the most painful things a player can do, has no immediate impact because you just aren't spending enough. :/
• 12 minutes in:
• Usage at 101%; no longer glutting, but you probably still want another caretaker to handle any reclaim/expansion.
• Top tier raid! Destroying so many caretakers is going to seriously imbalance your opponents' economy for a while. This would have been an excellent time to have used the rest of your army, to ask whether your opponent can properly handle multiple fronts at once. You even have a Lance, so Stingers won't be a problem.
• 13:30 in:
• Bombers are good, though with an enemy PlaneFac scouted you probably want some fighters as well. An Owl would also be a very good idea. This would not only spot enemy units, but also help sustain radar coverage of enemy units. Radar dots that have been in continuous radar coverage since there last scouted will be approached more intelligently by your units, similarly to if they're in direct line of sight.
• 14 minutes in:
• I don't think the Stinger was a good target to bomb, even if you did have enough to kill it. The selection panel will helpfully tell you what burst damage you have from the units you have selected, if you're unsure.
• I would have been very tempted to return to their main base to bomb the remaining caretaker plus any they had built/rebuilt, and also perhaps any constructors who showed up to help that aren't microed into stealth. Your opponent is still hurting very badly from that raid, and it wouldn't take much to keep it that way.
• As a fallback, if the main turns out to be too heavily fortified to raid, you could consider checking their natural geothermal. If they have managed to recover their buildpower, and it's at a state where you can't keep it down, they'll be trying to reclaim a lot, which makes their energy very precious.
• 15 minutes in:
• Very good engagement, though you really want a few hold-fire halberds to stop snipers forcing you back. Losing the ground where the battle took place means losing the reclaim field too, and there's a lot that died just then. A full 1k metal is sitting there waiting to be taken by the claimant.
• Scythes are out doing scythe things. A faraday at important locations- oh hey, you're already building one in response! Have I mentioned your intuitions seem pretty good? :) In this case, you're not going to be able to finish it in time, but starting one at each other cluster would go a long way.
• An alternative, when you have planes available, is maintaining enough of a fighter presense to chase down and finish off scythes whereever they appear. Since scythes are so expensive, they'll struggle to make cost if their every appearance means boosted swifts strafing them. I'd prefer this over scythes where available, since fighter screens let you play around many other options your opponent might otherwise have, from bombing runs to Dante drops, for roughly the same cost.
• 16 minutes in:
• Wow, the Faraday did finish, because those Scythes weren't microed at all. Sometimes lucky?
• Bombing Reavers with Ravens in places where there are that many Reavers is probably not a good idea. Not only is their health just over the point where you need a second Raven, Reavers (and most other riots, for that matter) exist to quickly chew through anything that gets close. Even Pheonix would be better, since it doesn't need to dive, though a Thunderbird with raider followup would be best, and probably game-winning on its own. Better targets would be the opponent's geo spot, or failing that, Mexes. There's a lot more you could have done that would have forced your opponent to spend money on on a fighter swarm (and Flails high enough alpha to be rather effective against planes, so that's very good for you), overbuild porc everywhere, or lose valuable targets.
• Your usage is back to 73% again. You have now cumulatively excessed enough to pull a Dante out of thin air, which would also have been game-winning on its own. :(
• 16:30 in:
• Ooooohh Thunderbird! You really are going to be terrifying once you're practiced with handling your economy :)
• Unfortunately, more AA has arrived now. It's going to be hard to air strike that with anything short of Likho now, and Cloaky doesn't tend to cluster in valuable enough balls like shield does. Even Likho could still die to a lucky sniper shot, which hurts more than your opponent getting lucky against other sorts of air.
• Trinity :( I suppose your opponent does seem to be pretty passive, so you might be able to get away with it, but I would still feel safer with a stronger army, and one that wasn't so heavily skewed towards Lance by cost. Any push right now could kill you.
• You're only just recapturing the Mexes in your territory, and you haven't taken your geothermal spot yet. There's so much more that could be done before gambling on a nuke that could be easily spotted by the Owls you've already seen flying around.
• 19 minutes in:
• Your expensive hovers keep getting picked off by Snipers. Hold-fire Halberds have 5000 effective hitpoints and are very fast; you could at least kill the scouting Gremlins if not the Snipers too.
• Your opponent still has your Mexes in the top left, which could be easily bombed, and is even brazenly capturing the geo immediately below your army.
• 20 minutes in:
• Destroying the geo was nice, it was a good target! There was also absolutely no AA on the way or back, so you could have done the same thing with Ravens a long time ago.
• 21 minutes in:
• You're still getting picked off by snipers :|
• Good work with the Ravens! You should be able to retake the top left soon.
• You're also finally taking your natural geo! As a note, Geothermal generators are so cost efficient they should usually be made (and morphed) on high priority. You might want to set the Geothermal nanoframe construction priority (for build) and misc priority (for morph) to high so you don't need to worry about setting the states each time.
• 22 minutes in:
• ...your... your opponent is capturing a Mex inside your main base? I guess since your front has crumbled, but there were so many in the mid that could have been taken first. :|
• I'm not sure what the Urchin on the west is far?
• I would have given up on the Trinity win condition long ago, but I guess now it's your main way out. I don't think your opponent is going to go back go being passive, though.
• 23 minutes in:
• Your opponent went back to being passive. Okay.
• 25 minutes in:
• Your opponent's owls spot that you have nothing, and they stop being passive.
• Despite being three minutes in, you're still 1:22 off your Trinity building a missile. If you're playing towards the nuclear win condition, you might want to change the misc priority of the Trinity to high priority, which ensures that metal gets spent on constructing a missile before anything else. If you make that the default state, you might also want to change the default stockpile amount to 2 or 3 missiles while you're there, so you don't end up overspending on nukes if you can't immediately use them.
• 26:24:
• wow that was a clutch moment
• 27 minutes in:
• Your opponent didn't actually have that much at their main base. I get that it's tricky to know where snipers are, but to come back from this situation I think you really needed to land a nuke that killed their army somewhere that you could quickly get the reclaim.
• 28 minutes in:
• That brazen geo turns out to be the thing that keept your opponent alive. Phantoms constantly on the move consume so much energy, though now I don't think you have enough to take them out even if they were permanently visible.
• 29 minutes in:
• Non-skirmishers against riots let the riots do their thing. This situation is pretty much impossible, but if you want to keep going you could try for a Hold-Fire Halberd to try to find and take out the snipers, then get at least two Scalpels to start picking off the Reavers.
• 29:43 in:
• I don't know how that worked, but okay. Really good job.
• 30:28 in:
• Yeah, that was pretty much impossible, but you fought to the end. gg

Closing

Overall, I think your intuitions are in the right place. A bit more practice would help you respond to crises faster. Possibly you're more accustomed to higher density games where heavy artillery and fire support become relevant much faster, and excess is never a problem because it goes to the rest of the team?

Watching the replays to see what your opponent is doing might be a good starting point. Though ignore the spending so much metal on Gauss forests from the first game, which also seems tuned for the heavy artillery meta of high density games.

What a service esainane, thank you so much! as the opponent in one of the games, I learned a lot from that comment.

Regarding my thought process during the game, at the beginning I felt pretty overwhelmed from the speed and power of those daggers and was surprized I managed to get a decent eco out of it (I scouted that you expanded less than I did).
I then stabilized a bit with reavers, but them being so slow, I still had lots of unprotected spots to get raided. Your dagger army was so frightening to me, and I thought you could have done so much more with it.
Once the game slowed down in the midgame with that single frontline, I felt fairly confident, because I thought I had army and eco advantage, so my plan was to be passive and win the game by just massing more army. Periodically you would suicide your daggers into my reaver army with the hopes of killing my phantoms, giving me an even bigger advantage (regarding reclaim/eco and military). That's possibly where you felt helpless that you couldn't counter my composition.
That nuke launch was so clutch! I didn't know about nukes as possible win conditions, so I continued being passive and mass more army for one unstoppable push. In hindsight I could have prevented that by being more active, ending the game earlier.
Thanks for the game Fumichan. I'm learning something new every game, and that game in particular was very insightful to me.

Cloaky easily ruin your not so heavily guarded expansion/base with 2 Scythes + 1 Gremlin.

I can't thank you guys too much, especially for esainane's lengthy analysis. This community is excellent.

I turned off autoqueue long time ago for fear of slowering economy down. Now setting priority solves it, I'll definitely turn it on in the next match. I do like the team room which you guys call lobsterpot but I don't reckon my ladder game is influenced(?) Anyway my midgame macro is absolutely a thing to work on.