#1 vs 2 BarbAI

1 messages · Page 1 of 1 (latest)

tardy cliff
#

Hey,

On Comet Catcher, I can win 1v1 Easy, 1v1 Normal and 1v1 Hard, however once I try a 1v2 (Easy, Easy) I just can't keep up. If I go out for other mexes I'm constantly defending and not producing units or expanding my economy (I'm only human) If I try desperatly with just 4 starting mex and a shitton of energy and energy converters I'm using all my resources for that and the AI is at my door with a gazillion units and my T1 laser towers die in seconds.

I've seen some fancy replays on youtube where a guy is reclaiming all his build up stuff just to get a T2 bot factory out but thats just annoying micro BS gameplay. Is there any way to play the game against multiple AI that works or shouldn't I even try ?

thorny bone
#

I like using liches to bomb coms to cheese strat a win, which works well vs two barbs on hard

fallen oxide
#

I recently put out a video on youtube on the scenario Catch those rare Comets which is 1 vs 3 Barbarian AI with global vision on Comet Catcher.
Eating a t1 lab to get t2 is just normal in this game if that is what you are referring to. Also, it's hard to give you advice without seeing a replay but based on your description:

  1. there are no 4 mex spots on CC, you are probably delaying your t1 factory to much, you should go for 2mex or 3mex opening

  2. it's a map without wind so you are going to starve for energy all game. It sounds like you don't have enough build power since you have to rely on converters. I don't recommend building converters until you have fusion

tardy cliff
rancid ledge
fallen oxide
#

You can totally beat those 2 AIs with some improvements in the first 10 min, because that's where you lose the game.

  1. The most mexes on CC are in the center, that's the area where you want to contest. That means that the best spawn is in the left mid position, as it allows you to expand north, south, and east at the same time and it's the shortest distance to the center of the map.

  2. With 2 solars you can assist the lab for faster construction bot with your com before moving out

  3. The rez bot is expensive that early on and doesn't help you at all, try 2 engineers instead of 1 engineer, 1 rez bot

  4. If you still want to spawn south, you can defend your base with just 1 llt north and 1 south, the gauntlet rush put you so far behind, you need to stop building that to have a decent economy. You overspend resources on defenses instead of army.

  5. Focus more on expanding, you got stuck in the corner while the AI took freely 3/4 of the map. Assist your cons with some units while moving out and you'll have a better eco.

  6. With the army that you had at 10 min you could push yellow and kill him. Even without making other adjustments, by just being more aggressive, you can defeat them. Give him no mercy.

  7. Amphibious tanks are not that good on land, replace them with medium tanks and your army will be better.

  8. Your first construction turret was at 15min, you are going to overflow resources like crazy unless you build con turrets earlier. You already have good energy production, you only need to remember build power.

  9. Once you get T2 labs, stop making t1 units and reclaim the labs. T2 units are so much better.

fallen oxide
tardy cliff
cold briar
#

there's been a few changes over the recent months that have seen the difficulty of BARB vary a bit, so not all replays and challenges are created quite equal...

#

that being said, looking at your replay you're being way too defensive...

#

CCR is basically made for BARB, it plays to it's strengths so you'll have to try to compete with it a bit on its terms.

cold briar
#

r77feed has good tips, i dunno if it'll help me putting together some commentary on a vid...hmmm...

cold briar
#

Ignore me, i got bored 😛

pearl hare
tardy cliff
# cold briar https://www.youtube.com/watch?v=ACBgrmQTm5c

Thanks for this video I learned everything I need from it, it might be boring for most of you but for me it wasn't. Shows it's decently different from the old way of playing TA, you had to put units out then on patrol to get LOS and guardians defending early game etc ....

fallen oxide
pearl hare
fallen oxide
#

the AI does the same thing every time, it tries to leak units in undefended areas or will find an imminent threat and throw all its units at you. I took control of the map secured the leak areas, created a meat grinder in the center and pushed on the sides.

cold briar
# tardy cliff Thanks for this video I learned everything I need from it, it might be boring fo...

No problem. Yes, the patrol function and guardians are deceptive traps. Like all things in the game they have their use, but for the most part CCR is a map about movement and control: guardians are too expensive and there's too much alternative resources available to make them that practical. Much better to use a minority of static defence to hold certain key chokes and mobile units to then push out and capture territory or respond to any threats. The line drag formation command really changes everything and when you get pressured too much the role of statics should be to defend against minor leaks and allow a logical place of retreat for your mobile forces.

#

The openness and access to metal in all directions is one of the reasons CCR is made for BARB: being a computer it can focus on multiple expansion directions at once and it can use its scouting ability to find leaks and move everything around without feeling overwhelmed. It's very common to feel like BARB is almost impossibly good at such and no human can compete fairly, but hopefully my vid can show that even without explicit cheesing or optimisation, it is possible for a human to actually be better at map coverage and manouvering than BARB. You just can't let them expand and capture more resources and territory than you, even the best players can then fall to BARB if it gets enough of a lead.

pearl hare
cold briar
# pearl hare The ai does the same thing every time per patch with remarkable consistency, the...

We had a lot of recent engine reworking and AI fixes that significantly changed AI behaviours for a brief period of time. Not to take anything away from r77feed, because any win over 3 barb on CCR is pretty good, but even some vet videos of wins against 3 barb have to be taken in context: different balance and different AI lead to videos that don't work any more or worked in a an extreme minority of cases or involved behaviour that the AI no longer displays.

pearl hare
# cold briar We had a lot of recent engine reworking and AI fixes that significantly changed ...

That video is very recent and is about the same time I tried it (couple of weeks ago). I don't replay scenarios often, but I do play a bunch of coop pretty much every week, so I have a decent idea of how the AI changes per patch. The AI in r77feed's game isn't broken in the way that recent changes often break it. Really the only inexplicable parts are the missing t1 army and the complete lack of nukes (which would've hit anti anyway, but still strange)

cold briar
#

And there's a bunch of older 1 v 3 very videos of vets being able to rush out and get the the commander snipes before being overwhelmed, but then a recent change made BARB more likely to retreat commanders or two to the safety of their bases rather than push out with all three. I think I saw another one recently where someone managed to go to tech 2 vehicles and push vehicle artillery forward into barb, which was new and I'm not sure it works any more...

#

I think my favourite one I've seen was a cat and mouse transport commander dgun battle that goes on and on until the player finally wins.

#

I've got a 2 hour slog lost to the annals of history that shows it's possible to win against 3 playing defensively, wherepossible is doing a lot of the heavy lifting if you ignore the 76 failed attempts.

#

And yeah, have the nukes stopped?

#

Because that was one of my constant struggles in my defensive proof of concept

tardy cliff
#

I played a big map 24x24 with my buddy against two normal ai's , it gave us a lot of time to expand what CCR doesn't give you and we had a one hour and a half game, eventually we won having faced a ton of juggernauts , and using nukes ourselfes. However I wasn't sure if the ai build any anti nukes at all. Isn't there anything in the ai that makes it build an anti nuke asap when it's available to him ?

fallen oxide
#

I finished the same campaign against brutal AI 1 year ago, the AI definitely improved. It used to have a lot of exploitable bugs, especially regarding pathfinding and army control.
It would stay on t1 for too long, mass t1 units, bunch up in balls and when he faced defenses around your base, it would make a giant push at the edge of the map. 
Units would get stuck leaving factories or trying to fight for position, on Supremacy it spammed amphibious vehicles and they all got stuck transitioning from water to land.

The AI in the scenarios has global vision, meaning that it instantly counters whatever units you are making. In skirmish maps you can add a barbarian AI, enable os vision and modify it's difficulty to make him more adaptable, maybe that's the answer to the lack of antinuke.
Current AI still masses t1 units if it senses that a push is in his favor. Try making ticks for 8 mins and you'll have 40 t1 units knocking at your doorstep.

In my video I destroyed some of its army, had 20 grunts, llts and made tanks at 8 min which prompted all 3 bots to go t2. This is why I had to stop bulls, tigers, tzars 1 min later with shuris and dgun. The bottom bot had t2 bot lab when I attacked him. The AI will react to whatever army composition you make, that's why you get different results than mine.

bronze marten
#

Yea when I won my 1v3 brutal ai ccr, I had to comsnipe. AFAIK, randy is the only one to 1v3 brutal with a defense game plan. But he tried over 20 times

pearl hare
# fallen oxide I finished the same campaign against brutal AI 1 year ago, the AI definitely imp...

The usual t1 blob is going to absolutely roll over 20 grunts, so it's interesting to see them opt out of it if they see the grunts. I don't recall their usual t2 timing, but it didn't seem to early for them? In my game I made a mace-resbot push to try to grind them down with efficiency, and that didn't stop the t1 blob. Or maybe that's precisely why they built it, because of me being too aggressive?

rancid ledge
# bronze marten Yea when I won my 1v3 brutal ai ccr, I had to comsnipe. AFAIK, randy is the only...

I would very much like to see these replays. All I have found are exploiting either walls or jammers which AI cannot handle. I do remember seeing someone do it (on normal, I think) with the old blitzes and raiding a lot, but in that replay the AI coms walked out and could be rushed in the middle of the field w/o assaulting the base. I tried it on normal yesterday and all the AI coms were hiding in bases (I had to use jammers to win), that is a bit of a problem because BarB likes to spread it's units around the map, so the resistance doesn't end after frontline (unlike most humans). How did you snipe coms, with air?

silk zealot
bronze marten
#

#💾｜bar-replays message here is mine

#

Just over a year ago now

rancid ledge
#

Thanks!

tardy cliff
modest karma
#

Oh wow

#

I actually have quite a lot of experience, I put a lot of hours into that particular scenario when my internet was down for a few days

#

I played on normal difficulty but was able to devise a tactic that gave me a ~75% win rate

#

Practically, the AI doesn't play like a human does

#

As its already been mentioned, the AI will constantly test you for gaps, its important to shit LLT's all over the place

#

If there is a gap, the enemy will abuse it and you'll be stuck trying to patch them and repair

#

The AI, however, fails to adapt properly to being out-ranged

#

At least as of 6 months ago

#

My tactic was to go core, first expand with plenty of LLT leaving no gaps or ensuring I had units in all gaps. (starting in the corner)

In T1 build up a raid force and take one of the 3 barbs bases, specifically targeting energy economy when I could.

I knew I couldn't match the 3 AI's for map control and metal, so by wrecking their energy economy I could make them less able to effectively use their metal advantage

#

I'd start making core asolar right after my raid force ended, so by the time I was done microing my raid on their energy economy I could T2

#

At this point, I'm somewhat relying on the AI's unwillingness to charge into LLT's, even if it had enough units to take them

I presume the AI calculates that if it charged a line of LLT's, it wouldn't be very efficient metal wise, so the AI is rather unwilling to attack LLT's until it has T2 units

#

The AI also primarily builds raiding units, such as T1 tanks and T1 raid bots, almost never building long ranged T1 such as rocket bots or missile trucks

#

Practically, even though the AI could probably kill you if it ordered all of its units to attack while your T2'ing, it never does

#

Core T2 has the unit which is critical to this strategy, spam morty as hard as possible

#

Morty cost relatively little energy, which is great for a map abundant in metal but low on energy such as comet

#

The AI derp out when encountering units that outrange them and fail to attack the mortyball with raid units effectively

#

A few roach and fiend held nearby and used on anything that gets too close with deter the AI from trying to attack the mortyball with its T1 spam it built up

#

Again, the AI could probably kill it, even if it wouldn't be efficient, which would buy it time to build more energy economy to utilize its greater metal income, but the AI cant make that decision

#

As such with your mortyball and a few defending units, you should be able to slowly stomp all 3 enemy bases

#

The AI will keep its units moving in and out of the maximum range of your morty, constantly poking for an opening, but provided you dont give them one by utilizing roach, fiend, or spy bot stuns, they wont react properly

#

A real player could split their units to attack the ball from all sides too, but the AI will try to concentrate its attacks, leaving it vulnerable to spy stuns or roaches

#

Anyways, TLDR: The AI reacts very poorly to LLT or long range units, if you are able to get T2 first by taking out the enemy energy economy, you can roll them with a mortyball

#

tested on hard and normal, I never tried brutal as once I beat them on hard my internet came back

#

Generally, I always tried to end the game before the AI could get a fusion reactor down, if they got one of those I found it was almost always over for me as they could spam T2 raid tanks which were not afraid of LLT whatsoever, so they'd constantly test and break through my flank defenses

#

I also did try to kill the AI's by comsniping early game, but after the commander update thats a lot harder

#

In the end, I do think the scenario makes for an interesting challenge, however it may not be the best practice for pvp games as the real challenge I find is developing a strategy that exploits the AI's miscalculations

tardy cliff
#

I'm currently trying vs 1 easy and 1 normal, I can hold out pretty well till around the 40 min mark then usually I'm focussing fighting of one wave with my units and somehow the ai find leaks all over the place somewhere else and trash my flanks. I did get to T2 in two games and was able to hold him of and expand pretty far in the map, but then I left my com unguarded for five seconds and bang game over ...

modest karma
#

If I dont win early, I dont win at all

#

I'd say AI's have near perfect micro in certain areas, such as constantly testing you for gaps as well as expanding quickly

#

But stupid overall decision making

tardy cliff
#

it's impossible to defend two lines

#

all I'm doing that game is send my army down to defend the lower half then race back to defend the upper half meanwhile my mex's get sniped left right and center and I can't rebuild cause I'm moving my army all the fucking time

modest karma
#

I'll take a look

#

3:00
You can use pawns to guard cons with the guard command, so the pawns follow the con and are quicker to react

#

8:00
I'd avoid T1 vehicles in this situation, scaling with asolar would probably be better so you could afford more kbot and once you get enough energy production you could T2

#

You spend a very long time not scaling energy economy at all in the midgame leading to an enerergy stall, you are unable to spend your metal despite the bar being full

#

You dont expand, but thats fine, I want to make a point of saying that you were fine to not expand anymore around the 6-8 minute mark, your energy is your limiting resource

#

expansion gets you metal, which you are full of

#

12:00
You attempt to rectify your energy stall, but to some degree its too late, you should have started 4 minutes ago

You were energy stalling from 8:00 to 12:00 and that significantly limited your ability to build units or further economy

#

At 13:30 you stop building energy and use com for defenses, using your T1 solar to now produce T1 units in large quantities

Your metal and energy are both now satisfied but you are lacking build power

#

I know this is painful, but nanos should have already been made as a midpoint to scaling your energy economy

being reactive is too slow in a situation like this where you need to outproduce 2 enemies

#

I'd also suggest rez bots for reclaiming, there's plenty of dead tanks in your base, by 16:30 your metal stalling on your T2 which that could solve

#

Anyways, I think the biggest thing you need is practicing resource utilization

Energy scaling on a map like this should start early, its one of the biggest advanages you have over the AI on a map like comet, the AI is slow to scale its energy economy in T1

#

More energy scaling and T2'ing earlier in my opinion

tardy cliff
#

I'll try next game now I tried me + 1 easy vs normal and easy but got outproduced on units again even with an ai help I have to constantly defend two fronts it's infuriating

#

and why are there static defensess in the game ... they are except in the very early 10 minutes USELESS, they get sniped out by long range the second you place them, it ried perimiter cameras to give them line of sight but what's the use one T2 missle unit snipes each one , one by one

#

I had the warning metal and energy overflowing multiple times but couldn't build more factories or anything cause I was just moving my army up and down

modest karma
#

Watching

#

As a starting note, yes trying to micro a 1v2 or 1v3 is somewhat painful, I know it is quite taxing as a player

#

However it is great practice to increase your management potential for actual pvp games

#

I'd advise trying to learn overall focus management from these games, as in managing you as the player focusing on what needs done, such as managing eco and units at the same time

#

Just dont try and learn pvp tactics from AI games, the AI plays completely different to real players, howevever that being said the AI is still quite a challenge, just in its own way

#

4:30, a radar would have allowed you to react slightly sooner, you were able to defend the attack without loosing any economy, which was good

#

But to note, a radar allows you to react faster and provides more information for less stress on you as a player

#

even if unecessary for defending, it can reduce the stress of you the player

#

I wouldn't recommend building beholders, especially on the backline and especially early game

they eat a constant 13e/tick, which is more than half of what a basic solar produces
on a map like comet where energy is typically the limiting factor, especially early game, it is significant

#

Plus the 2 more in front, eating 52 or 2.5 solars worth of energy

image.png
#

Use radars instead, T1 radars have 680 LOS to a beholders 560, plus the massive benefit of having radar

#

To admit something, beholders are pretty useless and almost a bait unit, you should almost never use them

#

They can provide value if you expand in T1 but are getting pushed back so plop a few down to provide vision of the enemy's frontline as they expand into you

#

8:30
You queue a nano to be made by a conbot by your kbot lab

I'd suggest building an asolar instead, even better would be more basic solar, then an asolar, then a nano

image.png
#

I dont mean to be an ass, but in a way this is partly the same problem as before, around the 8 minute mark you hit an energy stall as you did not build past your initial solars, despite having plentiful metal available

#

Your 5 beholders are also making things significantly worse

#

T1 radars dont consume any passive energy, for another note

#

By 12 minutes you've corrected the energy stall and have asolars up
Ideally you can plan for the 8 minute energy stall in advance and build more of it early so you can keep your resource utilizaton high

#

Perhaps even, if you want to practice quickly with new builds, run in and play out only the first 12 minutes, just practicing to avoid an energy stall and maintain high resource utilization

#

Honestly, I think you would have done massively better if you built 2 radars, 1 in front and 1 in back, instead of the 5 beholders, it would have significantly improved your energy economy

#

A quick fix which I think would massively improve your play

#

Still, I will note your improvement over last game, you solved the energy problem by 12 minutes and it remained solved

#

Going right into T2 con veh to build mex is good too

#

Overwatch are not an upgraded version of the LLT, overwatch are specifically designed to outrange rocket bots and missile vehicles to protect other towers from them

#

Overwatch have atrocious dps compared to their cosyt

#

You got screwed by unit specifics for a second time

image.png
#

Use beamers instead, beamers have increased dps for less than half the cost, with a bit less range

#

Alternatively, being T2, T2 popup towers like pitbulls are highly effective against anything that isn't a T2 siege unit

image.png
#

Speaking from experience of pvp games, T2 popup towers are very resource effective when fighting enemies who do not respect them

#

T2 popup towers can be outranged and sieged by veh arty, morty or sniper, but that at least forces the enemy to slow down and siege

If the enemy tries to run units past them, they're going to suffer significant damage to their forces

Well, excluding tiger tanks, tigers are special though being the most stats wise efficient assault unit in all of T2, T2 popups go about even resource wise with tigers

tardy cliff
#

thnx I'll take that with me in my next games

modest karma
#

Beholders and overwatch both suck ass

tardy cliff
#

I'll just keep one T1 bot in the back line and queue an infinte of solars with it , it will slowly build up energy over time I guess

modest karma
#

Thats a very good idea on this map

#

Personally, I do it a bit different

#

I'll get a replay of me scaling my eco on this

tardy cliff
#

I do have a little bit of past knowledge but BAR is a bit more unforgiving

modest karma
#

Yeah

#

GAAT got shit on in BAR, the ""upgraded"" laser tower actually got split into 2 units

#

Overwatch for range, beamer for dps

#

beamer still outranges rocket bots barely, so its actually pretty good

modest karma
#

I lost versus 3 hard, but I hope you can see a bit from the early game

#

I got wrecked by the enemy air ai bombing my metal eco to shit

#

I think my eco scaling was still decent, but I did fumble the later parts of it and my reaction to the enemy air

#

I must say that I have a particular hatred for the AI when it plays air, even if it kills nothing it damages your efficiency by forcing you to build aa wherever you expand

#

I should have responded with an early rush, but eh, I'm out of practice with the scenario

#

If the AI goes air, hit them early, especially the air players base

#

I will also mention, good on you for keeping up practicing against the bots, if you really want to get good at BAR, the management pressure they put you under does help you play pvp games

#

Most pvp games average at a lower management intensity than AI, so playing pvp you rarely spend a lot of time practicing high intensity management

AI is almost exclusively high intensity management and is a great training tool for it

#

Real ascendancy in this game is when you realize your 1v2 but completely comfortable and relatively fine holding them off

#

Turn on your best music and grind those AI games till you can kick the shit out of those bots, might take quite a few hours but its rewarding and many of the skills are highly valuable for pvp

#

Better yet, you can instantly surrender and go next when against bots, if you realize you made a big fuckup

#

Helps for learning, no need to wait for lobbies and other players, ect

tardy cliff
modest karma
#

watching

tardy cliff
#

watching yours now

modest karma
#

Holy shit the AI builds nukes, I havent seen that before

image.png
#

Anyways, on a more serious note, great eco management in this one

#

I'd suggest possibly switching to asolar after making ~10 normal solar

#

You were somehwat low on metal for a few bits of it and asolar are significantly more metal efficient

#

Still, you did good

#

Built a fusion at a good time too

tardy cliff
#

thnx was a nice one, but yes ai nukes , I played on a big map with a buddy we had an 1h33 minute game

#

where in the end our anti nuke was launching non stop

#

eventually we won sniping the last commander but we had to stop like 15 juggernauts and 4 behemots

modest karma
#

Yeah

#

A note is that the AI does get constant improvements, so it can actually play better than I'm used to

#

It still does seem to avoid LLTs even if it could kill them, but it used veh arty on me earlier than expected

tardy cliff
#

yeah I noticed it going long range very early which is so annoying cause a radar gives you where they are but only with vision do your units attack

#

I put ticks in my build order just to have a bit of vision

modest karma
#

ticks are good for vision

tardy cliff
#

I couldn't find that popup tower though I looked in the T2 build menu couldn't see it

#

only the big laser tower

modest karma
#

I have it highlighted here

#

Note: I use grid menu

#

Note: You probably should too

#

Its a bit annoying to get used to, but it gives every single structure a letter key combination so no clicking menus

tardy cliff
#

ah k I'll have a look next game now I know how it looks like

modest karma
#

IIRC its a default widget you can just enable at will ingame

#

Otherwise, T2 popups should still be there, they just look weird since at least in arms case, the picture shows the popup tower stowed

#

Another note is arm T2 popups can cloak

#

Core's cant but cores have a bit better stats

#

Additionally popups gain damage reduction when stowed, making them take longer to kill with artillery

tardy cliff
#

well as long as the cloak engages and disengages on it's own with energy fluctuations, I rather not have a cloak drain my entire supply

modest karma
#

T2 laser towers can stow too, but they typically outrange artillery and can fight back so generally stowing them isn't much of an advantage

#

Oh yeah, the cloak isn't as useful on a map like this...

#

Its occasionally just neat to know when your playing a map

#

like glitters or such with more energy and are getting pounded by arty and want to make it harder for them

#

If the enemy spams arty with few frontline units, you can mix T2 popups with jammers, as both units can cloak and jammers hide the T2 popups from radar too

#

Rarely will you get a chance to use that, but its a funny trick to pull off when playing arm and when your enemy over-invests into arty

#

Set t2 popups to hold fire, let the arty get close, then wreck shit

tardy cliff
#

I see, well thnx for all the information but 2 30 atm here so time for sleep got work tomorrow

#

I'll play a few more vs easy and normal until I'm confortable then I go for two meds

modest karma
#

fair, see ya around, I'll look forward to you posting a replay of you beating them

modest karma
#

To note, since I last played the scenario the AI does seem to be somewhat better

#

It seems to focus on bombing your metal extractors constantly

#

Still, it reacts very poorly to longer ranged units or even defending itself in general

#

The AI will continue to try and attack your economy with its units while purely relying on its tower spam to defend its base

#

I'll note that even now I didn't play it particularly amazingly, I should have focused on raiding them in T1 to hopefully slow their energy scaling

tardy cliff
#

I just had a very long battle, I should have been more aggressive mid gam, he managed to get 4 berthas up and was anti nuking my nukes

#

couldn't get through and my base got systemetically bertha'ed ...

#

so I just left that game

tardy cliff
#

2 v 2 I'm able to manage now but 1v2 is still to much I can't keep up with both of them

modest karma
#

Yeah, 1v2 is very hard

#

From my experience, I try to find a way to win without map control in cases of 1v3

tardy cliff
pearl hare
# tardy cliff I had fun on this one but lost cause the ai got a juggernaut and behemoth out .....

If you want to actually win the race, you can expand forward a little earlier and rely on radar and moving your army to catch any raids. But that's not why you lost.

Two things:

  • Cut your t1 unit production when you have access to t2. They are extremely cost ineffective, you're sinking mad amounts of metal into nothing. You can even reclaim the t1 lab for that extra bit of metal. This alone will probably get you a win.
  • Reclaim. Make resurrection bots and either eat or resurrect all the wrecks on the battlefield. Try not to have meaningless fights on the enemy side of the map, because they get all the wreck metal instead. If you're going to hit them, make it count, or retreat.

Also, get an antinuke by like 20 minutes max, or you're rolling the dice on losing instantly

tardy cliff
#

thank I'll try in my next match, just played another I could hold out a hell of a lot longer but then 5 jugs and 2 behemots yeah ... game over

#

and he went all along the last pixel of the south border to finish it ....

#

thing is I do get a few good pushes out but I never manag to push through, by the time I reach his base most of my units are gone

pearl hare
tardy cliff
#

yeah hadn't read your note yet

#

I was playing

#

I'll just build a bunch of T1 con bots then reclaim it next match

#

I always seem to be without T1 bots when I need them

pearl hare
#

you can always plop down a new t1 lab if you need more bots

#

keeping a t1 bot lab around is honestly not even a mistake. It often pays for itself in resbots. Producing maces out of it is a different matter

#

I also don't recommend sprinters as a fighting unit. They're not made for that

tardy cliff
#

what is a good bread and butter unit for the T2 bot lab I see a lot of youtbes with the fatboy

#

but he builds slow and shoots slow I need an early blob defense

pearl hare
#

fatboy is an entire loaf of bread in itself

#

usually you build Welders if you want to fight cheap junk

#

Fatboy is amazing because it can trade very efficiently, you just keep blowing up stuff and repairing

#

A few sharpshooters are also great to kill some of those mammoths you're struggling with, but I don't recommend building them before your first fusion

#

Hounds are great skirmishers, but they need more attention than welders

tardy cliff
#

sofar I'm building fusion and two energy convertors for each fusion to keep my metal up

#

I'll try my early defense with the welders and add an extra fusion plant for the sharpshooters

pearl hare
tardy cliff
#

true but isn't it one fusion for 2 energy converters ?

pearl hare
#

building units have priority over conversion, unless you put your conversion slider at the top all the way down.

#

You're not going to struggle with energy because of converters

#

also don't be afraid to plop down 20 nanos at that stage. You have to be conservative with them early on. By the time you have 200 metal income per second, it's better to overbuild buildpower than to float 10k like you're doing

tardy cliff
#

ah k so just go overboard with nanos , i've always been careful with those sofar since I didn't want to drain my entire eco

#

I'll just plop them around my factories and try to bundle my factories around them

pearl hare
#

In the early game, they will drain your eco and you have to be careful. Towards the lategame, go wild.

#

I also strongly caution against these. They are amazing value if you build them in range of the enemy base. On fire at will shooting at random radar dots in the middle of the map, they're a waste of metal. Build more fusions instead

image.png
#

The "go wild with nanos" point also applies to your economy. You should be assisting your economy scaling right now.

tardy cliff
#

k so no berthas until the 20 nanos and build them in range of end of map

pearl hare
#

I'm not a TA player, but I think they are supposed to have longer range than the rapid-fire version in that game. It's not the case here. Usually you either know exactly what you're aiming at, or you skip the LRPCs and make units/economy. If it reaches a stalemate, you can then build the rapid-fire one to close out the game.

tardy cliff
#

need a little break now but I'll try all your suggestions next match

tardy cliff
#

his bertha's just whittled my base away

pearl hare
#

Alright, here's what I'm going to do. A lot of the improvements I'm going to suggest boil down to "just play better". I'm going to write these in cursive so that you know they are possible, but might demand more of you. Anything not in cursive is something you should be able to act on by just knowing it, instead of taxing your apm and the like.

First of all, you're too slow to expand. The AI is perfectly beatable this way, but ideally you'd be keeping on pace or ahead of the AI in metal extraction. Having every mex covered by 1 llt is enough, you don't need to build a forest. Catch bigger raids by moving your army to the radar dots.

If you're going to beat the AI while down in economy, you have to prepare accordingly. You will not be able to beat down their front door, and will have to outscale them using fusions and converters. Any favourable trades in the meantime are welcome. This means you do not need to build more army than it takes to defend. Either cut back on the amount of pit bulls you're producing, or the amount of t2. Invest into fusions instead. After the first few fusions, switch to Advanced Fusions. These are very explosive, so build them in the back corner.

The army control needs work. Have separate hotkeys for long-ranged and short-ranged units. Move them accordingly.

Alternatively, switch to single-unit compositions. Build only hounds, or only welders, you're having trouble managing both. Have sharpshooters behind the lines, but waaaay behind, not in the same blob. Sharpshooters don't have to be moved as often.

Don't stop scaling the economy, make sure you have fusions always queued towards the lategame. If you don't want to think about it, literally spend 30 seconds setting up a queue of afus-4converters-afus-4converters.... like 10 times.

Dont's neglect resbots towards the lategame, keep them active. If it's a good unit, it often makes more sense to resurrect it instead of eating it, especially if you're going to hit the overflow.

#

Or, and I don't believe I'm saying this, build a metal storage to eat as much as you like.
Switch to Titans earlier. If you're planning on sitting it out against juggernauts, you need their bulk.

Also, cut the t2 production at that point for the same reason you stopped producing t1

#

Also, Plasma Shields deflect LRPC shots. It doesn't make sense to spam your entire front with them, but protecting your key areas like buildpower is a must lategame

tardy cliff
#

I'll give it a shot, also I didn't knew plasma shields where a thing good to know, but the AFusion tip is hopefully going to help a lot as well. I'm also going to try gradually build out my defensive line to encompass more metal spots, might work or not

tardy cliff
pearl hare
#

I also don't recommend building that many t2 constructors for buildpower. Get 1 per task to place the buildings, and assist with Butlers from the same lab. Much more efficient. Or in a bind, making a few more t1 constructors will also work. T2 are expensive.

tardy cliff
#

I kind of wished the reclaim bots would start reclaiming after being build

#

I always forget to manually assign them an area to reclaim

rancid ledge
#

Use fight command for factory rally point. Rezzers will reclaim everything in their path. You can even make a long zig-zag.

tardy cliff
#

seems the AI is really dependant how well it plays on the map, I tried vs 1 hard on supreme isthmus and it was with all I learned sofar , just an easy walk

pearl hare
tardy cliff
#

true and the metal is relatively far apart of eachother compared to CCR

#

the fighting in the choke was pretty descent but once you break the choke I just wiped it's back lines with one push with about 8 T3 and a bunch of welders

tardy cliff
#

I tried that STAI or how is it called again, lost my first, but on my second I switched up and instead of pawns I only made the plasma bots and rushed to each metal spot I could find to pay for it. After beating back the initial plasma rush the AI send to me it was a piece of cake. Played differently than the BarbAI though

pearl hare
#

Wait, since when is STAI publicly available? :O

pearl hare
#

It did play a lot more like a real player, but a very bad real player. Fun

compact frigate
#

For what I have seen, STAI tends to crush/break multiplayer games (for some players). And are not stable enough jet.

tardy cliff
#

you can enable it in options with a checkbox , you get a few more ai's , think it's developer mode or so

#

I'll have to boot up bar to know the name of the checkbox