Figuring this strategy out was sick :D Support me : / uthermal Coaching : metafy.gg/@uth... My Twitch : / uthermalsc2 My Twitter : / uthermal My Discord : / discord Clips : / uthermalclips
@@bcsa8079 There are several maps for 16 players. Non have been in the ladder rotation but a lot of 16 player free-for-alls have been part of community events throughout the history of the game.
A note about the SCVs: the AI doesn't pull SCVs to get rid of them, they just go to your mineral patches if they are mined out themselves. This is because they have full vision of the map. The upside is that you know that at least some of the AIs are out of income, the downside is that they usually have a big bank. But they will actually replace the harvesters they lose.
I definitely think armor upgrades are worth it. Not necessarily because they will prevent your units from dying, but because they will end up more than paying for themselves by reducing how much repairing you have to do. If your units take less damage you spend less money on repairs.
if you're correct about how the AI determines whether to attack based on map-hack + army value, that might be an additional reason to keep your army supply as low as possible to incentivize attacks. Perhaps if you're at 200 supply the AIs will attack more intermittently with larger waves. I wonder if it's army value or just strict supply. Love your videos! This brings me back to my childhood playing 1v7c on BGH, rolling the dice hoping i get a "dumb zerg" that just builds two overlords and chills.
At first it is based on timings depending on their build orders, the build orders themselves depending on positioning of your units/buildings and what they scout. I am pretty sure they always know what you have resource wise but that doesn't give them full information so they make choices depending on what they see building. Outside of specific timings they will attack if threatened/trying to punish you or needing to defend their base or base trade even. Upon max supply they tend to be much more aggressive.
This is one of the most epic Starcraft videos I've ever watched. Incredible! It's like watching two heavyweight generals duking it out in an epic, all-out war.
The decision to keep your army slim is so smart. I could see this principle applied in late game pro play. Not an overall army count of course, but the army count required to defend some positions.
It actually is already applicable for overall army count in pro play. In grind it out macro games, where tempo doesn't win and it's going to come down to mining out, pro players tend to mass static D so that single quick game winning pushes are unlikely. And purposely not max out so they have flexibility to add the right units when situations change. And of course, refrain from committing fully to battles.
@@panner11 yeah the flexibility is a good point. Now that I think about it, something that pro protosses often do is to keep some army supply available for warp-in so they can reinforce some positions late game (against doom drop for example), or harass with the warp prism. But I never thought about this concept for terran.
Incredible stuff! A minor point - an unsieged tank actually outranges a PF (without the range upgrade) and does more single target damage than one in siege mode. So if you just want to gun down the PF its better to leave your tanks unsieged. Something else I thought of. I wonder if the other AIs would defend your neighbour if you reduced them to one building but refused to accept their surrender.
That's actually a really smart idea, if not accepting surrender still forced the enemy to try to help, you could get 4 bases of mining while also hopefully breaking up the enemy army.
Congrats! Vs protoss u might wanna consider adding like 4 5 ghost to emp. Its gonna be hard to keep them alive but maybe saving them with medivacs is an option since ais aren't the smartest. (What I know from the campaign ai is that they like to target healers so maybe disabling the healing ability for the medivacs works?)
I think immortals might be the biggedt issue due to the sheer bulk, so some emps might work, but I feel like interference matrix on some high value stuff might help more.
STORYTIME! rts games were my first computer games i played when i was like 5 and 6. although i played mostly warcraft and command and conquer. i never played against players.i was "too young" to play online lol.its really cool to see these challenges, it was mainly all i did when i was young. of course not at this level but its really cool to see you figuring out ways to beat them. thanks for all the effort you put into making your videos. cheers
This is sooooooo awesome! just started watching your challenges roughly 2mos and made me want to play starcraft again! Looking forward to the 7AI protoss match! Keep em coming!
Regarding the units lost: you had a ratio of 441,109 : 17.484 wich is about 25.23 : 1!!!! That is, to me, surprisingly better then the vs Z ratio at 16.85 : 1 (that is 150%), but when you think about it the AI mined more in total and with the AI behind your choke still alive they came in smaller waves wich makes you lose less. On average you had a 3.6 : 1 Ratio per AI and 5.36 : 1 against the AI that lost the most. I think it is also cool that you can clearly see at 44:06 wich AI had a base mor than the others in resources lost. It is great to see you still imporving and bringing great content to us! Keep it up!
Congratulations!! and thanks for making such an effort to produce entertaining content, I thought it would be months before this came out!! P.S. PiG said you didn't complete the challenge because it wasn't random but maybe if 7 Protos proves too much of a headashe for now then it might actually be easier to do random so there are loss protos?
PiG also said he didn't complete the challenge because there was a time limit. The deadline was several months ago, during the subathon which introduced the challenge and the bounty for completion.
Hi Marc, awesome video as always. One thing I guess would be very interesting is if you would make a video with your first attempts trying the challenge. Showing the mistakes, the things you discover while you are discovering them, similar to the GM challenges.
Amazing video. It's so fun to watch this haha, can't wait for the Protoss version! I wonder though, if you got +3 Armor you would take less damage, and need to repair less, in addition to having tougher units.
For the protoss, did you try ghost emp in combination with the wall ? + Raven armor shred ? Otherwise you can try nukes ! Against 6 armies in a single choke point, they cannot fly away that fast, your nuke will necessarily land on some %age of their combined armies (which would be cost efficient). You just need to deal with their ways of detection. That would be very cool to watch ! If you circle them with nukes (ghosts in your walls + sneaky ghosts outside coming from behind your ennemies when they go torward the choke point) the IA will panick and not know where to go. Easy to say hard to do but well.. it could be cool to watch ! I tried this against 7 hard IA and those traps were very cool to setup ! I don't have the skill to try it against higher difficulty IA tho. (I know the ennemies have vision on your level of difficulty but you can still trap them !)
Cool! Great job! 👍 Strange to watch a video without Marc's face, hope we see you next time! And yes, interesting to see same challenge VS Protoss and VS Random! If VS Protoss it is not possible - should we complain on imbalance? )) But I very much believe you can do it!
I am so looking forward to "against 7 insane protoss" if that is even possible! The only possible answer is luring them in, that might be attempting a wall off that you abandon, or something else! Give it a try and see if something works. Regardless of if you win or lose i would love to see a video!
Wow. This is really great. The commentary was as good as the game. So much explanation and detail. I loved it. Looking forward to the Protoss challenge and if it's impossible a video with one of your attempts. Would really like to see it. Keep up the great work. Thank you for the content.
@uthermal, how do you think, would it be more interesting if there were endless resources on bases? So you cannot just wait until bots run out of resources, but have to attack them. On the other side - you can invest as much as you need.
congratz, awesome to see how you figure it out and still hav the nerves to try protoss further =) , best of luck and i enjoy your channel a lot =) so thanks, for me you da kindest, sweetest SC 2 youtuber =) love it that u keep going here and raise you subs so fast , grateful greetings
Awesome win. Even if the Protoss turns out to be impossible, do you think you’ll post any of the attempts? I’m pretty sure if it’s as hard as you are saying, it’s going to take the AI derping a lot like when Heromarine played the 7 random AI and won. He was getting hundreds of kills when the Protoss was just running in circles at the bottom of ramps and stuff for no reason and the only Terran AI didn’t even make an appearance until the game was almost over so that was very helpful as well, not that it was Big Gabe’s fault lol. So I think for the Protoss challenge you’ll just need a little luck or to find some kind of bug but it would still be fun to see some of your closest attempts. Thank you for all the great content Sir!
That was beautiful! Really awesome to see someone do the ""ïmpossible". Thermy is that guy who did the 4 min mile and then people around the world started doing it.
@uThermal u mad man :D yesterday on your stream u toldme to check your yt today and here is the next crazy vs 7 insane game :D congratz ! Very nice play ! And for the 7 Protoss : - Consider Ghosts , it would be hard to keep them alife and it cant be to many, but like 4-6 ghosts that when big chunks of toss try to enter the choke can drop an EMP should massivly impact. - Raven may be another thing - same as with the ghost not to many just maybe 2-4 - if you can make good use of Interference Matrix on immortals (which should be the biggest problem in that describe crushing through you situations) - Since protoss will impact with chargelots, the liberator/tank might do alot of damage but i feel like some of the chargelots will make it through what finally will make your tanks shoot each other. So maybe a small amount of Hellbat in front could change that. They are repair-able, and if you can spend the little gas for blue flame they might do a good job I actually would be most worried about tempest + high templare in a big mix. Even if in theory you can rip the tempest using vicings, if the protoss randomly storms your vicing stack it might end it very fast. And tempest will otherway hardcounter your sieged position. Gl anyway and looking forward to the Video :)!
Greetings @uThermal, I have completed all three challenges and let me just say, the key to success against protoss can be found in the phrase "nuclear launch detected". But seriously godspeed, this is such amazing content and I can't wait to see what kind of delicious exploit you are able to cook up to combat the dreaded 1k supply a move protoss army.
Extraordinary skills! Amazing that you found a way to get them to feed themselves slowly through a choke point. I admit I still don't understand why the AI doesn't just send huge armies with diversified units.
That was awesome 😎, i know it's probably been said by others in the comments but i would not misuderestimates the power of the turrets upgrade AS WELL AS / AND STRUCTURE UPGRADE! You give you an incredible boost, if you would open early ebay.... with on all the work have done the turrets in this game it would give you such and edge also considerate that Ai's doesn't upgrade much😎😎🚀 GG! WAS an insane game!👻💪
I wonder if you can leave a very small pathway with the liberators at your entrance, so the enemy thinks they can go through but they'll die to the tanks instead.
You might be able to do the protoss run if you micro'd ghost EMPs as well as you did the vikings here. no shields and energy in the middle of their a-move would blunt the worst. Absolutely fantastic work. your videos are excellent!
Wow! This challenge turns out to require some counter-intuitive brilliance -- the kind only UThermal can come up with: don't upgrade and don't max out. Win. Crazy.
Good job! By the way, I think a siege comp to GM would be fun! (Tanks and liberators only.) Don't know if it would be possible, but it would be interesting!
I'm curious, around the 30 min mark, you still had 1300 gas and around 2k minerals. What if you made just one battle cruiser and flew it over to your neighbors base to do a little damage and jump back? It would most likely draw the A.I. back to defend and cost minerals in repair. The only problem with that I can think of is the stress of your multitasking.
Repair is actually quite expensive: you have to pay a quarter of the build cost of the unit. SO even when repairing, you must trade really effectively. Battlecruisers, alone, don't really. They can TRADE, but trade profitably? A 0/0 battlecruiser against 3/3 marines is going to REALLY struggle, and what else are you hoping to trade with? a viking ball? even liberators would mess you up. Really, it's also just not needed. He is already killing huge amounts of enemies in the choke, and he doesn't need to deny mining, as he has no plans to close the game before the enemies mine out. It's also a ton of micro.
I love these videos its very satisfying to watch entire armies get shredded. I wanna see you do it with zerg so that we get a proper lurker blender at the choke.
I know this video is few days old now and it was excellent quality. You asked for any ideas that might help my only thought was using ravens with thier deploy turret to help bait the AI into your choke death zone so you dont loose any units next time.
I feel like because the majority of protoss units have so much more health than their zerg or Terran counterparts, attack upgrades for seige tanks and libs would actually be much more important.
Hey uthermal, one thing I noticed, when you were attacking to eliminate your opponents, you did put your tanks on siege position, which caused splash damage to your own units, maybe to risky even if it was clear you had won. And I also think this could cause problems for the 7 Protoss challenge, since their units are so "tanky" and manage to get close. Maybe the best strategy would be no tanks, only planetary, liberators and Valkyries, and build reactors for the Starports with your barracks, so you can spare resources. And instead of marines in the beginning, maybe some reapers, since they have more damage against basic units (marines/zealots/zerglings) and workers and also regenerate life and also have the grenade which could be useful to defend? And the last idea I have is, since we are looking for efficiency, maybe some ravens to deactivate carriers, colossus and tempest and auto-turrets to defend/harass and finally kill your opponents in the end?
Brilliant game! Question: Could it be interesting to scan the mineral lines of your opponent, and drop a bunch of mules in order to weaken their economy?
Yes, I absolutely would like to see more content like this. I just wonder what might be next, now that all the races are done. 7 cheater AIs all set to rush? Seems impossible, but so did this one. ;)
It may be an exploit but from what I've seen the Ai looks like it stops producing workers when you finish all the resources. Which in turn means they can create bigger armies. So maybe just keep 1 mineral patch on the map so they keep producing workers
The reason the A.I. keeps an army in their base is it is waiting for SOMETHING to change so they can be effective. As soon as the choke is gone and you move out it registers that your choke is empty and it tries assaulting again. Until then it'll keep more than enough to wipe you hidden where you have to move to get to it. It has Patience. And long-term planning.
Used to do this in StarCraft 1. But I used two cheat codes… operation claw and show me the money. I don’t remember the map name, but it was an 8 player map with one choke for each base. I basically did the same thing, just lined the tanks up and had Valkyries for aa, but I also used cheats lol. Idk, you should experiment into an island maps versus insane AI. I feel like island maps are an untapped way of playing. Because it requires you to play so much differently then the traditional ground connected playstyle that’s always shown and advertised for matches. I’d be very much interested in seeing an island map play through. I feel like they were so much more popular in StarCraft 1 and Warcraft 2,3.
Beautifully executed. It's wild that they just completely ignore your bases because your army is on the low ground. The AI could definitely use more advanced harrasment behavior.
I think versus protoss you probably have to have 4 or 5 ghosts to emp clumped armies trying to a-move through you to make it holdable, but I’ve never tried the challenge so who knows. Seems like removing thousands of hp in shields with some emp’s could be good though.
EMP shot is probably a necessity against protoss, it cripples archs/immos/HTs, but it's probably a nightmare to micro. Because of the sheer variety of army outputs they'll presumably send, it might not actually be possible without forcing builds or a very lucky match. Observers, carriers, voids, collossi - they have so many ways to retaliate against the high ground ghosts with EMP would love to abuse. They easily counter libs, often counter harass mineral lines with phoenix, destroy bio with aoe... Honestly, victory over the protoss probably depends very heavily on what they build as a determining factor in your own build, and is going to require heavy simultaneous micro on multiple different small unit groups. But, aside from ghosts, probably don't build bio and you'll probably need all the mech
@@kozatas so how would that comp work? You're going to need A LOT of emps but too many ghosts and you won't have enough vikings for anti air. You could sacrifice libs instead but then you have less ability to capitalize on the lack of shields. Also the problem with introducing bio is you have to make medevacs to heal or just bite the bullet every time a ghost dies.
The only thing I could even fathom to improve this is a single dropship. There are times that you could have utilized it's ability to move SCV, Thors. That said, every 1% a dropship heals an SCV will be a step to compensate the 100 minerals it took to create the dropship. As well as the situations where you want to move a tank or destroy a turret. There is the fact that an SCV used to repair a dropship from inside. I don't know if that's the case anymore (could have been patched out) but still an interesting idea to consider.
A Theory I have regarding the armor on building/units is something to consider as an economic choice. As the repair of a unit is 25% of the original cost, saving 2 damage per hit, over the course of the game would have a chance to save a significant amount of % health. That would translate to the cost of repairs over time. That means that at least the first and second armor increase, if early enough, could easily pay for itself over the course of the game. The question is if there's an opportunity to gain it early enough to not mess with the tempo of your strategy needs. While in 1v1, the ideal of 1 armor preventing death makes sense, I think researching the amount of damage a unit can take per % of repair cost could help justify at least 1 armor upgrade.
So I decided to run some numbers on this for the first armor upgrade: Vikings cost 150 minerals and 75 gas and have 135 hp, so after 720 repaired hits against vikings you will have saved 200 minerals and 100 gas, paying for the first armor upgrade Liberators cost 150 minerals and 150 gas and have 180 hp, so after 480 repaired hits against liberators you will have saved 100 minerals and 100 gas, paying for the first armor upgrade Siege tanks cost 150 minerals and 125 gas and have 175 hp, so after 560 repaired hits against tanks you will have saved 120 minerals and 100 gas, paying for the first armor upgrade Thors cost 300 minerals and 200 gas and have 400 hp, so after 800 repaired hits against thors you will have saved 150 minerals and 100 gas, paying for the first armor upgrade Note that this assumes the AI does the same thing against you in either scenario; if the AI responds to you making armor upgrades by making more weapon upgrades (I have no idea whether that's the case, but it is strange that in this game the AI chose to stop at 1/1), then it's a net loss because weapon upgrades add *at least* one damage per hit, often more.
@@tanonevacm Thank you for running the numbers. It was a theory and I appreciate you taking the time for the rudimentary testing. Good point about the potential AI adjustments. I feel more now that this is a good concept to test. As a note, the numbers for SCV and Turrets are relevant, even if they aren't gas. Only to put into perspective the potential econ, that's all. Another thing to consider is that this math covers the entire army as a whole. Such as 200 repaired hits on one unit + 100 repaired hits on another unit + 50 hit on a third unit, etc. I do believe that (assuming AI adjustments are not responsive to it) at least a 1/1 upgrade is a potential opportunity for an economic move moreso than simply unit survivability.
