Gumiho had a better solution years ago. It’s medivac ghost drops. Instead of clicking all those changelings down you make it a priority to EMP them every 50 seconds. Batz trying to convince us all to become mech players. smh.
Mech is the solution rts’s problems. Siege and forget units allow you to simplify the map down to a MOBA style rts interaction by taking control of territories and then forgetting about them. This pushes the interactions into the disputed territories, aka the “middle lane” of the MOBA battle. The difference is that this is very strategical because you have to decide which positions to take, with what units, when to take them, and when to give them up.
The challenge is making zone-control powerful without creating deathballs. The solution to that is splash. If you put all your zone control into too tight of a space, splash annihilates it. SC2 designers didn’t like this interaction and called it a “game ending moment” when in reality it was a positioning mistake to focus too many resources in a single location. No more splash and now deathballs and f2 are the solution to the APM problem, which is where the strategy goes away.
Solving the apm issue via zone control and splash is a much better way than solving it through deathballing with f2. If I could have my way, javelin missiles would be removed and thors would do splash on its punisher cannons with the same radius as the javelin missiles. The disruptor would be an flying unit, no attacks until it seiges, and that sieges by landing on the ground. Once sieged, it has 14 range with a low base damage (1-2) and a huge bonus damage vs light (+30). The attack would be a catapult like projectile that fires at a location and not a unit. Why? Units could easily dodge UNLESS they are stationary aka not actively controlled. Bio could dodge it if managed, but would be annihilated if it were standing still to attack a cannon or whatever. So it rewards high skill terrans. It also rewards high skill protoss via focus fire. If you can predict where units will be, you can kill them as they move.
Void rays would also be reworked to have anti-air splash. Prismatic alignment would be an ability that does line-splash like the hellion attack, but on a cooldown like ravager bile. One void would do almost nothing, but 10 voids could one-shot a cluster of BCs. Again, the goal is to punish people who cluster too much. You can’t put 15 liberators into 1 spot and win. Zone control is for territorial control, not for deathballing.
1 Like
Sounds like there’s not enough skill expression in your ideas. I want micro to be the deciding factor in engagements.
You kidding me. When void rays can one shot a stack of air units with a splash ability, it will be as high micro as roach v roach wars are w/ bile. Because players control territory with siege units, they can put their focus onto the main attack channel where micro is important.
The prismatic alignment would also drop the void ray’s shields, by the way. So if they time it wrong, the voids die for free.
1 Like
I was talking about the automated disruptors and thors that have even more EZ attributes. The voidray idea would stop the phoenix vs phoenix wars in PvP.
Disruptor siege with no management would be shooting balls into nowhere. It’s a slow moving projectile. Units that stop to attack a building would be sitting ducks. But if properly microed, they would be useless. So what would happen is that toss would apply counter pressure, zerg or terran would stop managing their counter attack & the disruptors would clean it up. The exception being if there is a non light unit in the tank, like a thor, in which case the disruptors would struggle since they do 1 damage a shot.
This incentivizes players to attack. Protoss defense is best managed by attacking. This makes the game action packed.
1 Like
Zone control that translates into worker harassment would be interesting.
Disruptors would be useless vs mech, armored, or air. They’d be hyper vulnerable to liberators or BCs. No more bio drops or ling harass unless you have spare APM to micro it like crazy. Unloading a medivac on a single spot would be a sure way for it to die for free.
Obviously thor liberator would be used to do the primary pushes. Immortals beat thors but liberators beat immortals but void rays beat liberators but thors beat void rays. That would be the basic rock paper scissor dynamic.
Because both sides have vicious splash options, who wins comes down to tempo and positioning. A counter-attack with BCs could easily cause void rays to be out of position for a thor drop which the void rays are useless vs, for example. But if in position, prismatic one-shots the bcs and one-shots the medivac drops. The medivac drops would be interesting because they’d be useless if intercepted but overpowered if landed, because the splash of the thors would be very good vs a ball of void rays.
It becomes a game of somehow getting mech units in position to wreck mineral lines. It would probably involve mostly BC harass, but I’d buff marauders vs buildings so that four 3/3 marauders could one-shot pylons. Stalkers would be buffed vs buildings, too, but their ability to blink to the high ground would be locked behind an upgrade with a long research time.
There is simply too much emphasis on balling your army and not enough zone control. Fix those issues and I think an RTS game could do really well with the MOBA crowd. But it would have to be marketed as a different category, aka moba-adjacent rather than rts-adjacent. Don’t worry, you’ll see what I mean at some point.
One of the fun things about getting a video game off the ground is that you have to program bots to play it so that players have someone to queue against. Alphastar uses neural networks, but I’ve been a fan of markov chains which can be adapted for the same purpose. Basically you create a million chain links. The links are randomly assigned to 1-5 other links. The program starts at the first link. Variables are inputted to an n-polynomial where n is the number of game variables you want to train for. You calculate the x-intercept of the polynomial and that tells you which neighbor link to move to. At each chain link, event tags can be added that cause actions to be taken in the game, such as building a unit or using an ability. The program executes whatever tags are attached to the chain link it is currently at, inputs the game variables into the polynomial, computes the next chain to move to, then moves to it and repeats. So now you have a system of 1 million chain links that can be tuned on game parameters to create decision trees. When alphastar came out, I wanted to do this project but I knew I’d be wasting my time because I’d get a chance to do something similar in the future. One of the obvious benefits of this system is that it allows for recursion aka loops. Neural networks really struggle with loops and recursion. It also implements memory naturally, on its own, with no need to add it to the model. Memory is an emergent behavior for this system, which is interesting because it has to be manually implemented for neural networks. Memory happens via the structure of the chain links, so the only way you end up on a node with the tag “target fire probe with a marine” is by following down a tree of links that brought the game state to that point. So you use these bots to play vs people, and people play only vs bots for their first 20 games or so. After that, they are added to a ranked ladder where the bots are tuned to be silver and bronze league. So most human players end up being gold and above, boosting their ego. It’s very clever. The true magic is in how you tune the structure of the links as well as how you tune the individual link’s polynomial terms.
With this i agree.
I was always skeptical about his skills and never his fan. For many years i attributed his accomplishment mostly to the fact that he played zerg. The proof for that is at least 2 other zergs on a very similar level to Serral which also won many tournaments.
1 Like
Same here, I felt he was such a creep abuser (before the nerf), queen abuser (again before the nerf).
I never seen him do almost nothing special besides: scout, react, a-move, and repeat.
Plus his ZvZ was never top tier in my opinion, yes he was not bad, but not even to close to top tier.
I mean, you say this, but he sees things that almost nobody else sees either, either from the information he has, or the general lack of intel. His game-sense is pretty phenomenal. His scouting in particular is really good - better than literally any other player on the planet, I’d argue. Baiting the best players in the world out of position to fly slow overlords in to know exactly what they’re doing.
The constant rogue Infestor Fungals that pop up literally anywhere there is an engagement, is something that only Dark has really been able to replicate - and even then less successfully - and only Clem has really been able to deal with sometimes.
The fact that Serral has maintained his position as one of the most dominant Zerg players on the planet despite being in military service currently, is also pretty telling of his skill.
This is the crux of literally every player we see playing at the highest level - it’s exceedingly simplified for all of them no matter the race they play, but every single player must be able to do this. Serral does it better than everyone else.
3 Likes
These 3 things are important, but are not enough if you are T or P, even for top 5.
Things like multitasking, insane micro, and speed are also required if you want to be in top 5 of world or better with the other 2 races.
Although I agree with a lot of what you said.
1 Like
Well, yeah? That doesn’t change that he is the best at them. For all 3 races, they’re not the only things that are important; Serral - and Zerg - are no different either.
Every race prioritizes certain aspects, sure. But no one thing is nonexistant in any of the races, not even protoss, who players love to meme on for being easy and a-move-y or whatever.
Positioning, engagement planning, timings, game knowledge, micro, speed; these are just a few of the frankly numerous things that every race requires.
3 Likes
I get it, you are a Serral fan, and understand how you feel.
I respect that you are not biased and actually read and understand other players experience and thoughts.
For me Serral is very good, however for me he is not the best, I may be wrong, however you know the GOAT discussions will never have an end or consensus.
Frankly speaking, I’m not, though I’ll freely admit to being astounded by his 2018 run. Personally I much prefer Clem.
True.
3 Likes
Me too, brother! Go Clement!
I would argue that his greatest skill is abusing queens and zerg defenses against harassment. For years nobody and i mean nobody could hurt Serral’s in the first 6 minutes of the game and then if his eco is untouched it is not uncommon for him (and to be frank any other zerg) to transition into late game with thousands of mins and gas and then good luck killing zerg brood\infestor\corruptor\ultra ball.
3 Likes
What is weird for that era, it was normal for Zerg to have a significant bank like 4k 4k Zerg and against no bank Terran or Toss, and sometimes not even max out.
How was that balanced?
1 Like
Agreed. And most people did not see this before.
The best zerg was and always will be startale life. During the period of zerg being somewhat underpowered in heart of the swarm vs mech and extremely underpowered vs Protoss, he took zergs play to the next level and continued to win several top end tournaments.
1 Like
Zerg needs solid bank to replenish their army but not to the point they have like 10k min/7k gas because opponent could not kill a single drone. That was never addressed by team which controls balance.
Typical TvZ looks like this:
- terran opens hellion/banshee/lib/BC
- terran cannot do any meaningful damage to zerg eco, thus he is limited only to clearing creep
- zerg transitions into late game with giant bank, allowing them instant remax on whatever unit comp they want.
- terran knows he has no chance to directly engage zerg army, because even if terran wins the battle, zerg reinforcement will crush remainder of terran army.
- terran camps in his own base behing PF, tanks, ghosts and libs hoping zerg will keep bashing against it and eventually run out of money.
1 Like