|
@OP
/signed
The pylon idea is great. Not just great, it's awesome.
The current high ground mechanics are aweful. I miss the old ones.
The worker part is the least important to me. Against melee units it will always be possible to withdraw damaged probes thanks to having nearby and easily clickable minerals. You're screwd versus mnm though.
|
The way I see it having the PvP really short and tense is just like ZvZ. It doesn't really matter because a similar percentage of ZvZ's get to high tech as PvP's do, and we're fine with that, so I don't see how it's a problem. PvP might also evolve into a rock-paper-scissors match up.
|
Hi guys, when I calculate the required amount of average hits I get a result of 3 shots.
Here is my calculation:
We want to calculate P(X >= 2). We need 2 or more hits. This is equivalent to 1- P(X < 2) = 1 - P(X = 0) - P(X = 1). This is an Bernoulli experiment. So using P(X) = Bp,n(X) with p = 0.5 and Bp,n(X) = (n over x) p^x * (1-p)^(n-x) :
1 - B0.5,n(0) - B0.5,n(1) which results in 1 - (n over 0) * 0,5^n - (n over 1) * 0.5^n.
Furthermore 1 - 1 * 0.5^n - n * 0.5^n = 1 - (1+n) * 0.5^n
Now this formular depends on the amount of shots we want to do in total. But we're interested in the average value, so we want this probability to exceed 0.5.
0.5 <= 1 - (1+n) * 0.5^n => (1+n) * 0.5^n <= 0.5. As n can only be an integer value, we find:
n = 0: 1 <= 0.5 wrong n = 1: 1 <= 0.5 wrong n = 2: 0.75 <= 0.5 wrong n = 3: 0.5 <= 0.5 correct
Therefore we have 50% probability to kill the rine in 3 shots, no?
|
About the PvP thing, does the rushing player typically have to use the Chrono-Boost on his Warpgates to succeed? If so, another viable solution would be to disable (or nerf) Chrono-Boost for Warpgates.
|
South Africa4316 Posts
On March 10 2010 19:49 Kickchon wrote: Hi guys, when I calculate the required amount of average hits I get a result of 3 shots.
Here is my calculation:
We want to calculate P(X >= 2). We need 2 or more hits. This is equivalent to 1- P(X < 2) = 1 - P(X = 0) - P(X = 1). This is an Bernoulli experiment. So using P(X) = Bp,n(X) with p = 0.5 and Bp,n(X) = (n over x) p^x * (1-p)^(n-x) :
1 - B0.5,n(0) - B0.5,n(1) which results in 1 - (n over 0) * 0,5^n - (n over 1) * 0.5^n.
Furthermore 1 - 1 * 0.5^n - n * 0.5^n = 1 - (1+n) * 0.5^n
Now this formular depends on the amount of shots we want to do in total. But we're interested in the average value, so we want this probability to exceed 0.5.
0.5 <= 1 - (1+n) * 0.5^n => (1+n) * 0.5^n <= 0.5. As n can only be an integer value, we find:
n = 0: 1 <= 0.5 wrong n = 1: 1 <= 0.5 wrong n = 2: 0.75 <= 0.5 wrong n = 3: 0.5 <= 0.5 correct
Therefore we have 50% probability to kill the rine in 3 shots, no? You have a 50% probability of killing them in 3 shots, but you will kill it on average in 4 shots (because occasionally it will take you 5 or 6 shots to kill it). Think about it like this: If a tank shoots 80 times, it will hit 40 times and there will be twenty dead marines. Thus, on average, it will take four shots to kill a marine. As you say though, there is a 50% chance that the marine will die in the first three shots, but that doesn't mean that it will take an average of three shots to kill the marine.
|
Or sometimes, the unit might not die at all. ;P Anyway, on average 4 shots is correct but I oversimplifies the case and is not fair to compare both methods in such a way. However this is neglectible since the main article emphazises to change something at all. I don't care how it is fixed, as long as it gets addressed.
Btw, nice writeup anyway.
|
On March 10 2010 06:18 Vasoline73 wrote: Dude what the eff with people saying "plz no randomness."
How many times have you watched a SCBW match and shouted "omg come on how lucky can you get?! He totally got jewed by fighting from the low ground!! how lame!!" Seriously, I've never seen high ground "luck" appear broken in BW. If you attack up hill and lose, it's not luck, you're just dumb for fighting a battle from the low ground without taking into account high ground advantage. Low ground has a disadvantage in battle, it just makes sense.
How does damage reduction make sense? Does a bullet hurt 30% less if you get shot from low ground IRL? If you get hit, it doesn't matter if it's from low or high ground, the damage is the same. Low ground miss % makes more sense strategically because IRL that is the disadvantage of low ground. High ground has better vision and an easier time picking out targets... low ground has more problems to deal with and thus has a lower % of hitting their targets.
And before anyone with 18 posts says "hurp durp are warp gates and aliens IRL? nothing has to be realistic", I'm just saying strategically the game is much more interesting if low ground has a miss % instead of a set damage reduction. The game is deeper and battles will plays out more as it should, where high ground is just one aspect of "advantage" in an encounter. A set damage reduction is really shallow and like Chill alluded to once, makes the game easier to fully comprehend. Attacking from the low ground should be a risk. When you know "ok im fighting with 30% damage reduction but I know how much stuff he has with my obs so I know I'll win the fight regardless of the ramp and wall in" it's not a question anymore. Attacking from the low ground should be a choice.. by making the damage reduced you take away that choice because people will always know when it's right to engage from the low ground and when it's wrong. Which is boring and shallow IMO. And less realistic in a battle.
TLDR: Miss % is going to make the game deeper and more interesting than a flat damage reduction from the low ground
man, you must be joking. Don't you see the difference between luck that's always the same with same input and luck that's simply random? Imagine losing a game - not because you did a mistake, but because missed two times in a row. Or three times. The skill to count that in isn't any near the worth to negate luck. Did you play wc3 or smth?
|
Nice write-up. I agree that some sort of high ground mechanic should be put in place (I'd favor a random miss chance over straight up damage reduction, similar to what bw had.)
Right now the defenders advantage just doesn't seem significant enough.
|
so starcraft 2 should be more like starcraft 1. great suggestions.
|
Absolutely spot on, great write up Nazgul!
I would also be interested to hear your opinions on the current dynamics of air in SC2. Is there too much air dominance currently, and nothing to really punish it? - should it be punished? The whole air idea coincides with your points on positioning/terrain advantages, and the advantages gained from having a mobile army. It's easy to compare it to broodwar and long for the drawn out battles fought on the ground for map control, but of course... it is a new game. But, it seems to me that it almost removes some valuable RTS mechanic that we saw in broodwar, where it was possible to control areas of the map with ground forces. In SC2 your ground is far too vulnerable to air attack, as there are no easy counters (corsairs/goliaths/valkyries/scourge etc).
|
i definitely agree with mostly everything said here, especially with the hit/miss chances, I often used things like trees and cliffs as an advantage, take Heartbreak ridge as a PRIME example, there are huge portions of high ground vs low ground, and lost temple has trees (may not seem like much) but they also have a miss/hit ratio that can save you early game...
I also believe this idea should be tested because of the added level of ground (HIGH-MID-LOW) this leaves so many options open for map makers who, guaranteed, would love this feature implemented.
the thing is that with how high ground vs low ground is now, it leaves no room for creativity beyond making a map that is all mid ground and has chokes everywhere, that is the only way this mechanic works right now, as an equal choke between both players...
Ive had many matches where i was maybe one or 2 of the same units down, defending a cliff and still got pwnd, this would not happen in broodwar (looked for the video of 3 marines shooting down at a sunken and winning but couldnt find it)
I hope you posted this in the Beta forums on Bnet, keep up the good work and if you want to try more testing I'm down for the cause lol
|
Good read. This is what beta is for! Beta is most likely going to be going on for like 4 more months or so. Why can't they change these things and see what happens? If they don't like the results change them back!
I can't help but feel like they should be patching so much more often and trying out different changes for balance changes.
Also that interview with dustin browder really makes it seem like all he cares about balance-wise is every race having a 50% win ratio when there is so much more depth to the game than just race vs race balance.
|
I'm sorry I'm not a regular TL poster and that I've missed the other discussions of this issue, but I have some questions and observations I'm not seeing addressed in this thread.
On March 10 2010 09:55 Daigomi wrote: The point of my stats was not that miss is better or worse than damage reduction, simply that it is easier to balance.
Correct me if I'm wrong, but adding a chance to miss only makes the game easier to balance "statistically." In general, however, wouldn't the addition of random chance be a factor that actually works against precisely predictable balance?
Sure, "statistically" you'd preserve the hit/kill ratio for tanks vs. marines on a flat battlefield. "Statistically" you'd just be reducing the effective speed of every unit's hit/kill ratio when the units are firing at higher ground. However, as far as I understand the world, the nature of statistics is that you can't count on them in crucial, isolated moments (statistics only follow norms on the large scale) and I'm personally hoping that winning in StarCraft 2 will rely upon those kinds of crucial-but-small differences (just as SC:BW did).
StarCraft and StarCraft 2 (the little I've been able to test it with my girlfriend's beta key . . . sigh) are games that can have drastic outcomes decided by tiny margins. Neither game is very forgiving in the sense that they give players plenty of opportunities to heal units or escape from a battle. As such, it wouldn't seem out of the ordinary for me to imagine a single, extra-unlucky miss by a single siege tank which could then yield a drastic result in the other player's favor. Sure, your random, high/low ground factor could offer the potential for other unlucky shots at other points in the same game, but I have a great deal of trouble believing that every possible miss provided by that factor will be equally as critical to the game.
Yes, percentage-based damage reduction is harder to balance, but at least it's potentially predictable and (again, correct me if I'm wrong) predictability is the bedrock of strategic balance. For example, you could flip a coin with a friend, and that's balanced statistically (both participants get one side of the coin), but in that same situation there's absolutely no balance that is affected by player choices, player skill, and player mistakes. All random elements Blizzard puts into SC2 will have that exact same relationship to player interaction. (At best, a player can only predict random results and strategically embrace or avoid them.) In other words, while, obviously, the effect you're talking about for SC2 would be smaller than the effect of chance in coin flipping, the fundamental relationship is the same, and that could be undesirable if, again, as I believe, small advantages in a single battle can often (and should often) lead to large results.
I think balancing for damage reduction would pay off in the long run by giving us a more interactive, skill-based game. In fact, we could say that the complexity of having hit/kill ratios altered in cliff battles (as they were in SC1) adds a greater potential for player mistakes (and greater potential for a really skilled, smart player to gain an advantage). Some of the people here who are whining about how they don't want SC2 to be a predictable game are actually desiring a game that requires no skill (and they should go play coin flipping or "shoots and ladders"). If the predictability of a game is too simple (and therefore too easy) for them, then that's the fault of the game design, not the presence of predictability itself. Making a game harder to balance is generally beneficial for strategic gameplay (as SC:BW proves).
Otherwise, there are ways to preserve predictability in an altered cliff dynamic without that dynamic being different from the unit balance on flat ground:
1. Blizzard could give each unit specific bonuses and negatives when firing at a unit on higher ground (which would be easy balance and very easy to learn). This could also potentially add greater level of fun to the game by providing whole new uses for underused units when they fire at (or from) higher ground. (They could also keep the fog shroud for smaller, quieter-firing units, and have larger, louder units—such as the Siege Tank or Collossus—give away their position in the shroud.)
2. Blizzard could add a simple-to-understand, but difficult-to-calculate modifier for damage that is based on hits to kill. So if a 40 hp Marine takes two hits to kill from Siege Tank damage, a 50% damage reduction by this modifier would lower Siege Tank's damage to the point where 3 hits always do less than 40 damage to a marine while 4 hits always do more than 40 damage. (It would take me a while to think of a formula which would do this elegantly, however, so I'm not going to bother suggesting anything specific here.)
Both of these solutions would do what you want without introducing more game-changing randomness into StarCraft 2. Call me crazy, but I'd prefer that; it helps preserve the kind of strategic gameplay I enjoy.
|
Both of these solutions would do what you want without introducing more game-changing randomness into StarCraft 2. Call me crazy, but I'd prefer that; it helps preserve the kind of strategic gameplay I enjoy.
Agreed!
But the changes will also add another great "feature" back to Starcraft that many players seem forget these days. "The Power of Gosudancing"! Seeing two players not wanting to engage each other because of Low/Highground Ad/Disadvantage was quite fun to watch :D
|
Australia4514 Posts
I agree 100%
I'd add some other things.
When you madly click units to retreat, the unit ai messes up and it half freezes. When you click too slowly as it gets attacked, it starts to retreat, then it runs back to reengage.
That annoys me a lot. When i give a move order i want it to move, i do NOT want it to go off and do stupid things. If i want it to engage enemies, i would have given it an attack move order.
|
On March 11 2010 23:34 Legionnaire wrote: I agree 100%
I'd add some other things.
When you madly click units to retreat, the unit ai messes up and it half freezes. When you click too slowly as it gets attacked, it starts to retreat, then it runs back to reengage.
That annoys me a lot. When i give a move order i want it to move, i do NOT want it to go off and do stupid things. If i want it to engage enemies, i would have given it an attack move order.
It feels a bit weird to be in the presence of legends.
Good points though, the unit control and the high-ground mechanic must be improved.
|
nice read, im unhappy about all that things also : (
|
On March 10 2010 23:05 son1dow wrote:Imagine losing a game - not because you did a mistake, but because missed two times in a row. Or three times. The skill to count that in isn't any near the worth to negate luck. Did you play wc3 or smth? I'm fine with the player who plays better not winning every single game. The player who plays better will still win the vast majority (>95% easily) of games. You can have good competitive games where the player who plays better doesn't win nearly that often (card games come to mind). Of course, you can have good competitive games where the player who plays better wins every single time too (chess, for instance).
If you lose because you're attacking up a ramp and get unlucky, you took a calculated (or not-so-calculated, if you're not thinking) risk and lost--it's your fault. If you lose because your opponent attacks up a ramp and gets lucky ... I would think that would be what would bother you, since that's not your fault at all.
|
Possible solution for blizzard on proxy gating: Make it where the nexus has some sort of max radius that proxy's are limited to building in that range? (Nexus has some sort of grid that shows where you can build pylons, multiple nexus more possibilities.
What makes this good is regardless if a terran tries doing a fast rush with marines they would get rocked against T | Z. I personally think that would work out best for the game.
-------------
I do like #1 solution though - 1) When two enemy pylon fields meet the warp function ceases to work. One thing I hate about this solution is, what about PvZ gate rushes, we don't have pylons this idea doesn't work out as well.
My full solution on blizz: http://forums.battle.net/thread.html?topicId=23710232306&postId=237080590111&sid=5000#0
|
On March 11 2010 11:11 Tiptup wrote: Yes, percentage-based damage reduction is harder to balance, but at least it's potentially predictable and (again, correct me if I'm wrong) predictability is the bedrock of strategic balance. For example, you could flip a coin with a friend, and that's balanced statistically (both participants get one side of the coin), but in that same situation there's absolutely no balance that is affected by player choices, player skill, and player mistakes. All random elements Blizzard puts into SC2 will have that exact same relationship to player interaction. (At best, a player can only predict random results and strategically embrace or avoid them.) In other words, while, obviously, the effect you're talking about for SC2 would be smaller than the effect of chance in coin flipping, the fundamental relationship is the same, and that could be undesirable if, again, as I believe, small advantages in a single battle can often (and should often) lead to large results.
I would talk way too much if I chose to respond with both exhaustive counterexamples and a theoretical rebuttal (as below), and I don't think either of us is looking for a fight to the death, so I'll just list some "soft" counterexamples as something to think about. Is Settlers of Catan predictable? Is it strategic? What about old fashioned marbles? Is Risk predictable, or strategic? I don't think the way you phrased the relationship between predictability and strategic balance is exactly accurate. I think players use predictability to assess strategies. But more fundamentally, your first window on strategies in any game is examining the possible lines of play, and what outcomes may result. Your certainty then plays a role in "scoring" those possibilities, sort of like expected utility.
The coin flip example doesn't translate into a complex game design as cleanly as you say. In your game, it's just a process, not a game. We agree you could also describe Shoots and Ladders this way. You're just flipping coins, so of course there's no interaction. But in Starcraft, if high ground provides chance to miss, the players, mostly the attacking player, choose whether to engage in "coin flipping", and how much and in what manner, which is a world of difference. More than that, once a battle begins, the tactics of cycling low health units out and in, and focus firing, etc, are modified to accommodate the miss chance. It's just one element of a much larger field of interactions. Players have tons of room to participate skillfully in that situation. As others have pointed out, when you have high ground damage reduction, the decision (given an informed attacker) comes down to "whose army wins?" which seems to limit decision making. But miss chance vs damage reduction aside, my main point is that random processes can be used to specific ends in design, to modify the strategic topography, if you will. Often you can use analogues, like damage reduction, and often the differences are subtle. But competitive play highlights subtlety, which is why we're arguing now.
If the predictability of a game is too simple (and therefore too easy) for them, then that's the fault of the game design, not the presence of predictability itself.
The predictability is the result of the design; it's arbitrary.
There's another thing I'd like to point out. There are two substantial ways in which Starcraft has functionally random processes, discarding high ground things. Consider that when you play, there are things you want to do but which you can only approximately do--your control of the game elements is imperfect. That's why I used marbles, a dexterity game, as an example. You can't exactly perform the strategy you've crafted, and the keen player will even factor that into his or her strategy. This means that when a battle occurs, for instance, the incremental mistakes or brilliances that determined the outcome were in part accidental. Even with two equally very skilled players the unit control is imperfect, which you can imagine results in "lucky" good breaks and "unlucky" bad breaks. I'm talking things as incremental as a zergling here or splash damage on your goons there. We conceptualize these incremental things as part of the melee of an engagement and write it off. But you can view it as strategic event whose outcome relies partially on random elements.
Consider also that a complex game like Starcraft is a series of inputs that depend on iterated outputs. The players "perturb" a machine in order to nudge its processes in their favor, watching how they're doing in real time. It might seem esoteric to describe "zealot rush" as perturbation of a systems process, but grant me it's a valid perspective. The reason I'm going here is because I want to bring up a property of systems and relate it to Starcraft. It's possible for a complex system with perfectly deterministic rules to have entirely unpredictable future states given minor changes in initial conditions, or inputs. This is roughly what people are talking about when they say chaos theory. Thus, even if we had perfect control of our units and buildings on a frame by frame basis, it wouldn't even be clear then how to proceed optimally. To provide a rhetorical bridge to why this means our strategies involve randomness: the iconic example of an unpredictable system is a fractal object. If you calculate its divergence at a given coordinate, there's literally no way to predict the divergence at a nearby coordinate. If you map these though, you get pretty pictures with an infinitely fine grain of ever changing details. So despite the fact that you're using one simple formula to construct the object, its microstructure is essentially random upon observation. That's the second way in which Starcraft, and any other complex game, has randomness. Beyond enumerating all possible outcomes, which is roughly how chess theory has proceeded, you can't count on predictability. So our army positioning, the timing of our teching, our choices of when and how to attack and harass, are all informed by the gamestate as we perceive it, but that gamestate could have been wildly different depending a minor change here or there. Of course the general flow of our games is understandable and repeateable; that's the result of how the system operates, how the game was designed. It would be a bad game if it didn't operate the way we roughly expect it too. But again, the subtleties are where the skillful player captures advantage, and the most subtle things are unpredictable, which appears random to the observer. And both players are observers of the system, though they also are giving it inputs.
Both of these solutions would do what you want without introducing more game-changing randomness into StarCraft 2. Call me crazy, but I'd prefer that; it helps preserve the kind of strategic gameplay I enjoy.
No, call ME crazy! Yes, more than one way to skin a cat. (What a terrible idiom.) I think random elements are no less skill-testing, and often make for more interesting strategic landscapes and game play.
|
|
|
|