EPTPerfect micro with Phoenixes - Page 4
| Forum Index > SC2 General |
|
XanaNA
United States15 Posts
| ||
|
chairmobile
United States111 Posts
On July 25 2014 06:44 phil.ipp wrote: what ? why not its pretty easy, just kite in a circle .. Have you heard of fungal? Hydras? Queens? No? There's this great thing called liquipedia you might want to check out. That being said, perfect kiting with nix is hardly necessary given shields (note: I mean absolutely perfect or near perfect, not good kiting). | ||
|
KingWilhelm
11 Posts
It exist one Corrupter j with d(j,t)= min_i d(i,t), where d(i,t) is the distance between the Phoenix and Corrupter i at time t. For all k!=j with d(k,t)=d(j,t), we know that Corrupter k lies on a circle around the Phoenix. As the Phoenix and the Corrupters are linearly separable we know further that a sector of a circle enclosing all Corrupters k exist. We choose one of the edge Corrupters k and can use the above micro in direction of this Corrupter. (At this point it is obvious that linear separability is stronger than needed. It would be enough that the Phoenix is not trapped by the set of Corrupters ... but I'm not in the mood to find the corresponding assumptions  ) If at any time T a second Corrupter j with d(j,T)= min_i d(i,T)= d(k,T) exist, we start to use the above mico in regards of this Corrupter (as we where traveling towards him) and so on.EDIT: For more than one Phoenix it gets messy if the Phoenix are not "close enough" to each other. See p_1-----------D---------------p_2 |-----------------------------| d-----------------------------d |-----------------------------| c_1------------c_3------------c_2 p Phoenix, c Corrupter, distance between p_i and c_i d, distance between p_1 and p_2 D>d. p_1 and p_2 should use different mico to avoid getting hit by c_3. | ||
|
Salient
United States876 Posts
| ||
|
Sholip
Hungary422 Posts
On July 26 2014 23:30 Salient wrote: Too bad there isn't a painted circle to follow in game. The Phoenix only outranges corruptors by 1, so this works better in theory (where you can focus your full attention on 1 unit against a non -microed corruptor while following a painted circle) than in practice. Of course. But it's definitely something that isn't impossible to learn, and I'm sure the top players have enough APM to do at least something similar. | ||
|
Lobotomist
United States1541 Posts
| ||
|
kramuti
4 Posts
In the end, the 'real' result is possibly best modeled using field theory, and pertubation theory...you could actually write a dissertation on a game of starcraft...and it is directly applicable to almost any field in science, math, economics, biology...games are fun! | ||
|
fenix404
United States305 Posts
| ||
|
Advantageous
China1350 Posts
On July 25 2014 06:38 Big J wrote: That's probably the most well-researched post ever in the history of SC2. :-) Actually... from the beginning of the phoenix hype in all match ups you can kinda tell that phoenix's faster firing rate would actually make them a counter to units like corruptor... but, nonetheless, props for doing the math and actual research! | ||
|
neptunusfisk
2286 Posts
On July 25 2014 07:23 Orek wrote: This is super impressive. I haven't seen anyone else in this community who has such deep knowledge both in math and game editor. Also, so much passion for SC2  I'm looking forward to reading your researches!!This isn't really advanced math you know | ||
|
eightym
United States76 Posts
| ||
|
hfsrj
Germany166 Posts
| ||
|
Banchan
United States179 Posts
| ||
|
timchen1017
37 Posts
In reality, suppose the corrupter indeed follows (so that the term kiting makes sense), you can kite with any radius larger than your calculation shows, including a straight line. You just need not to move at all times, effectively reducing your speed. Now this in isolation is probably harder to execute, as you not only have to place your move command at the right place but also at the right time. But in game it is probably seen more often, as there is no strategic need to move in a circle anyway (so that you just try as you can to keep the distance between your phoenix and the corrupter the same as the phoenix range.) | ||
|
Reborn8u
United States1761 Posts
| ||
|
timchen1017
37 Posts
On August 03 2014 06:22 Reborn8u wrote: If you wanted to perform this but head in a specific direction across the map, could you perform this in an "S" shape? Or would further modification need to be made to the path? You can just move in a straight line. But you have stutter step to keep the corrupters in range (and not too much to be too close to them). | ||
|
Sholip
Hungary422 Posts
On August 03 2014 06:22 Reborn8u wrote: If you wanted to perform this but head in a specific direction across the map, could you perform this in an "S" shape? Or would further modification need to be made to the path? At any point of the circle, you can decide not to continue the current circle, but to start another one, which is the mirror image of the original one with the line defined by the Corruptor and the Phoenix as mirror axis. That means, if you travel clockwise, you can turn by 2*alpha counterclockwise and start a circle counterclockwise (or vice versa). That is because only the size of the angle you have to hold matters, whether it's cw of ccw does not. The resulting curve will not be C1 continuous, i. e. it won't be smooth, but will rather have a "break point". It will not be an "S", but some weird looking curve, that's for sure. | ||
|
Cascade
Australia5405 Posts
I think that it is possible to solve this with a bit less math, but that is all fine. As you pointed out in your post just above, you missed an important part though. Cos(x)=y has the solution, x=+-acos(y). The plus minus is what you say above, the the phoenix at any point can change direction. And a bit of zig-zagging is probably often a better solution in a real game. Further, you could do a similar calculation for a corruptor moving away from you, and how you would have to move towards it at an angle to keep within range without getting to close. So it would still be a circle. And maybe also solve for a corruptor moving in a straight line, independently of the phoenix. How to kill retreating corruptor without getting in range. For groups of phoenix/corruptor, you could approximate the groups of units with circles (yay physics!) And see how quickly you deal damage/take damage at different distances. A safe distance would be when no part of the circles are within 6 units distance, but you can deal damage faster (still at a good exchange rate) if you allow the circles to partially go within range. How much phoenix damage are you willing to take to damage the corruptor faster? You solve for steady state, always same distance, but in practice it is equally important to know how to get there, and how to adjust, when you are getting too close/far away. The answer her would be to adjust the escape angle more directly away to slowly increase distance, and angle more towards the corruptor to close distance slowly. This would be a lot easier than what many people are doing now, which is running directly towers or away from the corruptor, giving you a very small time window when you are at exactly the right distance. By adjusting the angle you can increase or decrease distance slower, and it'll be easier to find the right distance. | ||
|
DnCL
86 Posts
2: I try to reformulate your post in my words, how close is it ? '' I've formulated an hypothesis, then verified in game, that there is a way for Phoenixes to never stop attacking OR shooting a target that follow them, and it is to travel on this circle.'' (shown above) 3: I find very interesting the answer Then you have to adjust accordingly to another circle. It shouldn't cause you to lose Phoenixes because the Corruptor can only attack your Phoenixes if your opponent recognizes the pattern in your movement and intecepts the Phoenixes at another point of the circle. Worst thing to happen is that you can't kill the Corruptor. I feel it is sometimes what the best players do already. A: The zerg tries to do swift movement toward the Phoenixes or away from them in a attempt to ''break the circle'' B: If feel that's the important point: The protoss should not adjust with a movement on the same direction, but simply try to modify it's circle. 4: I'll try to notice that in replays and post them to see what we can see at high level of play. | ||
|
Sholip
Hungary422 Posts
3: It was a poor choice of words on my part. Again, you have it right: the Corruptor will stop following the Phoenix at some point, trying to move away. Then, theoretically, you should immediately start simply following the Corruptor (you will still be able to attack it and it won't attack you). If the Corruptor ever turns back to attack the Phoenix again, you will have to start moving in a circle again. This will be another circle, obviously, since the positions of the units have changed. 4: I don't think you will really see this in actual games, because I think it requires far too much (constant) attention and you are pretty well off with just good, not perfect, micro, due to shield regeneration (I may be wrong, though). | ||
| ||
