EPTSpecial Additions 2011-04-28
On April 15 2011 16:44 RibsNGibs wrote:
Would be nice to see main->opponent-nat distances on maps that are not symmetric (rotational symmetry). e.g. Delta quadrant main->nat is ridiculously short if going CCW, which is important for, say, defending a fast expand from a rush.
Okay, so here are the main2opponent'sNatural distances going CW and CCW on DQ, Backwater and Tal'Darim Altar. I included air distances, too, and calculated the CW-CCW disparity like this:Would be nice to see main->opponent-nat distances on maps that are not symmetric (rotational symmetry). e.g. Delta quadrant main->nat is ridiculously short if going CCW, which is important for, say, defending a fast expand from a rush.
main2oppnat CW by ground = X
main2oppnat CCW by ground = Y
disparity = (X - Y) / MAX(X, Y) and expressed as an absolute percentage
So the disparity is how much shorter, as a percentage of the longer distance, is the shorter distance?
Note: ignore old map pictures from Liquipedia, I did run the analyzer on the latest Blizzard versions of these maps.
![[image loading]](http://imgur.com/Y2D0U.jpg)
![[image loading]](http://imgur.com/K4Ak6.jpg)
![[image loading]](http://imgur.com/X2YeG.jpg)
On April 20 2011 01:12 ZeroTalent wrote:
Thanks for updating this for the new map pool!
I checked to see if this is covered in the thread, but I've heard that the walk distance for close air is actually longer than cross positions on some (all?) maps with 4 spawn points. Is that true, and if so for which maps?
Shakuras is the only map like that because it has a very unusual (and subtle) layout. Check the last spoiler called "Wacky Shakuras." You'll see that the 4 spawns are a rhombus meaning the top-left spawn and bottom-right (we normally say cross for this, right?) are on the shorter path across the rhombus. This path is shorter by ground than going to the close-by-air positions. NOW, if you spawn in the other cross-positions (BL and TR) its not true anymore, this spawn is longer than the walks to the close-by-air positions.Thanks for updating this for the new map pool!
I checked to see if this is covered in the thread, but I've heard that the walk distance for close air is actually longer than cross positions on some (all?) maps with 4 spawn points. Is that true, and if so for which maps?
On April 20 2011 01:12 ZeroTalent wrote:
Also how do I go from distance (in map cells) to time (in game seconds)? Divide by the unit's speed as listed in Liquipedia?
Yes. But don't use this as gospel for several reasons:Also how do I go from distance (in map cells) to time (in game seconds)? Divide by the unit's speed as listed in Liquipedia?
- The analyzer does not consider the exact path a unit would take (it only considers 16 possible directions to take each step.) It's within a few percent of the distance a unit takes.
- The distances between equivalent spawns (like there are two ways to spawn cross positions on Metalopolis) are in general not actually the exact same distance. I don't list every one individually, so whatever number I picked is just one of the possible spawns of approximately the same distance.
- Units bork each other up a little when there is more than one. Especially if there are doodads on the shortest path to where they are going.
Updated 2011-04-14
The data for this chart was generated with the SC2 Map Analyzer. main2main is measured from start location to start location, choke2choke refers to the main chokes, nat2nat is measured from natural to natural.
![[image loading]](http://imgur.com/hZAKu.png)
Some interesting points:
- The Shattered Temple's shortest choke2choke is shorter than Delta Quadrant.
- Shakuras Plateau has really tight distributions in each measurement because the shortest paths are very similar. This suggests timings can be very reliable in any spawn. Typhon Peaks has pretty tight distributions for such a big map as well.
- Metalopolis and Slag Pits have a lot of spread, so pick this map if you want your opponent to sweat the variations.
- Xel'Naga Caverns has a solid reputation; maybe its not a coincidence its data point is practically the mean for each measurement across the maps.
- Backwater Gulch is interesting because in choke2choke it measures medium-short compared to the map pool, but main2main and nat2nat is measures medium-long. I suspect this is a good pattern for a balance between aggressive and greedy styles being effective.
- Tal'Darim Altar's longest choke2choke is longer than the longest choke2choke on Shakuras Plateau by a good shot. The biggest maps in ascending order are Typhon Peaks, then Shakuras, then Tal'Darim.
Link to the raw data.
+ Show Spoiler [2010-10-12 Ladder] +
![[image loading]](http://www.jamesjenista.com/sc2melee/RushDistance-Ladder-2010-10-14.png)
Note: Desert Oasis is still on the chart in light grey just as a point of comparison.
+ Show Spoiler [2010-08-05 Ladder] +
![[image loading]](http://www.jamesjenista.com/sc2melee/RushDistance-Ladder-2010-08-05.jpg)
+ Show Spoiler [Beta Ladder] +
I searched around for this and didn't find what I wanted, so I compiled a comparison of rush distances for the Blizzard maps that I will use as a reference point for maps I create and post in my blog. I figure the comparison is interesting enough other folks might want to take a look at it.
![[image loading]](http://www.jamesjenista.com/sc2melee/RushTimes.jpg)
And a link to the raw data.
The experiments were simple custom games where I spawned as Terran with no opponent. Then, I positioned one SCV next to the CC and ordered it to move to another start location. If it was a 4 player map I sent an SCV, one at a time, from the same spot next to my CC to the other three locations. Then I checked the replays and subtracted the arrival time of an SCV from the departure time, pretty simple stuff. The SCV is just a convenient unit for this experiment, the point is to get data plotted so we can get a feel for the relative rush distances on maps, which have a linear relationship with the SCV rush times.
There are some interesting facts in the data.
-Incineration Zone is as short a rush as LT or Metalopolis, but you don't have to scout to find that out. Not a big surprise, but its nice to know for sure just how short it is.
-Desert Oasis is a whopping 36% further rush distance than the 2nd longest, Blistering Sands. I've been experimenting, like I'm sure others have, with larger maps and I expect my rush distances to come out somewhere just shy of Desert Oasis. It just goes to show you how "small" the majority of the Blizzard maps really are.
-Kulas Ravine's various spawns are very closely related in rush distance, which we can clearly see is because the paths converge on the center area for any spawns. This could be interesting information when designing builds for Kulas Ravine--you can count on timing attacks to be fairly consistent.
-Metalopolis has the widest variation. Again, this could be helpful for tweaking a build to allow for variations in timings, or exploiting the short spawn when you get it.
-I don't know if other people realize this, but it feels to me as though Scrap Station has a significantly longer rush distance than Blistering Sands, when in fact Sands is a little longer.
And a link to the raw data.
The experiments were simple custom games where I spawned as Terran with no opponent. Then, I positioned one SCV next to the CC and ordered it to move to another start location. If it was a 4 player map I sent an SCV, one at a time, from the same spot next to my CC to the other three locations. Then I checked the replays and subtracted the arrival time of an SCV from the departure time, pretty simple stuff. The SCV is just a convenient unit for this experiment, the point is to get data plotted so we can get a feel for the relative rush distances on maps, which have a linear relationship with the SCV rush times.
There are some interesting facts in the data.
-Incineration Zone is as short a rush as LT or Metalopolis, but you don't have to scout to find that out. Not a big surprise, but its nice to know for sure just how short it is.
-Desert Oasis is a whopping 36% further rush distance than the 2nd longest, Blistering Sands. I've been experimenting, like I'm sure others have, with larger maps and I expect my rush distances to come out somewhere just shy of Desert Oasis. It just goes to show you how "small" the majority of the Blizzard maps really are.
-Kulas Ravine's various spawns are very closely related in rush distance, which we can clearly see is because the paths converge on the center area for any spawns. This could be interesting information when designing builds for Kulas Ravine--you can count on timing attacks to be fairly consistent.
-Metalopolis has the widest variation. Again, this could be helpful for tweaking a build to allow for variations in timings, or exploiting the short spawn when you get it.
-I don't know if other people realize this, but it feels to me as though Scrap Station has a significantly longer rush distance than Blistering Sands, when in fact Sands is a little longer.
+ Show Spoiler [Wacky Shakuras] +
Now, there has been some debate about Shakuras Plateau so in case you've been living under a rock, a really important fact about this map is that the closest spawns (NW+SW and NE+SE) are disabled! I looked at the Advanced Teams tab to verify this, and anyone can create custom maps with this spawning behavior. And it turns out that Blizzard did something really clever and cool by including this former 2v2 map in the 1v1 pool, in terms of spawn possibilities:
![[image loading]](http://www.jamesjenista.com/sc2melee/ShakurasPlateau.png)
I'm not sure if this phrase is geometrically correct, but Shakuras Plateau has 1-axis rotational symmetry, like Blistering Sands for instance, except that its a four player map. Look at how the ramps below the natural form a rhombus; usually on a 4-player map (rotational or mirror symmetry) the choke points form a rectangle like on Lost Temple or Delta Quadrant.
What this means is that Shakuras Plateau has three distinct rush distances, but in a different pattern than LT or Metalopolis. On SP if you spawn NW and SE then its the short path across the rhombus, while SW and NE is the long path across the rhombus. If you spawn both N or both S then the rush distance is the same, but the terrain is slightly asymmetrical (look at where the ramp for the empty base is for each player--either jutted in or out).
So to recap, there are three rush distances (short cross, mid, and long cross) but there are four isomorphic spawns from one player's perspective: short, long, nat ramp pushed in, or nat ramp pushed out. Overall the rush distances range from Xel'Naga Caverns to Blistering Sands, and are clearly the longest (on average) for a four player map.
I don't know if this map will hold up in the long run, but I think that Blizzard just showed us a new family of map layouts we can experiment with, and one they might drop on us in future maps.
I'm not sure if this phrase is geometrically correct, but Shakuras Plateau has 1-axis rotational symmetry, like Blistering Sands for instance, except that its a four player map. Look at how the ramps below the natural form a rhombus; usually on a 4-player map (rotational or mirror symmetry) the choke points form a rectangle like on Lost Temple or Delta Quadrant.
What this means is that Shakuras Plateau has three distinct rush distances, but in a different pattern than LT or Metalopolis. On SP if you spawn NW and SE then its the short path across the rhombus, while SW and NE is the long path across the rhombus. If you spawn both N or both S then the rush distance is the same, but the terrain is slightly asymmetrical (look at where the ramp for the empty base is for each player--either jutted in or out).
So to recap, there are three rush distances (short cross, mid, and long cross) but there are four isomorphic spawns from one player's perspective: short, long, nat ramp pushed in, or nat ramp pushed out. Overall the rush distances range from Xel'Naga Caverns to Blistering Sands, and are clearly the longest (on average) for a four player map.
I don't know if this map will hold up in the long run, but I think that Blizzard just showed us a new family of map layouts we can experiment with, and one they might drop on us in future maps.
Desert Oasis must be so bad
