|
On November 25 2025 04:29 Dakota_Fanning wrote:Enthused by this nice analysis today I also added 1v1, 2v2, 3v3 and 4v4 location stats to all maps over RepMastered. These contain games of all races, and all games (not just Blizzard ladder games). For example 투혼 1.4 1v1 locations: 1 v 4, 880k games, 52% win ratio 1 v 8, 879k games, 52% win ratio 4 v 11, 879k games, 50% win ratio 1 v 11, 877k games, 52% win ratio 4 v 8, 877k games, 49% win ratio 8 v 11, 872k games, 50% win ratio Based on this data, top right location (1) is the favored, and bottom right (4) is the most unfavored location. Polypoid 1.75 as an another example seems more balanced regarding starting locations. And Eclipse 1.2 also seems almost perfectly balanced in this regard.
An excellent and considerate addition. Thank you very much!
Would you consider adding spawn location data for each racial matchup? That would greatly help us understand how spawn location affects the matchup winrate too.
Again, thank you very much for the data update!
|
For the ladder bug, I found this thread., but it seems like what was broken was the fact that colors weren't random but instead fixed by the spawn location. I think this wouldn't have broken the spawn stats, but not sure.
Still, on FS1.3 we see some spawn-vs-spawn variations being way more frequent than one would expect to happen by chance. I looked at a few other maps too, but FS1.3 seems to be skewed the most.
What is also kinda interesting is how there are more than thousand different hashes (versions) of FS1.3 that repmastered has detected. Is it possible that people were tinkering with maps and this is why we see the stats being somewhat off? Like, everyone remembers joining a game on a modified FS1.3 where the opponent had infinite minerals etc; maybe there are versions where people changed or removed some spawns?
Also this locations stats addition to RepMastered are not just for 1v1 games and are not just for ladder maps: they are for all maps and for 1v1, 2v2, 3v3 and 4v4. A lot of us play regularly or occasionally team games, often with and against random races, and this might give a clue what to expect on start.
It is useful and interesting in itself, I liked browsing the bgh stats and the stats for fastest map a lot.
|
On November 25 2025 21:21 Kraekkling wrote:For the ladder bug, I found this thread., but it seems like what was broken was the fact that colors weren't random but instead fixed by the spawn location. I think this wouldn't have broken the spawn stats, but not sure. Still, on FS1.3 we see some spawn-vs-spawn variations being way more frequent than one would expect to happen by chance. I looked at a few other maps too, but FS1.3 seems to be skewed the most. What is also kinda interesting is how there are more than thousand different hashes (versions) of FS1.3 that repmastered has detected. Is it possible that people were tinkering with maps and this is why we see the stats being somewhat off? Like, everyone remembers joining a game on a modified FS1.3 where the opponent had infinite minerals etc; maybe there are versions where people changed or removed some spawns? Show nested quote + Also this locations stats addition to RepMastered are not just for 1v1 games and are not just for ladder maps: they are for all maps and for 1v1, 2v2, 3v3 and 4v4. A lot of us play regularly or occasionally team games, often with and against random races, and this might give a clue what to expect on start. It is useful and interesting in itself, I liked browsing the bgh stats and the stats for fastest map a lot.
This occurred also when playing ladder and the ladder maps were always provided by the Blizzard server. So no rigged maps were used in ladder matches, only the correct FS1.3.
And don't forget, it happened on every map, FS is just the most popular map that's why we remember that more specifically. If you would evaluate only the time frame in question, you should see huge discrepancies from the expected result for all ladder maps at the time.
In custom games, some people might have used rigged maps, but I suppose many of them are simply UMS maps that used FS as basis.
|
On November 20 2025 21:17 [sc1f]eonzerg wrote:
Any terran player that could explain why Polypoid bottom right is such a good spawn for you ? I think top right and bottom right what they have in common is the same mineral lines. Sim City is different. Is it by being at the right side any beneficial with defense and mobility of your army any different to the left side ?
i'm not like a good player and i havent played in like 1.5 years. but theres something about this part of the map right here that just feels a lot easier to use as a terran in the bottom right than in the top right. i don't have a good explanation really. but i feel like it's just a lot easier to defend against any air units? if i place a high ground siege tank it just seems stronger somehow too.
if i had scmdraft installed i would try to compare the walkable terrain and see what's going on. or it might be just an emergent property of how a terran base is going to be laid out on the two spawn positions, since buildings dont rotate. i'm not sure
|
Unbelievable this game is so well balanced.
|
On November 27 2025 17:44 quaristice wrote:Show nested quote +On November 20 2025 21:17 [sc1f]eonzerg wrote:
Any terran player that could explain why Polypoid bottom right is such a good spawn for you ? I think top right and bottom right what they have in common is the same mineral lines. Sim City is different. Is it by being at the right side any beneficial with defense and mobility of your army any different to the left side ? i'm not like a good player and i havent played in like 1.5 years. but theres something about this part of the map right here that just feels a lot easier to use as a terran in the bottom right than in the top right. i don't have a good explanation really. but i feel like it's just a lot easier to defend against any air units? if i place a high ground siege tank it just seems stronger somehow too. if i had scmdraft installed i would try to compare the walkable terrain and see what's going on. or it might be just an emergent property of how a terran base is going to be laid out on the two spawn positions, since buildings dont rotate. i'm not sure
Hey, did you play TvT on Eclipse often?
Top Right of Eclipse has more than 55% winrate against Bottom Left in TvT. Maybe you played TvT on that map when it was on ladder?
EDIT: nevermind, it seems like TvT on ALMOST EVERY MAP seem to favor Terran spawning on the Right quadrant of the map, not just Eclipse. Probably due to the Comsat Station.
|
hm... so why is it so much better to spawn top right for TvT on Eclipse?
images for reference (all players, high MMR):
+ Show Spoiler +
On the other hand, for TvZ its the other way around and Terran prefers the bottom spawn. Probably something related to mutalisks? I wonder how much of this is "I got the better spawn" vs "opponent didn't get best spot for himself". This will be the most relevant on 2-player maps specifically.
Eclipse TvZ:
+ Show Spoiler +
|
On December 04 2025 01:29 Kraekkling wrote:hm... so why is it so much better to spawn top right for TvT on Eclipse? images for reference (all players, high MMR): + Show Spoiler +
I did notice TvT seems to favor right-quadrant spawns on almost every map, not just Eclipse. I guess it's due to the configuration of the Comsat Station, or the Siege Mode starting orientation?
|
from what i remember mineral income for terran on top right eclipse is just better than bottom left, especially with mineral boost, but again, i havent been playing actively and i might be misremembering
from my memory generally right side minerals mine better than left side on most maps, though this depends a lot on the map's pathing, and not just the layout of the minerals
|
On December 04 2025 02:01 Kanzzer wrote:Show nested quote +On December 04 2025 01:29 Kraekkling wrote:hm... so why is it so much better to spawn top right for TvT on Eclipse? images for reference (all players, high MMR): + Show Spoiler + I did notice TvT seems to favor right-quadrant spawns on almost every map, not just Eclipse. I guess it's due to the configuration of the Comsat Station, or the Siege Mode starting orientation?
Tank orientation to a degree. Not sure it is a noticable thing even for pros. I guess occasionally, very rarely can matter a lot, most of the time it's just a tiny thing that adds up to a tiny advatage compared to other positions.
I think comsat being on the mining side is almost always an onconvinience and a reliability.
|
If I recall correctly, Siege Tanks shoot farther in some direction, I believe up? Might be off on that. Anyway, I feel like I have seen it happen in pro matches (one tank outranging another), but I don't recall it having a significant game-changing impact.
|
On January 04 2026 20:00 Jealous wrote:
If I recall correctly, Siege Tanks shoot farther in some direction, I believe up? Might be off on that. Anyway, I feel like I have seen it happen in pro matches (one tank outranging another), but I don't recall it having a significant game-changing impact.
Correct. I don't recall seeing a game where it mattered either.
Out of all things I saw Fantasy's tank missing on Flash's tank on the 1/256 miss chance, and he lost the game because of it.
|
Siege tanks look slightly farther towards the right and bottom direction but that only matters in the early game when there is no rax to give vision. That doesn't explain the better win rates on right side spawns.
From my gut feeling from playing a lot, I would assume the cause is faster mining in the early game before the comsats are up as others have pointed out.
|
On January 05 2026 11:37 Cryoc wrote:
Siege tanks look slightly farther towards the right and bottom direction but that only matters in the early game when there is no rax to give vision. That doesn't explain the better win rates on right side spawns.
From my gut feeling from playing a lot, I would assume the cause is faster mining in the early game before the comsats are up as others have pointed out.
They have tried to optimize mineral patches on most maps with bottom right spawns but they cant entirely compensate for the reduced income. a Terran pro once showed me how everything can be up to 5-8 seconds slower, which is a huge advantage/disadvantage.
|
On December 07 2025 00:36 quaristice wrote:
from what i remember mineral income for terran on top right eclipse is just better than bottom left, especially with mineral boost, but again, i havent been playing actively and i might be misremembering
from my memory generally right side minerals mine better than left side on most maps, though this depends a lot on the map's pathing, and not just the layout of the minerals
I tested it, most patches mine between 167 frames and 176 frames on average on the right side, with the exception of the topmost patch which is almost 200 frames unboosted and the 7th patch which autoboosts 149 frames
a lot of the patches on the left side are about 180 frames unboosted (2,3,4), but it has more "better" patches at 6, 8 that are around 158 frames
the worst is #7 which mines at 190ish frames, which is still better than the right side #1 unboosted
it's basically a wash on the modern versions, unless you boost, but then it becomes "how many patches can you boost at the same time"
|
|
|
|
|
Thank you!
Amazing work. Especially with the game time graphs. Win rate at different game tlmes.
I don’t feel as bad about my ZvT anymore. Seems like other Z struggles as well.
|
|
|
|
|
|
|
|
EPT![[image loading]](https://i.imgur.com/eyPmC2S.jpeg)
![[image loading]](https://i.ibb.co/846JKwY1/spawn-winrate-radial-2p-ALLplayers-tvt-Eclipse-1-3.png)
![[image loading]](https://i.ibb.co/S47YGSX2/spawn-winrate-radial-2p-MMR-above2100-tvt-Eclipse-1-3.png)
![[image loading]](https://i.ibb.co/8LQDqLc7/spawn-winrate-radial-2p-MMR-above2100-zvt-Eclipse-1-3.png)
![[image loading]](https://i.ibb.co/xSsbZZ7d/spawn-winrate-radial-2p-MMR-above2100-tvz-Eclipse-1-3.png)
![[image loading]](https://i.imgur.com/aVaHCje.png)
