Loading...
  OR  Zero-K Name:    Password:   

rover - tanks and kodachi

35 posts, 846 views
Post comment
Filter:    Player:  
Page of 2 (35 records)
sort
So, i played some games against ROrankSigero to find out about the new rover-tank matchup.

http://zero-k.info/Battles/Detail/946422
http://zero-k.info/Battles/Detail/946426
http://zero-k.info/Battles/Detail/946428
http://zero-k.info/Battles/Detail/946443
http://zero-k.info/Battles/Detail/946449
http://zero-k.info/Battles/Detail/946451
http://zero-k.info/Battles/Detail/946453
http://zero-k.info/Battles/Detail/946455
http://zero-k.info/Battles/Detail/946458

you can see how kodachi seems really op first, then sigero manages to develop counterplay to the point where i now think that rover has the upper hand in the matchup. next thing to look at is if tanks can counter this counterplay and how your skill affects this. can the counterplay only be pulled of with superiour skill or can even i pull this off against Siger if i play rover and he plays tanks?

i hope this illustrates that i am not against changes but i think testing is done way too short to really gain info with weight.

To be continued!
+6 / -0

4 years ago
The problem with organized testing against a single individual is that there are too many constants (factories are known, playstyles and tendencies become observed etc). That's not to say this wasn't worthwhile though, I'd just like to see a couple more people recruited into the experimentation.
+1 / -0
agree. it`s also the same map wich is further narrowing the scenario down.

i know that this is by no means a perfect test, but it`s what we could do.
involving more people is fine, i just think that the time-dimension is a bit
underrepresented, aka the same players playing a meaningful number of games.
i played 2 games with GF recently and we switched facs after the first game.
that gives useful info, but this info is more "broad", less "deep".
to get "broader" data, everyone is invited to do those games.

Knowing fac beforehand in this case is a bit less of a problem, since i think
tanks start witch a kodachi in almost all cases, as well as it is safe to assume
that rover-players begin with dart/constructor/scorcher.


+ it was a spontanous thing, so i did not think about inviting more people.

+0 / -0

4 years ago
Yeah mate, not trying to tear you down just encourage an increase in scope.

It's funny, it's a really popular testing thing over the past couple of years - Kodachi VS XXX. Always happens. Most contraversial unit?
+1 / -0
This is a much higher standard of balance post than normal with all the replays, nice work. Would welcome more replays by other 1v1 gang for me to watch over my morning breakfast flakes
+3 / -0
I'll admit, the scorcher + mass dart strat didn't feature in any of the test games that I saw during the couple days of testing kodachi balance changes. But it looks like a higher micro requirement to pull off than using kodachis is, which probably is okay even if it ended up being dominant. That leaves more room for mistakes and more room for the tank player to devote time to other tasks. I'll have to see if I can try pulling off the scorchers and darts thing, I am kind of doubtful tbh :D Probably should be noted that any time the kodachis + scorcherdarts are clashing the kodachis are at a significant cost disadvantage, for example in the last video the first clash between the two factories occurs at 340 metal in kodachis vs 520 or so in rover units. It might be that you were just making too many welders. Also when you have multiple welders near the kodachis you're running them away from the fight, when you should probably be trying to use them as a wall to absorb damage and help your kodachis last longer. So I would say the results are definitely not clear that this is a direct counter. But it is certainly valuable testing.
+1 / -0

4 years ago
TAAANKS !
+5 / -0
yeah, for the first few games i am not playing well because i admit that i almost never play tanks in 1v1.
anyway, as it stands right now i think it`s relatively safe to say that a rover-player can choke a tank-player in the very early game on alien desert, given that there is spot-on micro involved. Meanwhile they can naked-expand because the tank-player can`t raid. Change anything in the scenario described above and i don`t claim anything anymore.
Maybe it`s just because ROrankSigero adapts to me faster than i do adapt to him, but at least i couldn`t find a way in and i don`t see many alternatives to a koda-start. I guess i could try an ogre-rush, but that is nothing i would ever do if i don`t know the matchup beforehand, so it seems like a niche-case. As GBrankDregs pointed out, knowing the matchup beforehand is a significant difference that i tried to overcome by using a start that would be considered a "safe and common" option.
+0 / -0
4 years ago
Yeah, I think that's the appropriate way to approach a game like that, try to adapt after you have scouted what they are doing. I wouldn't have started differently either, but I probably would've stuck with just one welder longer and maybe tried to build a blitz in one of those games sometime in the 3-5 kodachis time period. I don't know how strong you are compared to sigero, but that also certainly plays a factor.
+1 / -0
i think my estimated winchance was 2%.
at some point we should get two players of mostly equal skill to repeat the experiment, this would be interesting as well as changing factories as well as changing map and so on.
Anybody any good idea about what should be changed first for the tests to be as relevant as possible? we obviously cant test every possible scenario, not even a mere fraction. so it would be nice if someone had some thoughts about whats "important" to test.
+1 / -0


4 years ago
From my short experience, Scorcher can kite Kodachi and Kodachi can kite Scorcher. Except one of them kites the other by a larger margin and regenerates.

It's less autoresignable than before, but i'm pretty sure tank is favored.
+1 / -0
EErankAdminAnarchid is talking about this game:
http://zero-k.info/Battles/Detail/946977

so what the outcome was is that exactly because of that, tanks can`t really push, neither can rover, but rover has faster and cheaper builders. + they can afford a dart. so the tankplayer is more forced to build defenses on alien desert while the rover-player doesn`t have to fear anything and can naked-expand. ON ALIEN DESERT.

Because you asked me to describe it more detailed: the darts have to surround the first kodachi, as soon as 1 dart just lands 1 hit, the scorcher can kill the kodachi. this is admittedly pretty hard to pull off. As soon as I lost my first unit, Sigero makes use of superiour number and maneuverability to kill my expansion. There are 2 things in play i think:
Rover has cheaper and faster builders. Rover can build darts. So they have the option of very quickly building more units to get map-presence, wich the tank-player has to factor in. on alien desert, the welders alone cannot protect all mexes against darts, because the distance between mexes is too high for welders to control enough space. So you would need pork in the long term. The rover-player doesn`t have that problem. Kodachi is by far not as spammable as dart, so if you can deal with this unit, you don`t have to fear raiding at all.

On IcyShell, i guess this mechanic doesn`t work because the map is so small and the mexes are so close together that a welder is maybe enough to be dart-safe. I guess.

More importantly, you played different to how ROrankSigero against me. In the last games i linked, he focuses on pure expansion and uses the dart/scorcher to block my kodachi. If i lose it, because it got encircled and killed, the
snowball-effect kicks in and tank loses map-control completely. If i do not lose it and pull back, he will still expand faster (?). Your 2 llt`s next to your factory were the direct opposite of the general idea of Sigers play (wich was as "containment-aggressive" and defenseless as possible)
+0 / -0
Looking at replays... Some quick summary from spec's point of view.
Game 1: Early game is sort of back and forth, but midgame rover player has no answer to mass minotaur spam. Eventually mass scorcher overwhelms minotaurs, and tank player resigns, despite having 1K extra army value, 3K extra total value, and better static econ. Psych warfare win=)
Game 2: Rover player takes full map control with 3 darts, tank player focuses on eco. First few kodas make cost, but rover player just spams more units and takes mid decisevely, killing unfinished stinger with ~300 metal in it for free, as well as some other misc units. Tank player resigns again, still with superior eco.
Game 3: Koda sneaks out of tank base into rover base, kills a dart and a scorcher that were stationed to defend (making cost), almost kills a mex (negative micro). Then tank player decides to suicide a bunch of kodas to kill a bunch of scorcher-dart-fencer. Everybody dies but rover player gets to reclaim. Again, tank player enters midgame with superior eco, but nearly no units. GG is called by tank player.
Game 4: Early game is quite even, with kodas trading well vs scorcher-dart mix. Midgame transition for rovers is fencer spam, which deals well with kodas, but at that point tank player has over 30% eco advantage. Eco is barely raided by rover player, which transitions into more kodas than scorchers at some point. By some miracle (tank player suiciding kodas to dive on a com was a factor) the rover player stabilizes, and gets out an impaler and some static def. Tank player meanwhile sets up a massive push with 5 minos + ogre and a blob of kodas shielded by welders. While kodas die to rippers and welder spam gets thinned a bit, minotaur push wrecks face. While the first attack is held, followup is just relentless stream of mino-ogre spam. Rovers have generally no real answer to that anyway, so GG is inevitable. Nice dominatrix spam helps to hold the gates for a while, but then a huge wave of monos just rolls in and wrecks face .... except it retreats instead and gets captured by domis, domis capture EVERYTHING rovers win OMG WHAT DRAMA. Generally, upon seeing the domis tank player should have either switched into arty or gollies or something that has the range to deal with them. Instead he just tried suiciding kodachis which kinda worked, but never well enough to matter. GG is called for tank player. Again, losing with superior eco at time of resign.
Game 5: tank player plays defensively and dies to attrition with dart-scorcher and overall depression. very sad. Arguably this had more to do with micro mistakes from tank player rather than anything else.
Game 6: Very first koda gets caught with pants down by 2 darts and scorcher, dies. A lucky scorch gets into tank base and wrecks a welder there. GG is called.
Game 7: lotus and 2 welders die to scorcher and 2 darts. GG.
Game 8: Tank player rushes ogre and wins. except he does not, instead he just gets aggro with some kodachis. Midgame starts with rover player having a substantial income advantage, and few fencers on the field. Kodas split to sneak through def, get caught by dart-scorcher and die. GG.
Game 9: rover player applies relentless pressure, denying tank player any and all chance to leave the base, while naked-expanding behind. Economic win for rover player.

Game 10 (with Anarchid): Pregame - rover player picks a much better start location, getting instantly 2 mex advantage and easy defense in the corner. In early game a scorcher misses an opportunity to dive on tank factory, which would have been total GG for tank player. Despite this, first blood is drawn by koda 1v1 killing a scorcher. Second few darts and scorchers proceed to die to kodas also. Rover player chooses to go for a com push (guardian com drones are nice for keeping kodas from regenning too much), fencers also help a bit. But then an ogre pops... Rover player kills the ogre, but loses com and 3 lotuses, as well as few fencers. In tank's player base... At this point tank player has massive eco advantage, and GG is inevitable. Overall it felt like the ogre killing com was the deciding factor, if rover player was a bit more careful he could take the game. Also the fact that rover player (despite way lower rank) had a noticeable advantage in the game for quite some time sort of shows that rovers are probably easier to work than tanks.

Conclusion 1: tank vs rover matchup is fairly even, slightly favoring the rovers, and mostly depends on players decisions rather than unit qualities. Kodachi trades well vs scorchers if there are >3 kodachis, but in low numbers dies to fencer spam or bad luck. 3+ kodachis are actually formidable and generally win most fights. Fencer + ripper was tried, and worked quite well. Darts are critically important to keep tabs on where the kodas are. I'd argue that tank play in 1v1 is a hit-and-miss sort of thing, and not for the weak in micro.

Conclusion 2: need to try dart/scorcher vs minotaurs and other heavies more, looks like its fun. Overall seems folks underestimate darts in straight up combat. Also I think ripper vs koda was not really explored well. Spamming domis vs mass tank also did not look like a good strat for e.g. a team game.

Overall I feel like the whole koda vs scorcher matchup is a bit strange, and in some crazy way favors the rover player pretty hard. Maybe ppl just need more experience doing it in 1v1?
+2 / -0


4 years ago
quote:
blitz just destroys scorchers (not sure if darts are mixed in)
Mono blitz is insanely annihilated by a dart-heavy rover mix. As in, you won't even have a chance to run, Darts are faster.

Adding Kodachis changes this, of course.
+1 / -0
"Overall I feel like the whole koda vs scorcher matchup is a bit strange, and in some crazy way favors the rover player pretty hard."

this is the conclusion of what i draw from games 6-9.

At game 6 Siger said he now knows what to do. Wrecks the koda.

Game 7 is me trying to counterplay that by building well, just builders because i want to try if welders are maybe able to hold on their own.

Game 8 is DErankTopkack telling me to try it with splitting-up the kodachis

Game 9 is ROrankSigero having perfected the execution of his strategy. this is the game i base my opinion on most.


i would really like to see if there is anyone who can make tanks work against the rovers from game 9...
i know my micro is pretty bad, but with 1 kodachi i do not see any real chance to make it work.

FIrank[GBC]HeadHunter ofc player-decisions matter more than unit-qualities, and i think i agree with your view of the matchup overall. but i would add that if the rover-player has spot-on micro and positioning in the very early game, tank is absolutely chanceless.
+0 / -0

4 years ago
DErankkatastrophe, I agree that if both players play perfectly (500 APM micro) tanks have no chance whatsoever. But in such example, maybe it would also be true that glaives would wreck everything? =)
+0 / -0
It's an interesting set of games, particularly for all the interactions between Kodachi, Scorcher, Dart and Fencer. I'd say the greatest improvement would be to have ROrankSigero and DErankkatastrophe swap sides occasionally. This is a good way to check whether ROrankSigero is learning how to beat Tank with Rover, or whether he is just learning how to beat DErankkatastrophe.

I think playing the same map and matchup is completely fine. The aim of this sort of testing is often to check whether a matchup has a problem, not to somehow prove that it doesn't. A matchup has a problem if it plays poorly on any reasonably generic map, so we only need to care about one map. Sure, it is useful to know where the matchup works and where it doesn't, but for most testing a single map is fine.

Playing the same matchup is also fine. Some unrealistic build specialisation can appear as a result of it, but that is often not too hard to correct for in further testing. A matchup is still a problem if a specific blind rush or defense is required to avoid degenerate behaviour.

I wrote a few notes about the games.
[Spoiler]

My main conclusion from the games is that the matchup appears to be working so far. I think ROrankSigero was adapting to DErankkatastrophe's attempts to force aggressive Kodachi play, and the aggressive play did not seem too powerful here. If it turns out that Rover can handily beat Tank while Rover is making three unit types and Tank is only making Kodachi, then perhaps this is a reasonable starting point and it is time to work on bring Blitz into the interactions. I mostly think that we don't have the breadth of data to say much more than 'seems fine so far'.

DErankkatastrophe since you singled out Game 9 I took a closer look. What struck me is how Tank spends relatively little metal in Kodachi to tie up the Rover forces for a long time. I think Tank did pretty well to do this, but then did not use the advantage to make a better economy. Here is how the game looks at 3:00.

By 3:00 the Rover player has only seen three Kodachis, having not scouted any structures at all. Another Kodachi is about to appear in two seconds. I marked when and where the Kodachis appear on the map and played out an opening which hits those timings. My economy is a bit wobbly because I avoided using Wind and hitting the timings was a little awkward. Here is the result.

To be clear, the Rover player could not know whether they were playing against the game that happened or my expansion pattern. The radar coverage you see in the screenshot of the actual game is as far as has been scouted. To me this indicates that there is more to explore in this matchup.

I don't know if much more can be said about EErankAdminAnarchid's game. To me it looked like he gave up map control by starting in the corner, making two turrets and leaving his commander idle when it could have been pushing in. There is no point trying to catch out the Kodachi with Scorcher+Dart when there is no viable target for the Kodachi. On a map that small I would have liked to see more Ripper.
+3 / -0
let`see if i can recreate that as well later...

that is my attempt to recreate what you were supposing AUrankAdminGoogleFrog
http://zero-k.info/Battles/Detail/947422


then there are some replays nimor vs sigero:
http://zero-k.info/Battles/Detail/947423
http://zero-k.info/Battles/Detail/947431

nimor vs. topkack:
http://zero-k.info/Battles/Detail/947432

nimor vs sigero:
http://zero-k.info/Battles/Detail/947445

will write more soon
+1 / -0
4 years ago
Mmh, I think the impression I'm getting here is that Sigero is overpowered in this matchup. Even when Sigero makes mistakes, no one manages to really capitalize on it, whereas Sigero is winning games while being significantly behind at various points. I do think the scorcher+dart or scorcher+fencer combos take more skill/micro to effectively employ than what the tank player needs in the matchup, but Sigero is looking almost stronger playing as tanks vs playing as rovers. I'm going to say it looks evenish to me right now, but I've only played one game vs tanks since the patch so I am not sure exactly how I feel about the way the matchup plays out yet.
+0 / -0
because the question arose how hard the dart - scorcher thing is to pull off...
here are my laughable attempts:

http://zero-k.info/Battles/Detail/947447
http://zero-k.info/Battles/Detail/947450
http://zero-k.info/Battles/Detail/947462

conclusion: i fail miserably exept once iirc.

+0 / -0
Page of 2 (35 records)