Loading...
  OR  Zero-K Name:    Password:   
Title: MM 6074: 1v1, Rank Giant
Host: Nobody
Game version: Zero-K v1.6.9.0
Engine version: 104.0.1-287-gf7b0fcc
Started: 5 years ago
Duration: 44 minutes
Players: 2
Bots: False
Mission: False
Watch Replay Now
Manual download

Team 1
Chance of victory: 44.4%

ITrankBLISS
Team 2
Chance of victory: 55.6%

USranksrcerb
Spectators
AUrankAdminGoogleFrog

Show winners



Preview
Filter:    Player:  
sort
After this game, ITrankBLISS asked me about the purpose of territory, since USranksrcerb seemed to win by investing in overdrive and defenses instead of map control. I contend that territory is important, and that ITrankBLISS was winning due to their superior territory control, but did not manage to capitalise on this advantage and close out the game. I said this to ITrankBLISS and he said I should post an elaboration here. This is that post. I like how this game demonstrates quite a few aspects of territory control, escalation and overdrive, so most players should learn something here.


I don't want to focus too much on the early game, but USranksrcerb seems to be in a better position. ITrankBLISS has their constructors working in their base, whereas USranksrcerb is expanding. The result is an early production lead, and a sizable army, for the early game.


By minute six the economies have equalised. This is due to good raiding by ITrankBLISS and some expansion to the South. USranksrcerb has opted for an early push with Reavers, Knights and Ronin. It is difficult to assault a base this early on this map due to the long ramps leading up to the base. ITrankBLISS defends the push with Rogue and the defenders production advantage. USranksrcerb would likely have been better off taking their army South to clear out the expansion to confine ITrankBLISS to his base.


A lot happens between minutes 6 and 12. ITrankBLISS has beaten back the cloakbot army with some well controlled Rogues, Racketeers, and Snitches. USranksrcerb has raided the bottom left with Glaives, and gets a commander kill. Killing the commander will account for about a metal production disparity of about 7.2k over the course of the game. USranksrcerb also built two Fusion Reactors, but has been using snipers and cloakers, so only some of the extra energy is used for overdrive.

The biggest issue for ITrankBLISS at this stage of the game is that there are eight metal spots within their territory which do not have metal extractors. To be a bit facetious, ITrankBLISS is technically correct in saying that territory is not that important. The important thing is the additional income from the metal spots within the territory. Constructing these extractors would yield a net income of 20 metal, taking into account that two of the spots are being sneakily mined by USranksrcerb. The single largest point of improvement to be made here is the prompt construction of metal extractors. The pair of spots near their main base have never been mined, which is a loss of about 3k metal in income over 12 minutes. This is a significant proportion of the total metal produced so far, which is is 14.4k for ITrankBLISS and 18.7k for USranksrcerb.

There are a few more things to note at this point. The first is that this snapshot demonstrates the diminishing returns of overdrive. ITrankBLISS is using overdrive to convert 25 energy into 5 metal, whereas USranksrcerb is spending an additional 81 energy for a return of only 7 more metal. The difference in overdrive is not large at this stage of the game, even with the two Fusion Reactors. Another thing to note is that both players have similar values of total living units, which means that ITrankBLISS has done very well in terms of attrition.


ITrankBLISS spends the next few minutes pumping Rogues and Rackteers. This initially works well against the Ronin/Glaive composition of USranksrcerb. The later loss in army value is when the Rogues encounter heavy defenses.

The issue here was attack-move. ITrankBLISS could probably have won the game at any point in the last few minutes by just moving their army into the defending units and turrets, since the force disparity was so great. As it was, the Rogues were set to attack-move, which meant that the Rogues tried to engage targets at their maximum range. This allowed USranksrcerb to push the Rogues around by running small forces of fast units at the Rogues. In this screenshot, most of the Rogues are retreating from a Glaive while a small force of Rogues assaults a HLT. Similar things happened throughout the game, resulting in a lack of the type of coordinated assault required to smash the defenses.

Attack-move is most definitely not the "do everything for me so I don't have to look at my army" command. It implements a specific behaviour (retreat to maximum range against targets with less range) which must be used tactically. In many situations your top priority is not for your units to sit a safe range. In this situation the priority should have been to group together and assault USranksrcerb before they were able to make too many defenses. Attack-move was good for attrition, but not good for taking the initiative.


Both players have added a factory. ITrankBLISS picked Rovers, while USranksrcerb picked Planes. I much prefer the Plane factory because air units are very dissimilar to ground units. This forces ITrankBLISS to make some AA and respond to different types of threats. By comparison, the Ripper/Ravager/Scorcher/Fencer composition from the Rover factory is fairly similar to Rogue/Racketeer/Bandit, so USranksrcerb does not have to invest in different types of counters. More generally, your opponent probably has some way of dealing with your unit composition by the midgame, so one of the main purposes of escalation is to present your opponent with different types of challenges. This is what makes aircraft and striders such popular escalation choices; they are rarely seen earlier in the game.

It is also vital to respond rapidly to changes in your opponents composition, lest their temporary advantage snowball into something bigger. When I see enemy aircraft I like to use Alt+Shift to queue 10 mobile AA units at my factory. The Alt modifier inserts the AA at the start of the factory queue, and makes them one-time orders in the case of a factory set to repeat. If you slowly filter AA into your composition you may never have enough to keep up with your opponents planes.

The unusually small size of this map also allows for an additional type of escalation. A Missile Silo in the center of the map can reach anywhere on the map. If you can afford to built and protect a forward Missile Silo it very often ends the game. In this game, an expenditure of about 3k metal in a Silo and missiles would destroy bot Fusion Reactors and most of the Windgens.

At this point I have some advice for USranksrcerb. This game is the perfect situation for Phoenix as you are fighting an army of light units. Also, Swift is not particularly good against ground armies and a single Owl is extremely useful, since it is the best scout in the game. The Snipers were also very good as they resulted in a lot of attrition.


USranksrcerb has a clear advantage by the 30 minute mark. The low-weight armies of Shieldbots and Rovers have been whittled down by Snipers and are unable to push through the dense nest of defenses set up to the East. Glaives are now routinely raiding the side hills, nullifying most of the territory disparity. My last piece of advice is to put a few more resources into defense. Heavy investment in defenses is really not that bad if you use it to lock down 75% of the map. It is also good practice to diffuse constructors throughout your territory, both to respond to threats and reduce the time it takes to rebuild extractors.

This screenshot shows the effect the extended Rogue battles on the terrain and the vehicle-pathing overlay. This is another reason that Rovers were not a good choice.


Merlin is a good idea, as it is artillery and a strider. Unfortunately it is a bit late. By this point USranksrcerb has built up enough Snipers to kill the Merlin before it is able to make a significant dint in their defenses. The Dante was not that good by comparison, as it is bad against defenses and Snipers,but it could still be a good idea if built early enough. USranksrcerb built a Singularity Reactor later in the game, so it is was still possible to win by sniping it. 2.4k metal on a Missile Silo and two Eos would kill the Singu and could be built quite far back from the front line.


The game continues in a similar way until ITrankBLISS is overwhelmed by a massive force of Snipers, supported by Ronin. It may have been possible to come back by attacking up the ramp instead of into the heavy defenses. That said, I'm now at the point of nitpicking things which are only obvious in hindsight.

Here is a summary.
  • Mix constructors into your construction queue and periodically send them out to capture and reinforce your territory.
  • Attack-move is not the best tool for all situations. Sometimes it is better to group and push.
  • Escalate by thinking about how to give your opponent new types of problems to solve.
  • Be faster to react to opponents unit choices. Use the Alt modifier to insert one-time construction at the start of a factory queue.
  • Use slightly more defenses to defend your expansion, especially when you have pushed your opponent back to their main base.
+9 / -0

5 years ago
btw i noticed that badgers and recluses on fight-move are far more suicide-willing then in the past. The "lock" an attack-command and try to follow units into porc... maybe that happened here as well?
+0 / -0

5 years ago
tl;dr
+0 / -0

5 years ago
It seems like recluses aren't as good at hitting things as they used to be.
+0 / -0

5 years ago
Wow googlefrog thank you very much for your detailed analysis, I'm very glad you did it.

I understood that attack move was wrong while I thought I Had to do it to especially for the turrets cause I would engage at maximum range possible but seems like I was wrong. It's not my first time losing despite having more map control , I had my metal extractors raided several times cause I lacked defenses and he was better than me at raiding. Rover was my second factory cause I thought it could give more hp tankines with ravagers and some annoying lines with fencers. It all went wrong I should have used airplanes. or gunships.

I learned that I have to build more constructors to replace raided extractors , that overdrive is not that game changing as I tought it to be (luckily) , that I have to use more turrets near the metal spots , that attack move is not always the best option (in one moment I was managing to break into his base with simple moving but it was too late ) , that I have to play in a more creative way to give my opponent a bigger challenge.
+2 / -0