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

Priority of learnings for 1v1 in ZK

28 posts, 1447 views
Post comment
Filter:    Player:  
Page of 2 (28 records)
sort
Any player of ZK who is seeking to improve is faced with a lot of possible avenues of learning. Not all of them are equal. When they look to more experienced players, there are still a lot of bad habits that are being shown to them that aren't worth replicating. Some are genuinely bad play because even the best players of ZK aren't in a position of expert level authority. Some are signs of advanced play because they're only worth doing at a high level of play. Some play such a large part of any gameplan that they are indispensable skills that should be prioritised immediately.

So I'd like to, as a community, come up with a list of priorities for the prospective pro. From zero to one hundred, describe and justify operations that help you win, and what level of importance they are. One hundred being the highest priority, one being the lowest, and zero being a non-priority. Feel free to disagree etc., I'm not sure how we'll decide who's right. Maybe we'll end up with different lists?

99: Ensuring you have enough energy to spend metal
98: Ensuring you have enough BP to spend metal
96: Taking mexes that are in your territory
- area mex tool
95: Taking additional territory
90: Controlling the map if possible
90: Scouting if you have the opportunity
90: Retreats micro
80: Splitting the map in your favour
80: Optimising your composition to better compare with your opponent's
75: Raiding if possible
75: Escalating your composition past raider when appropriate
75: Not committing to losing battles unecessarily
74: Consolidating captured territory
70: Expanding or constantly building with commander
70: Maintaining constructor productivity
70: Not trickling units (make sure they're together when they need to be)
70: Never leaving factory idle accidentally (otherwise known as using factory repeat)
- alt clicking units to manipulate repeat queue
69: Controlling large wreck fields
68: Upgrading commander so that it has a weapon
65: Taking advantage of strong wind on maps that have it
65: Adding an additional factory in the mid-game
65: Positioning commander in best position to be useful without losing him in 90% if cases
60: Coordinating air and ground units simultaneously
55: Line moving units
45: Taking advantage of mountains for wind
45: Building geo-thermal plants when appropriate
43: Using terraform appropriately (to be expanded)
40: Building moho when appropriate
35: Aggressive raider micro
30: Not building too much energy
20: Set target and target priorities
15: Building fusion reactors when appropriate
8: Building Singus when appropriate
5: Upgrading commander past the point of having a second weapon
2: Building storage
2: Building more than two static defenses at the start of the game
0: Building ramps with which to launch roaches at opponents

First draft:
[Spoiler]

Criteria should include:
- what proportion of games it benefits you
- how big a benefit it is
- the likelihood of you having the information to correctly commit to this decision
- how much effort it is to learn
- how much effort it is to execute
- other stuff I can't think of now

So criticise, or add, or ask questions, or just give thoughts plz.
+4 / -0

7 years ago
teach the AI the list and we are doomed?
+1 / -0

7 years ago
Raider (and general unit) micro rates somewhere on this list.

I would rank Fusion considerably above Singu and good wind above Geo in 1v1.

[Spoiler]
+1 / -0
7 years ago
Some tasks that have high use are quite complex and depend on mastering basic tasks.
For instance: I am surprised to see "controlling wreck fields" as the highest priority. This is something I would not teach a player before he can manage eco and units etc...
Highest prio would be: early game play: plop fac, grab mexes and set up some E, send raiders/scouts.
(Or even more important: get to know the user interface)
+1 / -0

7 years ago
BErankFlipstip
Hmm good point. I put it there because when there is a large wreck field, it's quite literally the case that the payoff for having units that then control that field trumps even the raw efficiency of building mex. That's a pretty specific criteria to put it above everything else though, so yeah I'd lower it's priority.

AUrankAdminAquanim
I'd put retreat micro very high, but other kinds of micro substantially less.

When I referred to wind I meant specifically when you move up a mountain to build wind, not when it's just good. If it's just a strong wind map then yeah, it's a very good thing to do. I honestly don't see much value in fusions at all in 1v1, past lowering APM and space needs. It's only 20% more efficient than solar while requiring a very large commital of resources to OD, not linking your grid, not addressing shortages, and providing your opponent with a point of weakness to target. Most new players would be better off if it didn't exist as an option for them IMO. Comparatively, singu is efficient enough to warrant the expenditure, even though the opportunities to build it are a lot less common.
+0 / -0


7 years ago
Things to add somewhere:
  • Not building too much energy.
  • Defending territory.
  • Using terraform, somewhere around 45 based on that list.
  • Not feeding armies.
  • Stuff about air usage/response.

Also there could be a lot about using the interface. I am not sure whether that is the intention of the list:
  • Repeat for factories.
  • Queue handling (Alt etc..)
  • Area Mex
  • Line move
  • Fight and tactical AI. When to use it and what it does. I think some people think it is supposed to implement "be smart" but the tactical AI can't read minds. All it does is implement a particular behaviour and the behaviour is not always appropriate.
  • Set target and target priorities.
  • Terraform.
  • Checking for idle workers (may be in the list above?).
  • Area commands.
  • Retreat.
  • Overkill prevention. This is similar to tactical AI. I often see Racketeer fail to keep something disarmed because it is set to be efficient with its shots.
  • Etc...

There is also a whole bunch of knowledge:
  • Weapon types and shapes.
  • EMP, Slow, etc... interactions.
  • Armour.
  • Overdrive.
  • Etc...
+1 / -0
7 years ago
I think the format would be a scheme with arrows showing dependencies etc.
You need low level tactics, rps knowledge, positioning, metal/energy/BP balancing skills,... Before you can learn how to take control of a certain point/situation, as well as the ability to recognize these situations.
Each task can be attributed a learning cost (including cost of previous steps) and a gain value. Then you can sketch some kind of "learning network"
+1 / -0

7 years ago
Yes those things can and should be on the list Googlefrog. Preferably operationalised. I'll wait a bit and then do another draft, maybe a 24 hour update if I'm in a position to do so.

Flipstip I was thinking about the dependency thing (specifically scouting and what you do with that information). I think I'd prefer to work it in a simple priority order if possible, just for simplicity. Any noobster can pick up a list and pay attention to the things that are on top of it. Flowcharts etc. are great but I'd want them as an addition, not replacement.The medium is important. I want people to run their eyes down a large number of claims, spot the ones that they don't agree with, and think 'wait, why is that there?'. That's how learning do.
+0 / -0

7 years ago
BErankFlipstip that sounds like you're planning a MOOC.
It can probably work with short 5 min. youtube videos or something similar.
+0 / -0
best things to learn for ZK:
- expand
- reclaim
- don't feed metal
- value defenses but don't ovebuild them
- make enough energy
- spend your resources
- repair
+1 / -0

7 years ago
-spam storages till enemy resigns out of dispair.
+0 / -0
Re: fusion - I would only plan on making fusion in 1v1 on a map like Comet Catcher, which is very large and has a lot of metal; the time your cons would spend walking around setting up solar chains, and the additional investment required to protect those chains, make fusion look more appealing. (It also fixes e-stall, which is particularly prevalent on Catcher with its high metal-per-mex.)

I'm not sure I have ever played a 1v1 where singu was practical. No more than one or two at the outside. Maybe on some silly (for 1v1) map like Deserted Gully.
+0 / -0


7 years ago
A couple of advanced things:

-Keep retaking and rebuilding mexes. Mexes only take 30s to pay for themselves on most maps, so even if it seems contested, it's worth rebuilding.
-Move the commander as little as necessary, because it is slow and small losses of time add up.
+0 / -0
Okay I'm going to adapt the list a bit, taking into account community input.

I'm narrowing down the criteria so that operations are phrased in such a way that they are a behaviour of the player that can be contingent on gamestate. For instance, controlling large wreck fields is a very high priority, but players that prioritise them highly aren't going to be as strong as players that expand for instance. While controlling wreck fields is often a higher priority, when thought of in terms of which habits win games, expanding is a much higher priority habit to learn since it impacts every game from start until finish.

I'll keep a spoilered version of each iteration of the list in the OP (so that the discussion makes sense), but only the current version will be unspoilered.

99: Ensuring you have enough energy to spend metal
98: Ensuring you have enough BP to spend metal
96: Taking mexes that are in your territory
- area mex tool
95: Taking additional territory
90: Controlling the map if possible
90: Scouting if you have the opportunity
90: Retreats micro
80: Splitting the map in your favour
80: Optimising your composition to better compare with your opponent's
75: Raiding if possible
75: Escalating your composition past raider when appropriate
75: Not committing to losing battles unecessarily
74: Consolidating captured territory
70: Expanding or constantly building with commander
70: Maintaining constructor productivity
70: Not trickling units (make sure they're together when they need to be)
70: Never leaving factory idle accidentally (otherwise known as using factory repeat)
- alt clicking units to manipulate repeat queue
69: Controlling large wreck fields
68: Upgrading commander so that it has a weapon
65: Taking advantage of strong wind on maps that have it
65: Adding an additional factory in the mid-game
65: Positioning commander in best position to be useful without losing him in 90% if cases
60: Coordinating air and ground units simultaneously
55: Line moving units
45: Taking advantage of mountains for wind
45: Building geo-thermal plants when appropriate
43: Using terraform appropriately (to be expanded)
40: Building moho when appropriate
35: Aggressive raider micro
30: Not building too much energy
20: Set target and target priorities
15: Building fusion reactors when appropriate
8: Building Singus when appropriate
5: Upgrading commander past the point of having a second weapon
2: Building storage
2: Building more than two static defenses at the start of the game
0: Building ramps with which to launch roaches at opponents


To be added:
- Area commands.
- Overkill prevention. This is similar to tactical AI. I often see Racketeer fail to keep something disarmed because it is set to be efficient with its shots.
- knowledge of states, armor, overdrive etc.
-Keep retaking and rebuilding mexes. Mexes only take 30s to pay for themselves on most maps, so even if it seems contested, it's worth rebuilding.
-Move the commander as little as necessary, because it is slow and small losses of time add up.
- expand
- reclaim
- value defenses but don't ovebuild them
- make enough energy
- repair


I don't really know how to approach adding these. I moved things around a bit. I expect to move things around a lot more, and to never achieve anything that everyone is truly happy with. But so long as the broad idea is refined then I think we'll end up with something useful. I'm thinking I want to move a lot of things down further still. ATM it's ordinal only, but that's not obvious. People could easily assume that line moving is more than 50% as important as expanding for instance, when it's really not. But I'll focus on getting them in order first, then perhaps think about representing them more in ratio.
+0 / -0
7 years ago
To be added:

- Use Radar
+1 / -0

7 years ago
Wow how did I miss that? Jeeeez.
+1 / -0

7 years ago
lol. i use default radar on 1st morph. only skythes are allowed to surprise me.
+1 / -0


7 years ago
Why is it 100 points. It is a bit TLDR. Is your intention that players focus on the top point (that they have not mastered) and by the end they will have high skill?
+2 / -0


7 years ago
AUrankSnuggleBass: You are making your life too difficult in an easily fixed way. Don't use a numbered list, instead group it into unordered sublists going from most basic to most advanced. That way you don't have to debate the exact positions of things, and it's easier to approach since someone reading through it can just focus on the group that is relevant to their current level of experience.
+6 / -0

7 years ago
Dammit I think we've had this conversation before and I agreed with you then while also failing to learn from it.

Something to do with tier lists.
+0 / -0
Page of 2 (28 records)