Banner and Crosspost

Home    Overlay   Discord   Mutators   Maps   Integration   Links   About

November 10, 2020

Transmission 02 - Player onboarding and lowering the skill floor

Transmission 02

In the previous post I wrote about settings, scale, heroes, social aspect, shareability. This post will focus on lowering the skill floor and player onboarding. Together with player socialization I see these as the most important areas where a new RTS can improve.

This is a good opportunity to organize my thoughts. Following ideas might or might not fit a particular RTS game, but perhaps some will be useful or shape further discussions.

Onboarding & skill floor

RTS games are typically hard to get into and have a steep learning curve. There are a lot of things to learn and master before players can "play the real game", "engage in the core experience" or however you want to call that. In first-person-shooters you move and click enemies, in MOBAs you start with a single unit and gradually unlock abilities. That's a lot easier than making sense of UI and learning how basic economy, tech tree, production, and unit control works.

Even advanced players might not truly engage with the core experience until higher ranks. In StarCraft games macro-mechanics are so important that improving them might be your best way to advance to the next league instead of actively scouting, learning proper responses or controlling your army better.

Even the basic base management is hard in Brood War

This has led many games to oversimplify either micro or macro mechanics or both. However, I think that with good design it's possible to make a good game for a wide variety of players. Solutions will have to come from all angles – UI, UX, game design, art & sound design. It's important to consider players of skill levels when making any change. What exactly are players doing, and how engaging is the core gameplay loop? If we automate something, does the game become shallow or less engaging for them? Does the change reduce skill ceiling where we don't want it to?


This concept was mentioned on the Pylon Show. It refers to the idea that an army should be somewhat effective even without too much micro involved (80% effectiveness). The actual numbers might be a nod to pareto principle, but they are not important. In some cases additional effort expended will lead to smaller improvements (95–5), and sometimes to much larger ones (50–50). The core idea is to lower the skill floor of controlling an army.

An example that was mentioned is Nova in StarCraft II Co-op. She has a solid army that does "ok" when simply a-moved. This is not uncommon in Co-op. The great part is how much more it can do if you control it properly – siege Liberators and tanks, use Raven's abilities, lay mines, micro Nova, EMP with Ghosts, place defensive drones and use airstrikes. This broad range of options provides tactical choices and rewards micro. A new player might be satisfied with simply a-moving and using an airstrike, more advanced players will try to do as many things and as good as they can.

Nova's wide variety of units

Another good part is that many abilities can be "pre-cast". What I mean is that a players can siege Tanks and Liberators, place Auto-Turrets, Defensive Drones, and lay Spider Mines in anticipation of the engagement. This lets slower players to engage with these abilities even if they couldn't use them when the combat begins.


One of the approaches mentioned was a more common use of autocast on abilities. The idea is that slower players will leave them on autocast, while faster players might disable autocast and/or use them manually and more effectively. Results are:

  • Lower skill floor
  • Fights on lower levels look more like those on top levels
  • Players encounter autocasted abilities more at lower skill levels

There are some potential disadvantages as well:

  • Autocasted ability is significantly less rewarding than manually casted.
  • Too many autocasted abilities might result in unnecessary visual clutter.
  • If the autocast is too smart, the player won't engage in decision-making where and when to use the ability until the player can outperform autocast and overcome the opportunity cost of spending actions somewhere else.

★ ★ ★

I've looked for some good examples in StarCraft II Co-op. Ambusher's Blink might be the best one. Ambushers autocast blink to escape enemy fire, but players are encouraged to use Blink more aggressively as it provides significant DPS boost. This works very well because the manual cast is used differently than the autocast (defensive vs offensive usage).

Avatar of Essence's Devolution Wave is not as good example. It's an autocasted area-of-effect debuff. Manually casting it has exactly the same use, but you can affect more enemies if you use it at the right time. I would guess that more than 99% of playerbase has never casted it manually. Vorazuns's Corsairs, Zeratul's Shieldguards and Abrogators are in a similar position. Their abilities have the same use whether autocasted or not, and are mostly not rewarding enough to use them manually.

Swarm Host-like units are a better example. They do spawn Locust-like units automatically, but with manual cast you can send them to hit targets outside their autocast range, or you can spawn them in anticipation of the enemy attack. This option to pre-cast and different manual use case are what makes this interesting and rewarding, even if most players will not actually use it.

Ambusher's Blink can be used both defensively and offensively

I think autocast on abilities work particularly well when the use case for manual casting is sufficiently different from the autocasted ability – as is it with Ambushers and Swarm Hosts-like units. Warcraft 3 has good examples as well, namely Sorceress' Slow and Dryad's Abolish Magic. With these abilities you are not trying to compete with autocast AI, instead you might use them when the autocast wouldn't trigger at all (e.g., against summons or when chasing the enemy), or target the high-priority units first.

Disabling and enabling autocast to preserve energy is a skill in itself. However, if players are encouraged to change the autocast state too often, it can quickly become a chore.

★ ★ ★

Two more thoughts about autocast:

Using abilities in fights is engaging, and players should be encouraged to do it manually. It might be a good idea to first look at macro mechanics to lower the load on new players, then try to improve the user experience when controlling armies. For example having all abilities in one command card for Tychus does reduce the mechanical barrier to using active abilities. Only after that it might make sense to look at how many autocasted abilities are actually needed.

There was an idea that autocasted abilities would have different statistics. This seems highly unintuitive and inelegant. I don't want to see a fight where some Psionic Storms deal 80 damage and others 50 damage. It's better to make the autocast AI less good – doesn't react immediately, stacks unstackable effects a bit, requires certain game state, doesn't hit the most targets with one spell or the most important targets, etc.

Simple unit control

This returns back to the 80–20 principle, but instead of increasing effectiveness through manually casting abilities, the effectiveness can be increased incrementally with focus on simple control based around moving and attacking. My thinking here is that you can more naturally improve your army control if it's based around simple concepts like moving, attacking or target firing.

This has few advantages:

  • Effectiveness is improved more incrementally.
  • Resulting battles are easier to parse – fewer effects, no knowledge of abilities required.

More focus is put on kiting, and army positioning – arcs and surrounds. Brood War did a good job by distinguishing units and interaction based on movement alone. Vultures or Mutalisks can shine with just movement micro. Warcraft 3 also heavily focuses on army movement and positioning.

Unit positioning is extremely important in Warcraft 3

These simple unit controls could include positional units like Lurkers, Siege Tanks, Liberators or heavy machine guns or cannons in Company of Heroes with their limited firing arc and setup time. All these can be very interesting and scale well with player skill.

Macro mechanics

The 80–20 concept could be the most useful with macro mechanics (economy, basebuilding and production). Most of new players will want only very few tech and economy decisions before going to play with the army. But I think that for experienced and competitive players there should be enough room to improve macro through high APM and multitasking.

In the previous post I mentioned that it's hard for me to return to competitive 1v1 due to not being satisfied with my mechanics. I believe applying this 80–20 concept could help with that. It wouldn't be helpful to just new players but also to anyone returning to the game.

★ ★ ★

How to do this correctly is a very difficult problem and will require a lot of prototyping and testing. StarCraft II's Chronoboost, MULE and Larva Inject work, but they can feel a bit forced. They do provide some room for decisions, and the inject raises skill ceiling as well.

Few other examples how to introduce some optional complexity to macro: adjacency bonuses in Supreme Commander, more efficient manual reseeding of farms in Age of Empires, transferring workers, overbuilding workers before transferring them to an expansion, collecting scattered resources or wrecks from previous battles, creep spread, switching add-ons in StarCraft II, updating resource drop-off points, workers returning to work, and more.

In Supreme Commander adjacency bonuses reduce operating cost

It would help casual players a lot if repeating tasks were reduced, and macro was limited only to important decisions. There is nothing worse than watching a casual player making 20 Pylons without using the shift key. This could include limited auto-queues for workers, auto-hotkey for some production, etc. Repeating tasks can be reintroduced for more serious players, and as a player progresses through ranks these tasks would become more important. This is of course more easily said than done.

More things to help

Now for some other things that might help with lowering the skill floor and onboarding. They might or might not fit into a particular game.

  • Already mentioned ghost mode would be great. However, it should not become a crutch for overly complex buildorders.
  • Hotkeys for select all army / all units onscreen / all army units not in control groups.
  • Global build UI similar to found in C&C games or Spellforce.
  • Shared ability command card similar to what Tychus have (no need to switch between units when using different abilities).
  • Easier default commanders in Co-op compared to StarCraft II where default commanders are not very casual friendly – especially Raynor.
  • Better after game feedback. In RTS games it's often difficult to see what you did wrong. It would be even more difficult to find this algorithmically, but perhaps some machine learning model could manage it decently.
  • Reducing the number of cheeses and rushes in lower leagues. Possible options could include a different take on stealth than in StarCraft games, free scout like in Age of Empire games, or something like Orc Burrow mechanic in Warcraft 3. It wouldn't completely prevent aggression, but the damage would be limited to not mining. Something like that could keep early aggression viable but not game ending unless one player messes up.
  • Good and engaging tutorials. They could be combined with a challenge providing bonus experience, other rewards and even leaderboards. They could focus on a variety of skills from basics to those aimed at the competitive mode.


I mentioned a lot of things, but I'm sure there are plenty of other ideas that could improve the experience for new and returning players. It's a difficult problem and solutions will have to touch and affect many aspects of the game. Thank you for reading. For discussion check this thread on r/FrostGiant.

Recent posts



Powered by Blogger

Main post