Widgets in competitive games

Save me some time and read https://zero-k.info/Forum/Thread/31189?postID=219329#219329 as it is mostly what I'd be writing here. In retrospect I came across a bit harsh there and we may even have had a player widget turn into a game widget in the year and a half since that post. Prefire was eventually solved by a combination of redesigning Kodachi and esainane digging up (and fixing in-engine) an obscure weapon tag that let units fire towards where a unit is going to be provided that it will be in range when the projectile hits. These changes removed most of the utility of the widget, making it obsolete.

The main thing I want to highlight from the post I linked is that

quote: My ideal world would have a list of other people's enabled widgets available at the end (and start?) of a match so that you can just grab whichever you like the look of. Allowing a rapid evolution of widgets as people share the good ones and improve/tweak them over time.

quote: Hot take: all allowed widgets should be publically available and the wiki should inform what they are and how to get them.

does not work. There is a massive gulf between "widget I can use personally as it fits my style and I know its quirks" and "widget that is user friendly enough for others".

quote: Since nobody mentioned the elephant in the room, biggest problem is that the widget enthusiasts are also ones most likely to be (potential) contributors to game development. So pissing them off by removing their game advantages will hurt the game more in the long run than some unfair matches. A kind of realpolitik opinion on why it is tolerated.

Yes. Maybe not the biggest problem, but it is there.

quote: One feature/design paradigm I particularly like about zero-k is that it arrives to have a "physics based combat model"; projectiles obey laws of physics, units don't have hidden bonus damages or armor types, and balance is achieved through tweaking these physical properties rather than artificially handicapping units. The prefire widget was a natural use of in-game physics to improve unit performance, and a physically accurate one at that. I would argue that the net result of the widget and rebalance is that the game is a more accurate combat simulation, which is something I value. The net result in terms of gameplay may be the same but there is value in staying true to the principles of the game's design.

Yes, this is good. Who is going to do this work though?

I made this ticket https://github.com/ZeroK-RTS/Zero-K/issues/4479 rather than weigh in on the Badger replay because it was "ruined" by the second and third (especially the third) posts. Trying to talk about anything reasonable on that thread seemed like a waste of time. It turns out I had already made a similar ticket a year and a half ago for Kodachi prefire. It seems like exactly the sort of non-specific additional command that PRO_Dregs can appreciate.

I'll have another go at picking apart the difference between player widgets and game UI. Well, apart from the fairness difference, which I think has been well covered. To start with, the UI is meant to be used by players to implement their decisions, not make decisions for them. The issue is that the act of creating anything involves making a bunch of decisions along the way. In making this comparison I may have conflated "creative/design decisions" and "ingame player/strategic decisions", but I feel like the natural mode of widget creation is to conflate these two types of decisions. Consciously keeping the design decisions within the widget separate from the strategic decisions of those using the widget is simply a lot more difficult than bundling all the technical and strategic considerations together, and is redundant effort when making a widget designed for personal use. The default is to, subconsciously, hardcode widgets with the types of decisions the creator tends to make, and to mesh well with their playstyle. Everything the widget does is in essence a real strategic decision for the creator because they can tweak or rework anything they want. Other players don't have this ability and it can't be expected of them. They'll not use the widget or be stuck having their units implement someone else's decisions. Letting a powerful widget of this nature dictate balance and the rest of the design puts everyone else in a tricky situation. Many people will either be left behind, or use the widget but noticeably degrade their UI in the process.

Imagine a master cobbler who shows up to a race with a box of magic shoes that double the running speed of the wearer. Unfortunately he only knows how to make them in one size - his own. Others can wear them with some greater or lesser degree of discomfort, and they don't fit some people at all. Having a whole box of such shoes doesn't make the race fair, and even many who can wear the shoes, but quite uncomfortably, would probably rather not have to.

In short, the naive unit AI widget decides that units should behave a particular way and implements it. The advanced version, the type that is polished enough for inclusion, gives players tools that let them decide how their units should act. For someone to really be deciding how their units should act they need to understand what the commands do, and the commands need to be obvious/easy enough to use. This is where all the polish and streamlining and UI design work comes in.

Some sort of technical ban on local widgets would make development harder. Eg PRO_Clopse was used local widgets, ones that issue commands, for a while. He would make some feature request, I'd see that it isn't in the options but could be with a small tweak, so I'd make the tweak and send him the widget. The feedback and revision cycle was tight. I'm also not so sure that the problem is restricted to the 1v1 MM. It might just be the canary in the coal mine - some widget that cause complaints in 1v1 might be noticed in team games a month later. Delaying this complaint isn't overall helpful to development. Regardless, I don't want to maintain two sets of balance for it.

Zero-K has been designed along the lines of what Niarteloc said above, and while it takes increasingly more work for smaller gains, I'd hope we can continue to do so. However, the work largely involves things like resolving conceptual paradoxes, such as firing at the ground being mostly free, and implementing generic commands/behaviours, rather than making specialised widgets.

One final example which someone might pick up. Shaman had a widget that could do Snitch bombing. I recall asking for it as a generic attack command for transports that tells them to fling their cargo to land at a location. This hasn't happened yet.

quote: But as PRO_rANDY said it's not the same as widgets that make units OP.

That's not a distinction that's possible to make.
Every one of the things I mentioned (and indeed, every UI or unit AI change ever made) makes units stronger or weaker. Scattering makes Likho less good, even distribution line move makes Aspis stronger, Glaives auto-retreating from Reavers makes Glaives stronger, the F-alt command lets your badgers easily lay a minefield (though most people aren't aware this exists), etc.

GoogleFrog:

quote: does not work. There is a massive gulf between "widget I can use personally as it fits my style and I know its quirks" and "widget that is user friendly enough for others".

It solves the "unfair" problem, which seems to be the main one people are unhappy about.

Also, I don't think this is any different to the current situation. A huge number of the default-installed widgets are finnickety in some manner, have surprising behaviour or interactions and don't always do what you might want:
* Auto Jump Over Terrain often lands your jumpcom in hot water.
* The Recluse unit AI will have it walking right into that Stinger.
* The Phantom unit AI will have it stopping outside of visual range when attack-moving vs a radar dot indefinitely rather than moving in for the kill.
* A Cyclops will sit next to that raised Stinger plinking away at it with the slow beam until it dies.
* A Minotaur will attempt to kite a circling Ravager, repeatedly swinging its slow turret away from its target and rarely firing.
* A Reaver will happily keep walking right into Rogue shots.
* Detriment on attack move used to sit at missile range plinking away.
* Badgers laying minefields won't target units that come into range or try and kite them.
* GBC is a thing.

Announcing widget names or descriptions is pointless because they can easily be changed. "Your enemy is using widget 1 and widget 2 with the description Get git and git gud." Full transparency can only be reached by publishing the full source code.

First of all, widgets add a lot of value to the game as it has been demonstrated by many examples here. You always have to ask yourself the question: If we already had access to all custom widgets, would we want to remove them again? Just like the widgets that we already have, you probably wouldn't want to miss most of them.

I also think that it's important to have a free space for testing widgets in casual games as dyth68 pointed out. This makes it more comfortable for widget devs to test things and does not spam a potential widget database with half finished code.

For MM games, I think there could be a widget archive where the source code of each widget used in MM is stored with a unique widget ID and published with the battle replay.[Spoiler]
If people agree on this, someone who puts in the time and effort to implement such an archive is still needed. For me personally, the effort is not worth it as I can anyway publish my widgets and replicate others. But it would be interesting to see what others use. So if you like it, go for it.

quote: Badgers laying minefields won't target units that come into range or try and kite them.

Unless you use my widget.

quote: It solves the "unfair" problem, which seems to be the main one people are unhappy about.

dyth68 I've sick of this ridiculous statement. Either interact with the actual points that people are making or concede the point. I cannot take your points at all seriously while you show such disregard for the responsibilities of communication that will enable us to move this issue forward.

Are you reading what I write or just parroting your original point? https://zero-k.info/Forum/Thread/31189?postID=219329#219329 and http://zero-k.info/Forum/Post/243633#243633

Here is some help:

quote: Imagine a master cobbler who shows up to a race with a box of magic shoes that double the running speed of the wearer. Unfortunately he only knows how to make them in one size - his own. Others can wear them with some greater or lesser degree of discomfort, and they don't fit some people at all. Having a whole box of such shoes doesn't make the race fair, and even many who can wear the shoes, but quite uncomfortably, would probably rather not have to.

While we're at it. When you say

quote: That's not a distinction that's possible to make. Every one of the things I mentioned (and indeed, every UI or unit AI change ever made) makes units stronger or weaker. Scattering makes Likho less good, even distribution line move makes Aspis stronger, Glaives auto-retreating from Reavers makes Glaives stronger, the F-alt command lets your badgers easily lay a minefield (though most people aren't aware this exists), etc.

How about reading

quote: I'll have another go at picking apart the difference between player widgets and game UI. Well, apart from the fairness difference, which I think has been well covered. To start with, the UI is meant to be used by players to implement their decisions, not make decisions for them. The issue is that the act of creating anything involves making a bunch of decisions along the way. In making this comparison I may have conflated "creative/design decisions" and "ingame player/strategic decisions", but I feel like the natural mode of widget creation is to conflate these two types of decisions. Consciously keeping the design decisions within the widget separate from the strategic decisions of those using the widget is simply a lot more difficult than bundling all the technical and strategic considerations together, and is redundant effort when making a widget designed for personal use. The default is to, subconsciously, hardcode widgets with the types of decisions the creator tends to make, and to mesh well with their playstyle. Everything the widget does is in essence a real strategic decision for the creator because they can tweak or rework anything they want. Other players don't have this ability and it can't be expected of them. They'll not use the widget or be stuck having their units implement someone else's decisions. Letting a powerful widget of this nature dictate balance and the rest of the design puts everyone else in a tricky situation. Many people will either be left behind, or use the widget but noticeably degrade their UI in the process.

quote: Also, I don't think this is any different to the current situation. A huge number of the default-installed widgets are finnickety in some manner, have surprising behaviour or interactions and don't always do what you might want: * Auto Jump Over Terrain often lands your jumpcom in hot water. * The Recluse unit AI will have it walking right into that Stinger. * The Phantom unit AI will have it stopping outside of visual range when attack-moving vs a radar dot indefinitely rather than moving in for the kill. * A Cyclops will sit next to that raised Stinger plinking away at it with the slow beam until it dies. * A Minotaur will attempt to kite a circling Ravager, repeatedly swinging its slow turret away from its target and rarely firing. * A Reaver will happily keep walking right into Rogue shots. * Detriment on attack move used to sit at missile range plinking away. * Badgers laying minefields won't target units that come into range or try and kite them. * GBC is a thing.

Then you've forgotten https://zero-k.info/Forum/Thread/31189?postID=226780#226780 and haven't been paying attention to what people have been saying since. I'll break it down.

• Auto Jump Over Terrain and GBC are probably around customisation level 3 or 4 so fall into the use at your own risk category. Pointing out that there are some barely maintained/janky widgets hidden in the repository doesn't help the case for adding more unpolished widgets.
• If you want to help with Badgers then do https://github.com/ZeroK-RTS/Zero-K/issues/4479
• If you want to fix things like Minotaur jinking frequency and Cyclops putting too much stock in its slow beam then actually make a pull request and fix them. I'm not saying the current unit AI is perfect or done, but being 90% of the way there on this front is no justification for pushing widgets with fundamental problems (read what I wrote above) over existing UI that has solid concepts behind it and just needs some refinement.
• Phantom only does what you describe if you set it to not fire at radar. By doing so you've said you'd rather it sit back safely and do nothing than go and melee its target, which would probably lead to it dying.
• What does "keep walking" mean for this Reaver? If the player told the Reaver to walk somewhere then it should walk there. If the player tells the Reaver to try to end up at a location, but fight stuff and try to avoid damage on the way, then it should do some jinking.
• Every other case of 'surprising behaviour' is the simple implementation of Attack Move which has units move forwards until they can fire, then stop and fire.

I'm really not buying this argument for including a whole bunch of unpolished widgets that harm the UI for general players. The core of the argument seems to be that, since there is some obscure, broken, or unpolished stuff in the existing codebase, there is no harm in adding a whole heap more.

I think that eventually, as some people mentioned, it will be very frustrating/time consuming to keep banning more and more widgets, and some people will just come up with another one that does the same thing in a different way, or they'll come up with more and more widgets that are unacceptable, or something like that, if you know what I mean.

So I think the easiest way and probably the fairest way to do it is to just remove widgets from Match Maker games, (I forgot a lot of what everyone else said so if this was already established then sorry I forgot.) Or, if people don't like that, then maybe it would be easier to instead of remove widgets that aren't allowed, what if you just removed all widgets except for a few that were allowed, and people could ask if a certain widget could be allowed. I think that would be easier since I am guessing there are more bad widgets then good ones out there, but maybe not.

quote: A problem (one of them) with removing widgets in certain modes is that it changes the expected behavior of the game. For instance, I have a widget that's basically the same as the default TTS widget (and overrides it), except I wanted it very verbose, so I made it speak numbers, spec chat and other info to me so I wouldn't have to keep glancing at the text chat all the time. Remove that in MM and I might miss some important information because I expect it to be spoken, since now "l8r m8" would be pronounced as "lr m".

Right, that is very true, that's why I thought about it more and said this:

quote: Or, if people don't like that, then maybe it would be easier to instead of remove widgets that aren't allowed, what if you just removed all widgets except for a few that were allowed, and people could ask if a certain widget could be allowed. I think that would be easier since I am guessing there are more bad widgets then good ones out there, but maybe not.

Which might make a lot more since then removing all widgets from MM games, since it sounds like there are some that are very useful and nice to have.

quote: Announcing widget names or descriptions is pointless because they can easily be changed. "Your enemy is using widget 1 and widget 2 with the description Get git and git gud."

quote: …at least you can call the person out on what the widget does...

...or you just get to know the widgets are there and get out if it's on principle. Besides, ZK doesn't have a big enough community to be going around being an acerbic ass to everyone.

If there's no trust to even put correct labels on widgets the end to all this will prolly be rather uninspired regardless. I don't think there's much hope that, left as is, custom widgets will be properly announced, used in test environments only and get properly distributed in whatever state they are in if proven effective. That being said I don't want custom widgets gone, hell, I use one constantly - namely esainane's playerlist. It slaps even in its unfinished state and I'd rather continue relying on it. I do feel like one has a right to know that a widget is there at least, whatever it entails.

I don't see how casual (assuming we are talking pot, small or FFA) is the place for the live "testing" of all this either. There are about more than a dozen widgets circulating among the people who never even touched MM already, and that's just the ones I know of. I smh highly doubt any of them are getting public spread in the near future. At the same time it's a constant point of contempt even in the pot as of late. Iirc both outside of mm usage and archive thing were touched on already btw, at least in previous big thread.

Announcements in-game are just an easy way to get everyone up to date. Whether you don't want to play vs people using widgets, or want to find out about an inevitable new strong one early in its use. I feel like if they are to continue existing in competitive modes the very least there should be is a clear heads-up for players. It also sounds like something which can be implemented relatively easily.

There's also a lobby setting for no local widgets already. Maybe another checkbox for MM would be nice, but I slightly suspect that no one in higher echelon has a burning desire to play vs widgets rn, so it'll be as good as outright banning them.

quote: *is sad to learn he's not in the higher echelon*

You're either in this picture or you're not:

quote: ...or you just get to know the widgets are there and get out if it's on principle.

Let's say I have this and only this enabled:
Name > Print Wind Speed V1.1
Description > Prints the current wind speed to the corner of the window

There's a widget. Would you still play or get out on principle?

Now the thing is... I copied that from the legitimate widget and pasted into my own. My (hypothetical, of course; I wouldn't do that even if it were possible) widget gives my Com a coupon at the start of the game for a free nuke silo that already has 10 missiles stacked.

So next time you don't play against someone who has Print Wind Speed V1.1. to be on the safe side. But that person had the legitimate widget that simply prints wind speeds. Whose loss is that?

quote: If there's no trust to even put correct labels on widgets [...]

Call me naïve, but I like to believe that honest people don't do dishonest things most of the time, so it's bad actors you really want this whole "reveal their widgets to me" thing to protect against. But it can't because they are the ones that will be putting the bad descriptions in. Also lazy ones, or those in a bad mood... but regardless, it seems to defeat its own purpose.

dyth68

quote: No, I'm just 90% sure that 90% of the people I'm talking to are unfamiliar with this nuance, so I'm using "widget" in a manner I think they'll understand (e.g. little bits of unit behaviour). I admit, in hindsight it might have been better to just post as if I was talking to you rather than trying to speak to a wider audience.

This doesn't matter. I was not concerned at all about how you used the word widget.

quote: The main point I was trying to make there was "Getting competitive at ZK already involves figuring out a whole bunch of UI and slightly janky behaviours, this bit extra is a drop in the bucket".

Can you actually demonstrate this? It seems like a lot of what I and others have been writing refute this point. What is your answer to my refutation of your dot points?

Are you seriously saying that top ZK players have to master GBC and this obscure autojump widget? If you're not, then don't use them to pad out your examples.

This flows nicely into what seems to be your main point.

quote: quote: Imagine a master cobbler who shows up to a race with a box of magic shoes that double the running speed of the wearer. Unfortunately he only knows how to make them in one size - his own. Others can wear them with some greater or lesser degree of discomfort, and they don't fit some people at all. Having a whole box of such shoes doesn't make the race fair, and even many who can wear the shoes, but quite uncomfortably, would probably rather not have to. I think this describes this game and every other game to an extent. Every game has bits of UI that you're not good with or don't mesh with your playstyle. Particularly games with a single main developer will have the "shoes" made to fit that vision (which other people may struggle with)... Which makes me think I'd misunderstood and that's the whole point you were trying to get at... That Zero-K is the game made with GoogleFrog shaped shoes (or at least GoogleFrog-vision shaped shoes), so the people who liked it and play it will be those with somewhat GoogleFrog shaped feet, so someone else coming in with, for example, Brackman shaped knee-braces puts the Brackman shaped people at an advantage? Am I getting this right? Is there also the fact that you make an effort to make the shoes fit as many people as possible (even if this doesn't always work, e.g. Superfluid) while most widget-makers won't (and those who do are probably going to have a PR fairly quickly putting it in the base game and making it no longer custom)?

Yes, to some extent games are tailored towards their developers interests. But you need a much stronger version of this claim to get anywhere close to arguing that 1v1 should be a unit AI widget development free for all. Essentially, you have to subscribe to the belief that any random person who can write a bit of code is no worse than me at keeping this game fun, fair and balanced. For a large proportion of players - not just whoever it is that is writing the code. You may personally prefer Brackman, but do you prefer everyone?

Put another way, a big part of game design is putting aside your own playstyles and quirks to make a game for a wide range of people, not just the designer. The natural state is to overfit games to a single person - the designer - because you are your own most immediate feedback. It takes work to push against this. Widget developers routinely do not do this work, and who can blame them, it's work.

TLDR: False equivalence between me and widget developers.

Edit:

quote: Is there also the fact that you make an effort to make the shoes fit as many people as possible (even if this doesn't always work, e.g. Superfluid) while most widget-makers won't (and those who do are probably going to have a PR fairly quickly putting it in the base game and making it no longer custom)?

I didn't parse this sentence on first reading, but I think I answered it above, and the answer is yes. Also I am really doubtful of claim that Superfluid didn't work. 1v1 activity improved significantly since then and overall player numbers are pretty good too. It's hard to tell because of the pandemic peak though. Anyway, I think the onus is on you to show it didn't generally work.