Auto-Play

Here's an interesting game design development that I haven't seen discussed here yet.

Auto-play. This is an increasingly common feature where you can turn on a mode where an AI plays for you. You become a spectator (at least, for the duration of that encounter/battle/level/whatever).

It sounds ridiculous, but apparently a lot of people like it. See this blog post from last year, Entering the era of Auto-Mode. And here's a thread on Auto-RPGs at TouchArcade.

I guess this supports my thesis that fun is about a sense of accomplishment, and not about difficulty. In at least some of these games, it sounds like you can play pretty much the entire game in auto-mode; there is no difficulty whatsoever. But you still get to name your character, watch him level up, buy better gear with all the loots he gathers, etc. So you get a sense of accomplishment. Not fundamentally different, I suppose, from farming/gardening games where you just click to gather stuff, and then click to plant more stuff. (Except perhaps with less clicking.)

But man, this is generating some fairly heated debates in the player community. I don't see anybody passionately in favor of it, but a lot of players saying basically, "You know what, it's surprising, but I actually kinda like it." And then I see people who hate it vehemently; they argue that if the game has an auto mode, that's basically admitting that the game design is so shallow there are no interesting choices to be made. Not a game at all, but more like watching a movie, so they say.

But on the flip side, gamers these days are more casual and multi-tasky than ever before, and maybe a game that plays itself is just what they need. You can still feel like you accomplished something, without all that pesky having to pay attention and do stuff.

What do y'all think? How (if at all) should we be incorporating auto-mode into our games?

 

I imagine most of this is from the tradition of JRPG combat that isn't engaging. Nine times out of ten no decision is being made, and you are just hammering confirm until everything is dead. In that sense, combat isn't really important. The more important decisions are all made outside of combat.

The only real problem with it is that in a sizable portion of the gaming populous, combat has been conflated with being the meat of games. Hell, just about any game that doesn't have combat these days usually has to deal with "this isn't a game" arguments.

I feel like there was a game (about ten years ago, maybe flash) where the idea was that you didn't actually do anything in battle, but the point was to figure out how to gear up the party right for the fight. I could be completely making this up though, because I don't really remember it well.

 

I used to "play" Progress Quest a bit "with a friend" when I was in school. That game didn't even have graphics, just text, numbers and progress bars I think. Between the two of us it basically was a competition who could manage to keep it running for longer throughout the days.

I can see the appeal, but I'd never ever buy a game that doesn't have any decisions to make.

I think the concept can have a place in a game if it is thematically fitting within the context. E.g. a management game where you build a fleet of robot servants that are deployed into homes of rich people in some future society where they work for you and generate an income that you can use to upgrade and expand. There could be decisions to be made as much as you want (some of them just being cosmetics like designing your bots, company logo, your home etc.) and the game could be designed in a way that just never anything goes too wrong and if you idle away you basically just earn money (at a reduced rate compared to when you actively micromanage) to then do something with when you want to interact again. The challenge in the design would be to still make the game compelling to play and not feel pointless. If it's straight up "this game has no interaction and plays itself" I doubt many would consider buying it. I'd imagine there could even be quite a lot of outspoken backlash.

 

Auto-play is brilliant, because it allows people who don't have time to play games to play games because of auto-play, simple as that. It allows your game to become more grindy and longer because it is auto-play which is keeping your players playing longer. It hides away the problems of "grinding" and "boring" stuff to the user because you can "auto-play" through it. But most importantly you don't ever allow the AI of auto-play to be smarter than the person as you want auto-play to lose against bosses or any new areas as it is only designed to get rid of the grindy stage, but favour the person to "jump in" at any time to beat the boss. Yes the game fundamentals should be balanced around the auto-play, that you can turn on or off at any point of the battle *this is key*.

Without auto-play I would not have played the mobile games that I do, and same goes for my father or girlfriend because auto-play exists.

P.S. I just implemented it in my game and it is amazing to watch.

 

I was about to say that this is really about mobile games, but I remember trying Eve Online for a few days. I got a surprising number of kills on people who happened to be watching a movie or something while they 'played' and didn't notice me shooting their ship. It was basically the only way I could blow up other ships, since I was a newbie.

But wasn't this idea really explored and discussed in depth back when the WII first came out and everyone decided that nobody wanted games you could die in anymore until Dark Souls reassured us that some players still like challenge? This was one of the reasons that Dark Souls was such a big deal. I believe that people were discussing semi auto play features - where the game would like automatically beat whatever challenge you were having trouble with or something.

 

Sports games have had this sort of feature for at least a couple of decades. It's very entertaining to watch, and although the Madden/Live predictions aren't always accurate to the real-world events, it also provides a way to stress test your AI by pitting it against itself.

Isn't there also a setting in Quake/Unreal that just throws a bunch of bots in an arena? Was also great spectating.

 

I like to approach JRPG auto-battling from a different angle - I think auto-play in JRPGs is actually a symptom of a design problem: shallow combat systems that offer few interesting choices.

I like to hold up Vivi in Final Fantasy IX, actually, as codified in Rule #147: Vivi's Spellbook Principle which reads as follows:

People consider JRPGs boring, because you make nearly no meaningful choices, ever, the game just uses a Skinner Box mechanism to keep you playing (gotta gain another level! I might learn another useless spell!)

JRPGs are the unholy hybrid* of the Visual Novel and the tabletop RPG (for reference, Dragon Warrior and Final Fantasy, the first two JRPGs ever.) The problem is, the combat system is the place where you could theoretically make meaningful decisions...except, that's a rarity that often comes up in gimmick battles.

Complicating this, is that we've had super-easy Mook battles in JRPGs for so long, that a JRPG like Lost Odyssey gets interpreted as being too hard, when in reality it's difficulty curve is merely 'above average' like the NES RPGs. Alternatively, if you were to use interesting mechanics for every mook, you'd soon have nothing left for the boss characters (I recently read a thread about boss fights on this subforum, so that's another topic I won't dip into.)

While I disagree with the prevailing train of thought in the modern games industry that JRPGs are a dead genre in the "they won't sell anymore sense", I will concede that they're dead in the "we can do new things with the format" sense - JRPGs are a well-explored space that have strongly established conventions that serious efforts at breaking will cause serious backlash - some of this is why Final Fantasy XIII failed, actually. That and the god-awful level design.

I think to avoid this, you have to have more clever mechanics in Mook battles, but more to the point, it's a better idea to de-emphasize combat, like later versions of Dungeons & Dragons did with skill check scenarios that allow for XP gain. Battle-heavy JRPGs are so well-explored now, you're more or less not going to make an interesting battle system that offers clever choices, but you can give the player a feeling of agency if you let them do something that feels clever and masterful.

TL;DR - If you're implementing auto-battle, maybe step away from battles, because they're uninteresting, and add something that is.

JRPGs are actually my favorite genre, and I truly love them. That doesn't mean I won't winge about their shortcomings, however.

 

Is auto play just the next step from twitch? Or is it for people that can't actually stream twitch?

"New AUTOPLAYYYYYYYyyyy. Why watch others play when you can watch your own game being played for you!!"

 

...I absolutely love the idea. Just throw a Cruise Control button on your game and say, "Are you sure you want to enable Auto Play? Can't make any guarantees that you won't die..." It won't work for all aspects of gameplay. But the truth is, there are a hell of a lot of programs out there that bot games already. From a programming standpoint... what an interesting little challenge. Endlessly interesting. We already have an idea of how our games should be played... but to actually dictate the play? Granted, the game can still be resumed manually. I am currently making a platforming game and a resource/building game and I can't see how you autoplay a platformer BUT I can see how you could autoplay the little war game and watch it all unfold with a can of soda. You could even fast-forward if it was all strictly autoplay! How weird would that be?

 

Make your settings, fast forward to see what happens, rewind, change some settings, run again.

Make it an experiment or something so that you can only rewind 3 times. When you rewind you press start & stop for the first one. For the others you press start & stop but it will automatically stop at half the rewind period of the preceding one. That way you can't keep rewinding to the start.

 

I guess you could argue that one of my favorite mini-games, Cooper's Little Adventure, is an auto-play game, since you don't have to jump or direct your flying for the coins, especially once you get a few upgrades. It's not long before you're mostly watching it, just as with any idle clicker game.

But just as with those games, the real game isn't the jumping for coins — it's the upgrades. You have real, meaningful decisions to make about which upgrades to buy and in what order. And of course you have the whole massive development curve that produces a sense of accomplishment. But I think that accomplishment comes only if you feel you have some agency in making it happen; thus it's important that you chose those upgrades. If you took Cooper or Cookie Clicker and had an auto-mode for the upgrades too, so you literally had nothing to do but watch, I don't think it would produce a sense of accomplishment at all.

Perhaps this is a key insight... a well-designed game with an auto-mode simply isn't really about whatever you're automating; the real game is the other stuff you have to decide.

@Azmar, as the person in this thread who appears to have most experience with it, does this fit with your view of auto-play, from a game-design standpoint?

 

I don't believe this game would fit with auto-play, because it literally plays the entire game for you and in this case the auto-play would probably be better than the user. It's only suppose to support and abstract away the boring parts of your game to keep ppl playing, not literally be the game. As the key part is to always allow the option to turn it off as it would never be as "smart" as the user when it comes to any real actions in the game. I don't think this concept would fit in any platform game.

@Asvarduil Yes the combat system naturally only offers a few choices for the user and those choices are generally properly made to support the system. But a few choices in a strategy system does not make the game shallow, that is like saying League of Legends or most moba games are shallow because you only have 4 skills, heck even darkest dungeon! Also more choices does not give the game more depth either. The brilliant part with a few choices (4 skills) is that it's easy to understand for the user, and this allows you to focus on a battle system that actually holds depth to keep them interested (like darkest dungeon), I would argue you can make a more balance able system this way by having constraints. It's weird you may lose more people or people may become less interested easier because you have so many skills in your system and its not very "new player friendly".

The system I am modelling even even went to the point of having only 6 simple stats in the game, and they even removed "dodges" in the game. Literally they removed so many complicated parts and made it so simple, and yet the game has more depth than most MMO's I've played in the last 8 years. It left me in shock for a long time when I realized this, and changed my whole view of making games.

 

I really disagree, although it certainly depends on specific games as JRPGs have a huge range in combat systems, there is actually far more tactical depth and decision making in most JRPGs compared to their western counterparts. The reason for auto combat is the raw volume of those combats and their lopsided nature: that you should easily win most if not all of them.

After doing massive amounts of analysis of many games, I think the conventional wisdom on how JRPG mechanics are weak, stale or boring is not remotely correct. There is way, way more innovation, experimentation and risk taking in JRPG combat design than western RPG. If you really want good or unique ideas on turn based combat, JRPGs are the place to look.

The problem is that JRPG combat tends to be more of a grind, not a lack of depth, but a lack of challenge.

 

Cooper's Little Adventure seems to be what you would call an "Idle Game". You can get more coins by jumping, or you can simply accumulate coins passively while idle. These games are also Resource Management games.

 

Exactly so you accumulate passively (this is the auto-play), and creating better auto-play mode with jumping etc and now the user never needs to play or even exist.

 

Well, certainly the battles themselves are simply battles of numbers and not so much of actual skill. You could determine the winner in milliseconds, even run a best 2/3 with randomized combat styles and decision making within. An auto-battle system could free the player up to actually do other things while the battle is taking place, or simply allow them to skip battles entirely so they never have to leave the Overworld map. You just walk into the enemy and you could play the sound of swords clashing (short ting) and then if you win they get tossed backwards and fall to the ground and you can walk right over their dead bodies. You can still monitor your party's HP and MP, heal and use potions and stuff. All the fun stuff without the grind.

Granted, you might be able to battle more effectively than the auto system. This seems a way we can have jrpgs in the 21st century.

 

Just another note on the JRPG auto battle thing. I think people are getting this example wrong.

A lot of games that have a strict transition to a tactical mode from a strategic mode, especially with very frequent combat have historically always offered an 'auto resolve' mode. This can be seen in classic games like Jagged Alliance 2, through modern day 4x like Age of Wonders 3.

In games where combats are frequent and often pit imbalanced combatants with a strict mode switch for combat, you will see auto battle options. This is very different from something like clicker games or progress quest, which are designed to play themselves, instead of offering the option to the player to 'skip a boring, repetitive, meaningless' battle.

I won't get into a massive debate on wither JRPG or Western RPG is more guilty of this, what I will say is that JRPGs have an order magnitude more variation in their systems than their western counterparts.

Western RPGs are generally built around the idea of character customization, self expression and almost always root in one of two very clear combat modes that have existed since the dawn of gaming. There is little change in western turn based combat since early Ultima games in the 80s.

 

Yeah, I'm "playing" while replying to this thread in another tab.

 

Neither Jagged Alliance or Age of Wonders are JPRGs?

 

That's the point. JA2 and AoW offer the same auto combat option, with the same goal as often offered in JRPG: to skip boring or monotonous fights.

This is not about a game playing itself, it's about skipping over boring or overly repetitive content.

 

All I know about JRPG's is turn-based combat involving blue-haired protags and slime monsters. I automatically think of Final Fantasy or maybe a Phantasy Star/Earthbound style combat system... maybe. But 9 times out of 10 I'm thinking of Final Fantasy 1-6.

 

There is more that just accomplishment to these games. Of the Bartle types, the other type of players who can get into these games are explorers. There is a reason reddit calls clickers/idles incremental games. They start off as mechanical mysteries that slowly unravel their secrets. What drives play in this case isn't accomplishing something, but figuring out what new things that 'something' created.

It's not the number of choices that matter, so much as what meaning those choices have. In most to all JRPG's, all you do is choose from a (mental) chart of DPS vs energy cost what will be the most efficient option. That's about it. You get a little bit of variability with elemental types, and once in a blue moon a status effect is useful, but the underlying decision is how to maximize DPS cheaply (sometimes for extended periods of time, i.e. in a dungeon).

 

Thats a really interesting and insightful comment. +1.

Although I think the JRPG comment that follows is wrong or based on too narrow a view of JRPGs.

 

I'm not seeing the broader world of JPRG combat where you make deeper choices than "10 damage for 2MP or 20 damage for 3 MP..." or "save MP but prolong the battle, incurring more HP loss and requiring healing... which takes more MP healing or attacking" and "I should battle near to the free healing location to save money and MP". "I can get more exp from these monsters in 2 fights vs 4 fights in the other area, but I have to walk across the bridge and back to town whenever I heal... which is better?"

That's all I remember ever thinking about.

Even in Pokemon, with its strict focus on type effectiveness charts, you are still just trying to use skill points efficiently and conserve health AND save time... sounds like an incremental/idle game where you have to manually crank the machine.

 

Grandia II - Combat system is really based on unique timing interrupts in a mixed mode real time/turn based system.
Valkiyra Chronicals - Command point system is either unique or very close to unique, heavily based on out of turn combat and control of space. More similar to chess than any other turn based RPG I've ever seen.
Eternal Sonata - Mixed mode real time turn based combat, with novel 'transformation' on shadow/light space.
Tales of Vespira - Mixed mode turn based / fighting game hybrid combat.
Agarest Generations of War - Extremely complex battle system with positional character linking and combinations allowing you to modify the turn queue.

Those are a few off the top of my head. These are way more innovative and experimental than their western counter parts, and these are like AAA titles not small experimental indies.

 

The only JRPG's that have any depth to combat are Chrono Trigger, FFX, and FFX-2. That is intentionally excluding strategy RPG's and Tales of/Star Ocean that are built on fighter conventions. Yes, the only good JRPG combat systems come from the Dream Team collaboration that was Chrono Trigger and the swan songs of Final Fantasy.

 

Show me a single western rpg with deeper tactical choices than Agarest Generations of War. I've never seen it. Including strategy/rpgs like XCOM.

That they happened to arrange their content such that all the fights themselves are trivially easy, grindy, and monotonous doesn't take away from the core depth of the design.

 

I think you may be misinterpreting what I was complaining about in most JRPGs...or, I did a bad job explaining.

I don't think limited numbers of options is a bad thing - personally, I think that, while Pokemon's battle system is simplistic, it provides a superior experience specifically because you have to pick up to four move per monster, full stop. You have a restriction that makes your choices more meaningful at both the macro and micro levels, even if the combat is just as grindy as any JRPG. You actually have to strategize!

Where I complain about most JRPGs is that they shower you with a bunch of meaningless options (that entire second paragraph of yours is the part where we both completely agree.) Another entry from the Grand List of Console RPG Cliches is Rule #76: The Magical Inequality Theorem, which illustrates the design problem perfectly:

As far as MOBAs are concerned, I don't know why you mentioned those, to be honest. While MOBAs are not newbie-friendly despite limited numbers of abilities, a MOBA moment-to-moment showers you in meaningful choices, the complete opposite problem of what JRPGs have: JRPGs give you tons of options and are easy for a newbie to get into, but you wind up realizing very late how few options actually exist at any given point...sometimes, to the point where people include this counterintuitive revelation in a list of cliches about an entire genre of games.

 

I am very much for an games having an auto pilot mode. a few months ago I was playing ARK and I had just tranked my first T-Rex. which take a few hours to tame. then my wife comes in the room and said it was time to go out for dinner. At that moment I desperately wanted an auto play option. Since then I've made several lists and ways to implement an auto play option into every game I could think of. I think the most important part of make this successful would be to have a the player go through a calibration level that ranks how you play in different situations then uses those metrics to determine how the AI plays while you are AFK. one Idea I'm working on involves being able to recruit your friends characters to play in your party with you while they are in auto pilot mode. this would be great for large clans or armies taking shifts on leading everyone through some grinding missions. I think Auto Play opens an whole new world of options to players.

you know maybe I find this so appealing because so many games these days require mundane grinding.

 

Ackowledged
Yeah, you're not just choosing an attack, you're also having to think about where your units are. That is a little bit deeper than what I was talking about. When I say JRPG, I mean turn-based, random battle, overworld map, party of characters, sleeping in the "Inn" and exploring dungeons, leveling up and defeating monsters on the way to the final confrontation with the big bad in all of its variations.

Tactical vs RPG
In my head, there is a distinction drawn between a regular JRPG, which is story-focused, centered around a protagonist and his/her personal growth and development vs. what I call a "Tactics" game (since the first one I saw was Final Fantasy Tactics). Tactics games are not really RPG's in my opinion, because you're not playing as a character traveling with a few friends, you're basically playing an army simulation on a smaller scale where the focus is on you learning how to formulate an overall strategy and then exploit the terrain and the rules of the game to give yourself small tactical advantages, which will allow you to achieve victory... as opposed to "highest stats wins" you can win or lose a battle in a Tactics game based on how well you play. However, it can still degenerate into the same phenomenon described by @RockoDyne above, to some degree.

Meaning of RPG Lost
Again, in Tactical games (Tactics Ogre, FFT, Disgaea, Shining Force, etc) the main difference is the cast of characters is generally large and even though there is a protagonist it's not like "you" are the protagonist, in the sense that when I was playing Final Fantasy 6 I was Terra Branford, the esper girl, riding in Magitek armor who fried 50 of their Magitek Armored soldiers in under 3 minutes but still did not know what it was to feel love *tear*. Sure, later on I got to be Locke Cole (aka Flynn Rider, Tangled) which was cool 'cause Locke is a told badass, but for the most part I was Terra. I think for some percentage of gamers the whole point of the RPG is lost because there is no role-playing going on, there is just stat accumulation and conservation of resources/dps maximization. So that's I think why we're aiming at different targets here.

TLDR
Tactical games are different than RPG's in my opinion and should not be included under the umbrella term JRPG, which is best defined by the early Final Fantasy series and then later by turn-based games with anime characters. Just not tactical games.

 

I remember seeing that on greenlight and I still haven't gotten around to it.

Is most of this depth dependent on space? Getting combos is pretty old hat for SRPG's and interrupting turns/stalling enemies is in just about every RPG to some degree. Even those two aspects combined isn't that novel, Gust games have done it for a while without the strategy layer, but even then the combat lacks any dynamism. I'm still willing to bet there are two master strategies on a shallow spectrum, with the conservative, zero cost strategy on one end and the nuke everything out of existence strategy on the other.

 

The combat system is based on linking characters together. Each character has a unique arrangement of squares from which they can link. By linking characters together they can act on the same turn.

Different moves will leave characters at different locations once completed. A really excellent player would be able to not only connect characters leading into a combination attack, but to predict their links after combos.

This is so difficult to do that very few players will be capable of it or even really take it up as a challenge. But it's possible, so the tactical skill cap is extremely high. That said, the skill floor required to complete the game is very low and the real gameplay borders on the most insane grinding I've probably ever seen in a game.

But even though the actual game play is mind numbing, the range between skill cap and skill floor is much higher than any western RPG I've ever played - including stuff like XCOM - which to me, means that there's more depth, even if the content doesn't demand use of it.

 

So all of the depth is dependent on space. It sounds as much like an exercise in minmaxing as memorizing resistance tables. Just because there is a high skill ceiling doesn't mean its mechanically deep. Shallow systems can produce an insane number of outcomes, and learning even a fraction of the combinations constitutes skill. The few mechanics in checkers create a simple system, whereas the more elaborate mechanics of chess do far more to play off of each other, creating much more complexity.

 

I kind of wished the thread would have stayed on topic to actually discuss auto-play and not focus on depth of JRPGs or tactical games. Yes we all know how monumental these games are, and how much depth they have. References of course are useful, but the thread quickly got off topic. This topic I deeply want more information on, so sorry for trying to push it in a different place.

What do you guys think of implementing auto-play in more steam based games?

 

Fair enough, the original point is that a number of PC games have always offered auto play options, although we wouldn't think of them as auto play games. They simply offer auto play as a method of skipping repetitive or boring content when the combat system is a different mode from strategic.

This kind of thing has been used in multi-mode strategy games for decades, from Total War, to Jagged Alliance 2, to Age of Wonders, etc. I would argue that JRPGs follow this form and offer a way to skip boring fights rather than a game that's really designed around hands off game play, like a Progress Quest.

 

Ok, and another question for you and other people to add on top of it. What is your expectations of an auto-play mode? As a newcomer expectations from never playing the game before, and expectations after playing the game for a long time? Please try to approach it in a general sense so logic can be applied to many games.

 

I think auto-play can work if the main thing the player is doing is macro-level tactics/strategies. Heck, Programming Games are a great way to do that; Final Fantasy XII is a great example of this in a way with the Gambit system.

 

So what exactly do you mean by auto-play mode? Half of the thread has used it to mean a battle auto-resolve mode/button/toggle, while the other half has used it to mean an AFK mode where something is happening in the game even if no one is actually playing it (even to the point where you make earnings when the game is closed).

The presence of the first is either a symptom of a boring and ultimately unengaging combat system, or a symptom from the inclusion of a combat system that interrupts the true focus of gameplay, or maybe even both. Either combat is pointless filler, or the elaborate combat is not of major importance. One hasn't been developed for long enough, while the other for too long. Both are design flaws.

The second case is just stupid. Time should never be an important variable to progression. Even games that are entirely about management should require the player's attention and never be about running the clock down.

 

I can think of several successful games where Time is the only variable for progression. Clash of Clans for one.

 

Let's expand this beyond simplly jRPGs. Auto play systems have been around for ever and are a fundamental part of many complex games.

The basic function of an auto play system is to allow players to ignore low level repetitive decisions when they don't matter. But at the same time allowing direct access to those low level decisions when it counts.

Take Civ as an example. Most of the game you set workers on automate. But occasionally you need access to a specific improvement now, or a road built to move troops to a specific location. The human player takes over the low level decisions when it has strategic importance. Once the job is done control goes back to the auto system.

Stop and imagine a game of Civ where the player had to make every single population placement manually. the game would be nowhere near as succesful.

 

Auto-mode can really cover a huge range of games, some fun, some less so.

Many games that offer a strict divide between a 'strategy layer' and a 'tactical layer' offer an auto resolve for the tactical layer. This has been done in games since Ancient Art of War in the 1984 and probably before that. This continues to be used in major modern AAA franchises like Total War.

Other kinds of games that offer some degree of 'auto play' are agent based games. These range from Little Computer People in 1985, to The Sims, to Fallout Shelter. How much influence you have on the little dudes in the game can range from directly taking control of the guys and making sure they pee, to indirectly influencing their actions like in Populous.

@Asvarduil mentions Programming Games, which can be pretty awesome fun and are also 'auto play'. I would add games like Dominions to the mix there (<3 Dominions), since you loosely organize your armies and tactics, but combat execution is handled automatically.

Most of these games (aside from the agent based) are 'modal' games, that is: there is a 'battle mode' and a 'strategy mode', the strategy mode controls what happens at a high level, while the battle mode zooms in. But not all are.

Other kinds of "Auto Play" games would include the wildly popular Fantasy Football, as in like, the game that grown men who don't play games play obsessively and devote huge amounts of time to.

At the most extreme end you have something like Progress Quest which is designed to be entirely hands off. Although Progress Quest was sort of a parody, a surprising number of people found it to be quite fun.

At a really high level, I think that games that simulate the 'action' and leave you to deal with the management side can be absolutely fun and complete games.

 

The Combat Mission series of games from battlefront.com have a lot of the setup/planning/strategy stuff with the tactical portion being automated in 60 second intervals. (I used to play them a lot, but they go overboard with the historical realism and I don't have time to learn the details of every single unit type and piece of equipment fielded during the entirety of WWII. Cool 3D user interface, too, loved the painted-on-the-terrain arrows.)

Back when Quake was at its peak I often wished I could throw it into auto-play mode as a screen saver (which was also back when people still used screen savers).

 

Personally, I both design and play games that require skill. These statements are from my view as a video game player:

I play games with skill. If skill is not required to advance thru the game, I don't play it. I don't mind investing time into a game if it is interesting and showcases my skills. However, if I spend 10 hours a week just watching a player play FOR me, I feel kinda stupid.

From a developer's point of view:

Auto-play opens up a whole new world of possibilities. You can provide different objectives for the player. For example, let's say you're making a tank game where you BOTH DRIVE and SHOOT your tank at other players. Well, that leaves us with some ideas from the start. Do we want to make the game all about customizing your tank and making it look cool? Adding the baddest weapons? Making it the fastest or best armored? All these questions are answered differently if we use an Auto-Play point of view, partial Auto-Play or fully manual.

Let's start from full Auto-Play, where the tank drives and shoots for you. This means that the player is probably interested in making the tank look cool since he doesn't control how the tank moves and shoots. So the developer should probably spend most time developing cool tank models and weapons and maybe types of armor.

Now let's talk about partial Auto-Play, where the tank drives itself but the players shoots manually. Now armor is less of an issue since you can't control where you go. So upgrading weapons is probably what the players wants most, since that is mainly how he is interacting with the game and amusing himself. However if the tank needs to be driven but shoots automatically, then the game is probably mostly about tactics. How can you out-drive the opponent, not how you can out-aim the opponent.

Finally, if the game is fully manual, then the player is interested in enhancing the driving capabilities, the shooting and is probably less likely to waste in-game currency on making it look cool.

Personally, I'd make the game fully manual but add options for certain auto-play elements so the game plays how the player wants it to play.

Of course, I spent most of this answer on a specific scenario, but I think it applies to many games.

 

This reminds me of the Halo Warthog dilemma. The vehicle takes two characters to pilot successfully. One operates the machine gun, the other drives the vehicle. In single player mode one of these tasks must be delegated to an AI character.

Shooting things with a big gun is generally more fun then driving around. However the AI does a decent job of shooting but a poor job of driving. So most of the time the player doesn't really have the choice.

So if you give the player the choice of which features to auto play, make sure the game is still fun and playable no matter which way the players choose.

 

Another example is Combots ... I can't even find a mention of the computer version any more but the boardgame version is here (I still own two copies). The boardgame was interactive, of course, but in the computer version you basically outfitted your machines with parts, then they fought it out automatically in the arena. The late-80s AI and graphics were terrible so it really wasn't much fun, which is probably why it seems to have disappeared completely, but it was an auto-play concept.

I did find this Flash game also called Combots, but it's a programming game -- also auto-play in a sense.

The boardgame had cool modular miniatures

 

relevant:

 

I don't think autoplay implies that you can't lose. Enemy AI is autoplay for other parts of the game, right? They lose...

 

I never use autoplay when it's offered to me, don't think it improves the game in any way, and think that the existence of an autoplay feature that actually is able to consistently win indicates a huge problem with your game design. In particular, if your game needs autoplay because it's just so grindy, that might be because, and I'm just guessing here, YOUR GAME HAS A BUNCH OF GRINDING NOBODY ACTUALLY WANTS TO DO AND THAT SHOULDN'T BE THERE. (And don't even mention monetization, because ads are usually most effective when somebody actually looks at them.)

Options that help the player go past hard parts when they're stuck, like Jim Sterling discusses in his video, are a whole other thing, of course. But things like Giant Mario in New Super Mario Brothers or Snake's chicken hat are far more sophisticated than a simple "solve this battle for me" button because they still retain interactivity.