But as you pass the first few hundred turns, the probability that your death is going to be entirely the fault of something you decided to do in the last few turns rapidly approaches one. You can acquire resistances and detection mechanisms quite early, plus your pet is so very useful in that period too.
Once you know what you're doing, every decision is an explicit trade, especially in hindsight.
By comparison, if you design in difficulty from the start, the difficulty will be interwoven with the game design. You can always make the game easier later, but you'll still have the areas that require Advanced Thinking to surmount—you'll just have a shortcut available as well, if you wish. For one example, the Mario series is designed such that every stage can be beaten as "small" Mario. Every power-up is, in ways, a shortcut—if you want to experience the game at its cleverest, play without them.
They would sit a wide range of users down in front of a new level or scene, and silently watch them play, with no help, and with analytics running in the background. If they noticed players taking too long on certain puzzles, dying too often from certain monsters, or just not grasping the expected flow of events, there was further work to be done; likewise, if players blew through a level or puzzle too quickly, didn't have any troubles with a "boss", there was also work to be done.
It's about the unusual design process Valve went through to produce the original Half-Life, and I too found it a fascinating read.
Bungie had a laboratory for testing players playing Halo 3. I remember seeing heat maps of where players died the most frequently.
Interestingly, they provide them for TF2 maps as well:
From that moment on, the designers could be confident that the player understood their new ability and only required some repetition to cement that knowledge.
The almighty example: Peggle.
Maybe I misread, but what you wrote is exactly the opposite of the point Jeff makes. Jeff is respected in the indie games community, and has created shareware games since 1994, while you just graduated college this year. I bring that up because he wrote a whole article explaining his point, while you simply wrote a statement saying it is the worst thing. Why?
the best thing you can do is reinforce every single correct user action with positive feedback and a reward.
Rewarding as many user actions as possible is known to be an important tactic and is not something the author disagreed with (and is likely something he believes as well.) Why was it used as a counter point?
Also, why do you believe that reinforcing every "correct" user action is on the opposite end of the spectrum of making games easy?
Rewarding user actions with points and spiffy effects is actually a part of making the game very easy. People want to feel successful for doing anything at all in casual games, and giving rewards and accolades for easy-to-accomplish tasks (as seen in games like Peggle) is part of making the game easy.
And what you call "correct" user actions in casual games such as Peggle would be better called "rewarded" actions. As players click around the screen, they are rewarded for coming closer to a goal, often having no idea why they got those points. Additionally, many successful casual games are designed so that even user actions that are completely ineffective are rewarded with points to the point that this is an inside joke between casual game developers. That is what I mean when I say that making games easy is correlated, not the opposite of, rewarding players often, and also why games are not necessarily about rewarding user behaviors as much as keeping them feeling confident and playing. Finally, this is also what I believe Jeff is talking about.
I realize he wrote a whole article, I did not read it all, just the bold stuff. This sort of idea about difficulty has been sprouting up since casual games hit the main stream a few years ago. Most people seem to think casual games got so popular because they are so easy, I totally disagree. That is what my comment was based on.
As for making the game "too easy" (notice I said "too easy") being the worst thing, how much fun is it to play an 8 year old in basketball? How fun is it to play chess against someone that has never played before? How many games have you been bored with because they are too easy. Making a game too easy is bad, it will just make players stop playing. The key is finding the balance between difficulty and skill (go ahead and graph it based on your playtesting surveys) and riding that curve up and to the right the entire way.
The reward can be independent of difficulty. The reward is juicy feedback. It is empowering. The problem is most people assume the reward must be something relevant to the game mechanics, like a bigger sword, but that is not true. The reason I used Peggle as an example is because user feedback basically has nothing to do with the actual game. The fireworks, soundtrack, bling-y graphics, all of that is what puts the game over the top. You think that the extravaganza that happens at the end of each level has any impact on the difficulty of the game? No, it just makes the player feel awesome. Sure, it is not as tough as Ikaruga, but have you tried the Master Levels? They are not easy. The feedback is not part of the difficulty.
As for the actual bold parts of that article:
"People will forgive a game for being too easy. They will never forgive it for being too hard." (I switched it since he disagrees with the opposite)
Eh, I guess. But there have been times when I have died while playing the game and I did not blame the game, I blamed my skill. It was not the game's fault. It is bad to make a game too difficult, just as bad as it is to make a game too easy.
"People will happily forgive a game for being too easy, because it makes them feel badass. If a game is too hard, they will get angry, ragequit, hold a grudge, and never buy your games again."
My whole argument is that you do not need to make the game easy to make the player feel badass. Even then, if the game is too easy players will think it is pedestrian and not play it.
"When a player is on the default difficult level, has built his or her characters poorly, and is playing straight through the main storyline with mediocre tactics, that player should almost never be killed."
I would think this is a problem with the tutorial or how complicated the game is. It should be straight forward to see progression of most players and know roughly at what point they would be at throughout the game. If they do not know how to play (have mediocre tactics) that is probably not a difficulty problem. I would look at what made the player think those choices were good, and then try to figure out how you can reinforce the right ones with juicy feedback.
TOO LONG DID NOT READ VERSION:
Your game might not be too difficult, it might just need more/better feedback for the user to make them feel awesome. If your game is too easy it will fail, probably the same way if it is too hard.
It's no good having a game that's too easy, because then it is boring. Conversely, if the game is too hard with no way to provide the player feedback necessary for them to improve their capacity, then there is no purpose in playing.
When Jeff is talking about games that are easy or games that are hard that are still satisfying, he's referring to the fact that both types of games provide adequate feedback to the player. The easy games that are good will still quickly defeat the player if he does not interact in an appropriate and effective manner even though the player still knows exactly how to engage in those interactions. A hard game that is good is fair about its difficulty and the skill challenges provide the player a way to attune their play to the required interaction with coherent and consistent feedback.
In its defense, it's saying that you shouldn't kill your player too much on the default difficulty. That's good advice, but it's missing what I believe to be the real kernel of truth: death shouldn't be excessively costly. Having to play through a long-ass level to fight the 99%-indestructible boss just because he got a lucky shot in will definitely turn people off.
I'm a very competitive person. I tend to be a little bit of a completionist. But, in an RPG like the ones he makes, I'm very likely to play through it once on the easiest difficulty level, do no sidequests, and just power through the story, so I can enjoy it. Then I'll go back and make a 'real' character the second time through, make it much harder, and go off on side quests.
So I think there's a balance to be struck here.
Another commenter replies that the difficulty levels should probably be better documented in this case, so you can tell who they're intended for. Fair enough.
The idea though is that you want to make things accessible. If you start out with ANY product targeting the top 1% you're going to either have to charge ALOT or just not expect to make much money. Having the default mode be not very challenging for newcomers lets them get their feet wet while providing a way to challenge themselves down the road, should they be looking for a challenge.
Ninja Gaiden is fine. It's a good game. It's hard, but it's not hard just for hardness' sake, it's not a jackass about it (usually). Ninja Gaiden was the best example I've seen of a game that makes the player rise to the challenge and become better at the game. There are a lot of games I've seen that are simply cheap and frustrating, losing sight of the 'fun & games' aspect of gaming just to raise the difficulty bar. These games shoot themselves in the foot by severely restricting their appeal. For example, the Medal of Honor series never interested me because too often, it uses the 'challenge' of making the player memorize with millimeter-precision a route through the level, with no way to find the path without getting shot in the face a statistically significant number of times.
However, the issue isn't really about how hard or easy the game is. The issue is appealing to different audiences. How do you appeal to the hardcore and casual audiences at the same time? One way is by offering difficulty levels (another is auto-adjusting the difficulty, like Max Payne), although some absurd fraction of people just choose "medium" and be done with it. Another way is having a more continuous range of success available, with different difficulties. This is why achievement systems are such a good idea. They let the user cherry-pick how much challenge they want to deal with, as well as when and which types.
This is not to say that hardcore tough-as-nail games shouldn't exist. They have a right to exist. However, they should be made with the awareness that they're niche products, and that the majority of the game-buying public (now that the majority of the public buys games) is looking for some entertainment and not a lifestyle.
There's also a lot that can be done to make difficult games more palatable, and even appealing. As as said above, Ninja Gaiden does a good job of making the player better at the game. Giving the appearance of progress is important, even if the player is nominally failing, they'll feel better if they're getting closer to succeeding. The most important quality is simply the appearance of fairness, which varies wildly and is totally subjective. Quite simply, if the player blames the game for being hard he'll quit, but if he thinks do better and beat it he'll most likely continue.
It would be hard to do the same thing in an RPG without a longer test period, but it would be really easy to do so in a more action-oriented game, or a platformer. Of course, one would want to be able to override the "autodetected" difficulty level, but it would simultaneously address the problem of casual gamers getting frustrated because the default is too hard, and the problem of tough-as-nails gamers getting frustrated because they picked one that turned out to be too easy.
It's an object lesson in humility! In every death after the first few hundred turns, you instantly realize the entire chain of totally avoidable bad decisions based on bravado.
So the problem for the game designer is not to make the game easy so much as it is to make the player feel confident - even overconfident - so that they are willing to see all of the game's features, and then to gradually slip in the challenging elements to disrupt those confident feelings without completely shattering them.
Reward-focused design happens to be the best way to go about this - "you can't fail but you can always do better," where the game puts a lot of emphasis on achievements of skill without making them a barrier to game progression, or at least makes it so that even if skill limits progression(as in most arcade games, or games like Nethack), opportunities are presented to polish up play at earlier points and reap rewards for it later. Resource bonuses like powerups or extra lives are particularly good for this, but even measurements like combo counts or accuracy can serve this purpose if the feedback is immediate and given regularly.
Designing all the game content around the introduction and tutorialization of mechanics, which Portal did, is another strategy, and it can work in tandem with reward-focused design.
This reminds me of the "Iwata Asks" article recently on HN, and I think they have a better take on the issue. http://news.ycombinator.com/item?id=970566 A couple quotes:
Iwata: So you wanted to know what it was that made players insert another 100 yen coin once the game was over and have another go?
Miyamoto: Right. And basically, I concluded that this was born of the players being mad at themselves.
Miyamoto: Well, since you've purchased it, it's surely better to be able to see the ending.
Its call to fame is the multi-player aspect where the level of difficulty can be as hard if not harder than what's available at the hardest difficulty setting for the single-player campaign (veteran difficulty).
For me personally Call of Duty is one of the few games I actually went through the campaign. It felt satisfying.
But like I said, if there must be a challenge, it should be against other players, and Call of Duty got that right too.
This is a game which is totally built around the concept of being impossibly hard, but still succeeds.
'Make a game easy' is what you do when you can't design good gameplay mechanics that work at multiple skill levels. If your game can only work at one skill level, you have to make that skill level really easy if you want to keep as many people as you can playing it after their first try.
The downside is that there is no reward for improving at the game. You play it a few times, and you do not improve. There's no newer corners of the game to explore, no new approaches to playing it.
Some games are so good that you can re-release them over and over for years with just a few changes and maybe some new content, and people will keep buying and playing it. Tetris. Puyo-puyo. Super Mario. Those are the games you should try to idolize, not games like Gears of War or Halo or random-indie-game-that-wishes-it-were-Braid #455321.
You can brute-force a game that sells a lot by playtesting it with countless focus groups, paying Hollywood-level for game assets, and marketing it to death on Gamespot and whatever, but that's the equivalent of hiring a giant workforce and grinding out your web app in some old Java enterprise WebPortletComponentModelerFactory thing.
Be agile, try small fun stuff and wander around until you get something that's fun to play, and then think up a way to make a full game out of it. Don't start the other way around, with the desire for 'some game' but no way to fill it out with the fun part.
Addendum: Don't be afraid to let the player die. That's why there are multiple lives (if your game works like that, I mean.) Most games I see today go as far as possible to prevent a player from dying during normal gameplay, to the point where you can clear many games on your first try without any sort of negative feedback -- Assassin's Creed 2, for example.
To learn to play Super Mario Bros., you had to die or watch someone else die. If you're setting out from the beginning in your design with "I never want the player to die", then just don't have lives and death at all, like Braid. It's ridiculous when you have a lives system in a game but contort your game mechanics to never use it. You end up creating an environment where a player never really gets negative feedback on when to not do something, and you have to constantly push them in the right direction with obvious and uncouth hints in order to prevent them from making an 'accident'. You're preventing your player from learning. Stop it!
I'm 29 and get off my lawn!