You did exactly what was instructed, and took scores from "barely passing" to mastery.

Emphasize that. If your boss won't bend, offer to take the game down.

Meanwhile, you may want to update your resume, taking scores from 68% to 96% is something that gives you HUGE bragging rights.

I would also suggest seriously thinking about making educational games and/or trainers a direction you can take your career path.

This is not something you should be scolded over or feel bad about. This is an amazing achievement. If your current employer doesn't appreciate you for this, others certainly will.

EDIT PER COMMENTS:

This is a management issue.

What happened with your game is a phenomenon called Gone Horribly Right There was no way to foresee the wild popularity and viral nature of your game, and you should accept no responsibility for it.

Counter any management pushback with suggestions that a lockout be placed after a time limit with a maximum of "X" hours, and any access beyond that require a grant of access from a manager.

Again, this is not something you did wrong, but something that management failed to limit.

Special thanks to Caffieneaddiction and jwenting for their useful comments.

While it's not immediately your fault that people are spending too much time on the game, as the developer, you could help the situation.

Turn off the scoreboard server/disable the game entirely for now. Go to your boss and show the increased test scores to show what benefit it has brought. Work with them to create a middle ground where employees are still engaged but don't spend unnecessary amounts of time goofing off.

I don't think you've correctly identified the problem. You weren't "too good" at your job, you simply performed a task without specific requirements. There were no guard rails. I don't think this is inherently your fault, but there are definitely steps you can take to avoid this.

Focusing on your successes and trying to pass this off as something to brag about doesn't seem appropriate - your boss has a legitimate complaint that is ignored by that approach. Yes, you raised scores, but you also caused efficiency to drop terribly. Employees are spending time playing your training game instead of doing their real jobs. Perhaps this isn't "your fault" but it's certainly a legitimate complaint for your employer to have.

Where does that leave us? That's pretty clear: This is a classic case of poorly specified goals - the terms of your work weren't detailed enough for you to be successful. Did your boss specifically ask you to make a game that people would only play for 4 hours, instead of stretching it to 3 days? Probably not. So, although he is legitimately angry, it's not because of deliberate action on your part.

sf02, Johns305, and others are focusing on steps you can take now to help rectify the problem that already exists, but there are clearly also steps you can take preemptively in the future to keep this from happening again. When your boss has tasks like this, ask guiding questions to help frame out quantitative ways you can ensure your work product is appropriate. In this case, it seems like the obvious question could have been,

"You want the training to be more fun - how can we know that we've met that goal?"

The way to patch things up is to offer solutions for the problems that upper management, your boss, and yourself have identified with the new training. If people are spending too much time, restrict it in a way that they can still complete the training. Show your boss that you are dependable to provide solutions to problems.

Also remember that there is nothing that you need to admit to. An employee choosing to spend 13 days playing a game is the employee's problem not yours.

First off, people skimming through training modules is...(shhh, don't say anything)...very common. That was never your problem.

As for the behavior of Boss, that's may not be entirely your problem. You were assigned a task and the outcome was not what management expected.

Here's you major decision point, did Boss review and approve what you delivered? If so, then it's squarely on them. That people are playing the game they approved too much is their problem.

In this case, with a little emotion as possible, tell Boss:

The training game is functioning as approved. What changes would you like? Here are some suggestions..."

If they object, don't press the issue, just get to the agreed solution as quickly as possible. Arguing with them will get you nowhere.

Or, did you ship the new training game without any management feedback?

In this case, the consequence is at least shared by you and your Boss. They should have reviewed it and you should have made them review it. Therefore:

Wow, I didn't see that happening...I can change it so they can only score during the first 4 hours. I can do that by end of day (or Monday or whenever).

You aren't being punished for "doing your job too well". You're being punished for optimizing your work for one metric without taking into account other relevant metrics. Your ultimate job is to create value for the company. Any particular tasks you are assigned should be considered to be guidance towards, not replacement of, that goal. If you interpreted "make this process funner" as "prioritize making this process funner to the exclusion of all other considerations" rather than "increase the value for the company of focusing on how fun this process is", then you weren't doing your job. When you're assigned a task, you should, to the extent that it is feasible, consider how it fits into the wider picture, and consider not just the metrics that you are explicitly given, but how it will affect the company in general. If you're assigned the task of reducing shoplifting rates in a store, simply shutting the store down will comply with the literal parameters of the task, but such a strategy would be unlikely to be commended.

You either didn't foresee, or disregarded as irrelevant, the wider effects of your work, and this is a valid basis of criticism. How much blame lies with you for creating a program that is decreasing productivity, and how much lies with management for allowing it, is debatable. Your discussions with your boss should be around expectations as to how much responsibility lies with whom to take wider considerations into account. You seem to have operated under the idea that you were to simply optimize for the explicit metric "funner" and it was your boss' responsibility to ensure that this optimization's effect on other metrics was acceptable, while your boss thinks that you should have considered this yourself. If you submitted this to your boss, and your boss implemented it, then you should have a discussion with them about what submitting something means (does it mean "I'm taking responsibility for this", or does it mean "I'm submitting this for your consideration, and it's up to you to decide whether to implement it"?). If you implemented it directly, then it's more clearly your responsibility.

Two things are bugging me about the OP's question.

First, you definitely did not do your job too well. Oh, sure, the training scores are up. But productivity is down so much that upper management got involved. That's not a success story, or a project that was simply done too well. Let's say you could go back in time, make three tweaks to your game that would've made it even more enjoyable. Do you think that would've made the project even better? No - obviously no.

Second, if you're going to ever move up the ranks in development, you have to take ownership. Stop blaming a boss, or blaming coworkers playing the game instead of doing their job. Take ownership of the projects you do and the results those projects achieve. Your project failed - not because it wasn't fun, but because it killed productivity in the office. Instead of admitting to that, and figuring out ways to either fix the situation or prevent something like that from happening again, you're too worried about blaming the boss. Heck, at the end of your question you say:

How can I patch things up without admitting any sort of fault?

... which is the exact wrong attitude.

EDIT: The OP's question is almost a perfect example of Locus of Control - basically, whether you assign responsibility on things you control or things you don't control (aka, whether you take ownership on an issue or whether you blame others/luck.) Because, both parties are unambiguously at fault for aspects of what happened. The manager had oversight on the project, and even if they couldn't have foreseen the problems with a leaderboard/etc, at least they could've fixed it before someone played the game for 13 days straight. And the programmer should have never added features that encouraged multiple playthroughs on something that was only supposed to take 4 hours.

But the manager has "External Locus of Control" and blames the programmer. And as mad as the OP is, they're doing the exact same thing, by blaming the manager and trying to figure out how to move forward while not admitting any fault.

First @RichardU is right. However, here is the thing to do.

Add negative consequences for taking too long.

You drop one spot on the leaderboard for taking longer than a day (or something like this) on completing a test.

You limit the number of times the test can be taken, or maybe just drop them 5 spots for taking too long.

Add physical limits to the game.

"You have spent 2 hours on it today, that is the max allowed per day". Come back tomorrow and train again."

"You are permitted 2 test per day. Come back tomorrow. Thanks for stopping by"

"You are permitted 5 hours a week. ....."

"You have fallen behind on your work, when you have caught up your account will be re-enabled."

Addendum

Reporting:

"User Tom has completed his training yesterday"
"Pete complete his training twice this week."

The exact details should be moderated to avoid snitching, but to still alert supervisors of people abusing the sytem.

1 Email to each supervisor, for all his/her employees.

Toxicity isn't likely to change

Honestly even if you patch things up, it's still a toxic environment (any management that responds this way is toxic), so the problem will continue to come up. It looks like lower management doesn't care about you, or about fairness, or even about your performance, but only about what upper management thinks. In other words, when they're stressed, they mistreat you. This is abuse.

My advice (which I otherwise never recommend as a first step, but which I think is warranted here): consider looking for another job, maybe one that leverages your development and design skills. You did an excellent job at making that training platform assuming the employees aren't somehow gaming the system and getting good scores without learning. A non dis-functional company will know how to use such a tool responsibly to help employees learn rather than allow them to become addicted and escape their boring jobs by playing. Your current management sounds like they lack the ability to control their employees and are taking it out on you.

Leaving a job is a personal decision that you have to make for yourself, but I would strongly recommend considering it. In my opinion it just isn't worth it to work for an abusive boss if you can avoid it.

Keep the peace and head for the door

Keep the peace at your current job while you look for another one. Start by asking your boss, "What kind of limits on play would you like me to implement?" and then implement them. But unless key details were omitted from the question, I'd simultaneously work on my exit strategy. Get out of the toxic environment if you can.

Without knowing the specifics of how you made the training more fun, my suspicion is you did more game than gamification.

Now the 4-hour training is taking on average 3 days for people to complete.

This is a problem.

(Considering how mundane most of their jobs are, I really don't blame them)

This is a subjective value judgement.

But in the end it feels like it was more game than gamification. Learn from it.

You're NOT being punished. You just have unintended consequences.

Once, North American was asked to build a bomber to fly subsonic (so the refueling tankers could keep up), to the edge of Soviet territory, then sprint supersonic at Mach 3 to the destination, to outrun Soviet fighter jets.

On their bomber, they discovered the engines used twice the fuel at Mach 3 but were going 4 times faster, so it was actually more fuel-efficient to fly Mach 3 all the way from Omaha to Moscow, and forget the tankers. This thing was awesome like that.

And then, the Soviets developed a surface-air missile that could go Mach 5. The project was a lost cause and was canceled.

So that's you. You did all the right things, spectacular job, and it didn't work out.

That's how it goes sometimes.

And your boss knows that, and it wasn't your job and you couldn't have reasonably foreseen what happened.

Just like the North American designers. It wasn't their job to know the Soviets were working on a missile. It was their CEO's job to know it and prepare the company for the consequences.

You should interpret his anger as anger at the situation, not at you. If he's mad at you for not anticipating this, he's wrong and he'll get over it.

It would help if you had some gating on the game so they couldn't replay content that is not constructive for them to replay. But that's a feature request.

There are a lot of answers already, so I'm surprised nobody suggested this yet:

If people prefer your training-game to their actual jobs - why don't "gamify" (parts of) their jobs?

It sounds like your co-workers' daily jobs contain a lot of tasks they don't find very interesting. You took one of those tasks and made it so much fun that they go out of their way to do this task. And do it well. It seems extremely short-sighted to see this as a bad thing. It could be a wonderful thing if you were able to apply this to other parts of their jobs.

(Though, sadly, some companies have a "if they're happy, they're not working at full capacity"-leadership style. If you're in one of those, this probably won't get implemented.)

With this new input from your superiors, couldn't you change the rules to stop giving points after too much interaction, limit number of attempts, and shut out users over a time limit? You have a target level of engagement of 4 hours: there is a such a thing as too much success. It's unfortunate that management didn't react in a healthier way, but with some simple modifications you absolutely can have your cake and eat it too.

Another thing you can do other than disabling the leaderboard or restricting time spent per day is to restrict how much game time can count toward working hours. In most companies I have encountered, internal training modules count as working time, and it looks like they do for yours as well. You can allow users an unlimited amount of time to play the training game, increase their leaderboard status, etc., but limit the number of hours of this that can count toward working time. That is, each user would get a certain number of "free" hours to play the game, and would not be required to provide any account to management as to why they took that long, as long as they actually completed it with at least the minimum passing score. Any additional time spent over this amount would have to come out of personal time (either non-work time outside the working day or vacation time), or be specifically arranged with management. If a user can genuinely demonstrate to management why spending X additional days on the training would be helpful to the company (e.g. that they need additional time to get a passing score, or that obtaining a higher-than-minimum score would provide some tangible benefit), they can do so, but if it is clear that the user just wants ego points, management can shoot the request down.

How many hours is reasonable to allow users to complete the training "on the clock"? You mentioned that the training is officially a four-hour training, so perhaps you could double this to account for the fact that some users will naturally take longer to master the material. If a user requires more than eight hours, they could speak to their supervisor to obtain permission to spend additional time "on the clock" working on the training. This could be necessary, for example, as a Reasonable Accommodation for a learning disability.

I agree with many of the other answers calling out management issues but I would avoid worrying about blame.

Your question was

How can I patch things up without admitting any sort of fault?

Your boss had a good idea to make the training more fun and you unintentionally made it a little too fun. I would approach your manager with something like

I wanted to get back to you on your recent feedback about the training game. I think your idea to make the training more fun was a great one and I loved the opportunity to work on this project. Hopefully it's clear that my intent was to complete the assigned task as well as I could and make the training more fun. The dramatic increase in test scores is more than I expected and I think that we can both agree that the unintended side effect of productivity decline was unfortunate. I've been thinking about some remedies that we can put into place now to prevent further problems this year and I would like to plan to set aside some time in about nine months to rereview this project for next year's training so it can continue to provide the intended results and prevent the unintended side effects.

At this point I would refer you to the other answers on how to fix/limit the game that you can discuss with your boss.

You're not actively being punished for something you did.

You're getting chewed out by your boss, because he himself got chewed out by his boss - he's mad and has to take it out on somebody*. That's completely unprofessional, but not at all unheard of. Unfortunately that means at least 2 levels of hierarchy above you are utterly incompetent.

In your situation you find some people who are in management and liked your game (not your boss), ask them to be your reference, and look for a new job with more competent bosses. You can fight 1 level of incompetence above you, but fighting 2 levels of hierarchy does not offer a healthy risk/reward ratio.

*That assessment assumes the boss didn't previously warn you about the issues you now were chewed out for.

This adds to the answers that say you should limit the time the users can spend playing the game. It sounds like the competition for the leaderboard is what's making it too much fun.

1) Take the leaderboard away from the game.

2) Make a second version of the game, that has the leaderboard, but that discourages play during work time. Maybe it's disabled except when the user is on break or at lunch. Maybe it only works when they load it on their personal computers. Maybe it's a phone app.

2a) If it's easier, keep the official version, but have the leaderboard functionality only enabled during off-hours (for the user or the whole company).