posted on Mar 2, 2024

What defines a "worthwhile" game?

Whenever friends ask for game recommendations, I have a pretty hard recency bias. In effect, the last game I played is likely the game I'll recommend. However, if we take a more holistic approach to assessing the games I've tried, what would my new recommendations look like? Let's look at the research and the data.

What makes a game fun?

We've introduced a few criteria for what makes a game fun. In the context of puzzle games, the answer should be obvious in retrospect (Why Wordle dies in 2028) and in a more technical sense, there should be a large number of possible actions but very few — if not only one — set of actions that leads to victory (Designing the Tic-Tac-Toe from hell).

These are specific definitions for specific categories, but for games more broadly, there are theories all over YouTube and the web, from definitions based on the emotions the player experiences to the degree of skill mastery.

• Nicole Lazzaro proposes "4 Keys 2 Fun" in a blog post: Hard Fun (personal triumph over adversity), Easy Fun (curiosity), Serious Fun (relaxation and excitement), and People Fun (amusement).
• Raph Koster proposes that fun is based on skill mastery, in the book "Theory of Fun for Game Design". Per a game design course at Plymouth State "What makes a game fun", the book above is based on "flow state," first proposed by psychologist Mihaly Csikszentmihalyi. According to the course textbook, flow state requires (1) a challenging activity that requires skill, (2) clear goals and feedback, and (3) uncertain outcomes that are influenced by your actions.

Generally speaking, a number of theories along these lines focus on why surmounting challenges are "fun" when designed in a certain way. For how these translate into specific practices for different game types, see a series of YouTube videos that explain "How to Fail at Making a X", such as this video on "How to Fail at Making a Platformer".

Is there ground truth for fun?

All of the above are properties of the game itself, but what if we look beyond the game? Let's say that we've owned the games for a year now, and we can now look at playtime. This is much more robust metric — did I love the game so much that I kept going back to it, over the course of the year?

To get playtime, I used the Nintendo Switch's built-in tracker, which reports playtimes in Settings. Values are rounded down to the nearest multiple of 5 hours, but that's granular enough for us.

Judging by just the playtime alone, it looks like Smash is the winner, followed closely by Overcooked 2 and Stardew Valley. That matches my intuition decently well: These are all the games I have fond memories of playing, with friends and family.

Let's refine this a little bit however. I noticed that the top entries on this list are a bit outdated, like "Overcooked" and "Smash". I've had those games for several years now, and by comparison, I've only had other games (e.g., Stardew) for a few months. This isn't quite a fair fight: I've had many more opportunities to play some of these older games. To this point, let's normalize by time to give newer games a fair fight. To do this, I can use Nintendo's account Purchase History page to find purchase dates.

Now, this lines up fairly well with my recent taste in games. Quite simply, my recent addiction — Stardew — tops the list. In fact, to a degree, time-normalized playtime is a "ground truth" metric for fun.

However, I'm interested in a different metric, beyond fun. Was the game worthwhile to buy? Since we have a proxy for fun (i.e., playtime), we can now normalize fun by dollars spent. To be specific, let's actually take its reciprocal and compute "dollars per hour of playtime" — the lower the better.

Let's break down what these columns tell us.

• Dollars per Hour: Although "Fun per Dollar" is interesting, it's a little harder to reason about. You'll notice I also added a "Dollars per Hour" column. In short, what would be the value of this game if I'd rented it instead of buying? Per this metric, I'd be paying a quarter per hour for my favorite games. Just judging at face value, this seems like a pretty good price to pay.
• Paid: I paid a total of $385 for the games I currently own. Given I mentioned renting above, let's comparing against game rentals. Over at GameFly, 1-game-at-a-time rental is $18 per month. I've owned the Switch for over 3 years, which we'll round to 40 months; all told, renting would've costed $640, or about 70% more. Since all games would technically cost the same under this model, I'd be paying $2.38 per hour of playtime — almost 10x the cost for my most worthwhile games "Cat Quest II", "Stardew Valley" etc. Granted, I'd also be paying 10x less for games I just want to try, like "Just Dance".

Across all of these lists so far, we do see consistent winners and losers. In fact, "Stardew Valley" is the only game to always stay in the top 3. Similarly, "Just Dance" is always last.

Which games do I recommend?

Given time-normalized playtimes correspond most closely to my internal sense of "fun-ness", here's the full table of results for my paid Nintendo Switch games¹.

Here are a few recommendations based on the above table of results, as well from our observations previously.

• Good couples games: Using time-normalized playtimes, I'd recommend the top three games strongly — "Stardew Valley", "It Takes Two", and "Cat Quest II". These are all great couples games!
• Good party games: The next three on the time-normalized playtimes list are also strong contenders. I usually play "Smash Bros", "Overcooked" and "Unravel Two" with friends and family, as "party" games. Smash in particular can support up to 8 players, which is useful for entertaining many friends at once.
• Don't buy: I'd recommend against Just Dance — it's a consistent last-place game across all of the lists above, which lines up with my lackluster impression of the game. We also see a consistent loser "Overcooked: Special Edition", but this is only because I have many other overcooked versions already.

We can also use time-normalized playtime to compare games I play on PC as well — combined from both Steam and Battle.net.

Taking both lists together, we can see that I do clearly spend much more playing games on the PC than on the Switch².

• Call of Duty: I believe my ranking for games — within the Call of Duty franchise — lines up fairly closely with the community's perception of recent editions. I recommend Cold War and Modern Warfare 2 — my top two, by a large margin. Vanguard and Modern Warfare 3 are both disappointing. Warzone is fun but not my cup of tea. The mobile edition was fun while I didn't have a gaming PC, but the console experience is markedly different.
• Helldivers: It's looking like Helldivers is shaping up to be my recent favorite, which agrees with my own intuition — especially with a large amount of playtime in such a short amount of time. Part of this is due to Helldiver's "unit of playtime" being much longer. A mission typically takes 20-30 minutes, in contrast to short 5-10 minute rounds in Call of Duty.

This about wraps up my list of recommendations for games. I'm sure there are many more worthwhile games, but if you're curious about "value" of a game, this list and data should provide you a bit of insight.

Takeaways

We've now looked at "fun" as an intrinsic property of a game. We've additionally looked at a ground truth metric for "fun" — namely, time-normalized playtime. You can pass this on too: Look at your own playtimes and purchase dates, then compute your own time-normalized playtimes to find out what your top recommendations are.

posted on Mar 2, 2024