There is a site and a chart going around that shows some games quantified in terms of return on investment. The original source is the Video Game ROI site, hosted by Ebay of all things.
Of course, this is a list, and we love lists! So I went to see the top ten value rated games, which are:

The Top Ten
To me that was an interesting list, if a bit odd. How did they come up with this?
Well, they are pretty up front with how they did. How they calculated the value rating is there on the front page.

Not bad so far. Hours per dollars spent multiplied by the rating percentage.
So the original Animal Crossing currently costs $6… this is Ebay, I guess they know the used price, so we’ll give them that… and the hours to beat is rated at 69.5 hours, while the average rating for the game is 88%.
So 69.5 divided by $6 gives us 11.58, which multiplied by .88 ends up with a rating of 10.19, which is the best rating of the lot.
Now, you might ask if a game from 2001 qualitatively delivers an experience you would want to spend nearly 70 hours on here in 2014. Fair point, and something not addressed as far as I can tell. And the cost of the game certainly seems to favor used games, but this is Ebay and they want to sell you some used games, so go figure.
I was a little more interested in how they came up with the hours to beat a game.
As it turns out, there is a site called How Long To Beat that is just brimming with this sort of data. I was curious as to how accurate it might be, but didn’t know how I could assess that. I would have to actually beat a game to get that number, right?
Oh, wait, I did just beat a game! I finished Pokemon Y, and all I really did was the main storyline as noted in my post. So I went and looked that up on the site and, naturally, found Pokemon X and Y listed with lots of data. But the essential bit, hours to beat for the main story was there.

So they peg the main story at 33 hours of play time. And I finished the main story in…

Pokemon Y finish time
… 31.5 hours. Pretty close. Close enough that I am probably willing to accept the H2B numbers. Meanwhile, the average rating is as close at MetaCritic, so I am good with that.
So it seems like we have some pretty solid numbers, even if they seem very biased towards older games, which are less expensive. There is Civilization in second place, from 1991. I am not sure, even if you could buy a copy for the $1 they show, that it would run on a modern operating system. The ROI on unplayable games should be pretty low.
Of course, I am interested in MMOs, so I went digging to see what they had listed on that front. Way down at 109th place I found World of Warcraft. Current price, $20, hours to beat, 11.2, and overall review rating of 93%, giving it a value rating of 0.52.

Now, I expected the value rating to be low because I figured that they would account for the subscription model in some way. But no, they figure you’ll be done with that free 30 days yet, since it only takes 11.2 hours to beat.
That seems sort of fast, 11.2 hours. I mean, I am running through the 1-60 on the whole Loremaster achievement thing, so it seems like that number should be higher for somebody new who doesn’t have heirloom gear or what not.
So I started going further down the list and ran into Minecraft at 127th place. The cost is $27 and the rating is 89%, but the hours to beat was 11.2, the same as World of Warcraft.

Now, if 11.2 hours seems very low for WoW, which sort of has a 1 to 60 main game, for Minecraft it seems very much off.
Reading through the site more carefully, I found that if a game is open ended or doesn’t have a well defined main game… which is to say the How Long To Beat site doesn’t show one… they went with the number 11.2 because that was the average of all the games measured.
Color me unimpressed.
Still, I suppose it is an interesting data point for discussing older games. And, of course, it markets older games for Ebay. But you’re not going to convince me that Pokemon Red and Blue, which ran on the GameBoy in 1996, provides a better return on investment than Pokemon X and Y for any qualitative measures.