RPG Sales Ranking Based on Steam Gauge, SteamDB and SteamSpy Data

[stony3k]

On the other hand, Darkest Dungeon has sold pretty well, and it's still in Early Access. That's big numbers for a roguelike.

 

[agris]

My bad.

Don't apologize, it's a nice presentation of the current data. That format is easier to read than LESS T_T's imo.

 

[Tigranes]

On the other hand, Darkest Dungeon has sold pretty well, and it's still in Early Access. That's big numbers for a roguelike.

DD made big waves, yes. Deservedly so.

 

[Crooked Bee]

Some moar, not all of them RPGs:

Dishonored: 1,884,501 ± 52,984
Deus Ex: HR - Director's Cut: 984,394 ± 38,371
The Binding of Isaac: Rebirth: 792,690 ± 34,447
Game of Thrones RPG: 421,827 ± 25,149 (data unreliable)
Deus Ex: The Fall: 312,058 ± 21,636 (data unreliable)
Mars: War Logs: 149,364 ± 14,974
Might and Magic X: 143,484 ± 14,676 (doesn't count UPlay tho')
Heroine's Quest: 141,916 ± 14,596 (it's free duh)
ToME: 101,144 ± 12,323
NEO Scavenger: 98,792 ± 12,179
Avernum: Escape from the Pit: 81,543 ± 11,065 (data unreliable)
Styx: Master of Shadows: 78,407 ± 10,850
Invisible Inc. (early access): 48,220 ± 8,510
CONSORTIUM: 45,084 ± 8,228
Loren the Amazon Princess: 43,516 ± 8,084
Winter Voices: 26,658 ± 6,327
Lords of Xulima: 18,818 ± 5,316
Quest for Infamy: 15,289 ± 4,792
SanctuaryRPG: 14,897 ± 4,730
Unrest: 11,369 ± 4,132 (data unreliable)
FRONTIERS (early access): 5,880 ± 2,972
Avernum 2: Crystal Souls: 5,437 ± 3,366
StarCrawlers (early access): 4,312 ± 2,545
Raven's Cry: 2,744 ± 2,030 (lol)
Clandestine (early access): 1,398 ± 1,936

Dark Souls games:

Dark Souls: Prepare to Die Edition: 1,841,769 ± 52,385
Dark Souls II: 981,650 ± 38,317
Dark Souls II: Scholar of the First Sin (just released): 88,207 ± 11,508

Also the best RPGs aka JRPGs:

Final Fantasy VII: 826,405 ± 35,169
Final Fantasy VIII: 386,152 ± 24,064
Final Fantasy XIII: 328,523 ± 22,198
Final Fantasy XIII-2: 157,597 ± 15,381
Final Fantasy III: 106,241 ± 12,630
Final Fantasy IV: 79,191 ± 10,904

Wizardry 7: 80,759 ± 11,012
Wizardry 8: 79,975 ± 10,958
Elminage Gothic: 4,704 ± 2,658 (surprisingly low tbh )

Recettear: 379,880 ± 23,868
Child of Light: 288,928 ± 20,820
Valkyria Chronicles: 271,287 ± 20,175
Ys Origin: 194,840 ± 17,100
Chantelise: 163,086 ± 15,646
Agarest: Generations of War: 155,637 ± 15,285
Ys: Oath in Felghana: 130,939 ± 14,020
Fortune Summoners: 116,826 ± 13,244 (data unreliable)
Trails in the Sky: 103,497 ± 12,465
Hyperdimension Neptunia: 54,493 ± 9,046
Agarest Zero: 48,220 ± 8,510 (data unreliable)
War of the Human Tanks: 40,379 ± 7,787
Agarest: Generations of War 2: 19,210 ± 5,371 (data unreliable)
LISA: 16,073 ± 4,913
Gurumin: A Monstrous Adventure (just released): 13,329 ± 4,474
Pier Solar: 5,488 ± 2,871
War of the Human Tanks - ALTER: 1,960 ± 1,715

Last but not least,

Moebius: Empire Rising: 5,096 ± 2,766
ASA: A Space Adventure: 1,392 ± 2,726

 

[Infinitron]

What would be cool is if there was a table with the number of owned copies for each game PLUS the number of discounts it's had including bundles (you could even have a column for each discount percentage) and how long it's been out.

You could use that as a data source for a heuristic that tries to estimate revenue.

 

[Crooked Bee]

I am disappointed in FF XIII-2 doing worse than FF XIII despite being a better game and just a great RPG overall. I hope they still bring Lightning Returns to PC too, like they originally promised.

 

[7hm]

I'm disappointed that Lords of Xulima didn't do better. Have the Grimrock numbers? (I'm phone posting)

 

[agris]

Crooked Bee, Abu Antar etc, I'm going to put on my analytical chemist hat and mention something: when you report numbers like

NEO Scavenger: 98,792 ± 12,179

Your uncertainty, 12179, should only have one significant figure, in the highest order (10 thousands place). So while you calculate 12,179 (or, rather, steamgauge does), since you're unsure in the 10 thousands place, there's no way your uncertainty can extend down to the single digits (9). That's true for all digits, up until the largest. That means your uncertainty is 10,000 because you're rounding to a single sig fig. Also, because of that, the number of owners can only be reported to the nearest 10k, or 100,000 (98.7k rounded). So NEO Scavenger and Dishonored should like this:

• NEO Scavenger: 100,000 ± 10,000
• Dishonored: 1,900,000 (rounded to nearest digit in the 10s of thousands) ± 50,000 (uncertainty is the 10s of thousands)

The difference probably comes across as autistic, but it makes larger data dumps like these easier to read, in addition to being correct.

ref: link (bottom of page)

 

[stony3k]

On the other hand, Darkest Dungeon has sold pretty well, and it's still in Early Access. That's big numbers for a roguelike.

DD made big waves, yes. Deservedly so.

Makes me proud as a KS backer

 

[stony3k]

Elminage Gothic: 4,704 ± 2,658 (surprisingly low tbh )

It's bad reviews + not been in a bundle or even majorly discounted on Steam

 

[tuluse]

agris

How often do you do random sampling in chemistry? 98k +/- 12 is not the same as 100k +/- 10k. You've just raised the lower bound 4k. Sampling error to sigfigs is not that simple. There is a way to determine it (and obviously they're not accurate to single digits), but I'm pretty sure it's not what you're saying.

 

[agris]

agris

How often do you do random sampling in chemistry? 98k +/- 12 is not the same as 100k +/- 10k. You've just raised the lower bound 4k. Sampling error to sigfigs is not that simple. There is a way to determine it (and obviously they're not accurate to single digits), but I'm pretty sure it's not what you're saying.

It's raised by 4k compared to an incorrect baseline. You uncertainty isn't 12k, since all numbers in the thousands place and under are uncertain due to the leading digit being in the 10k place. That's how error analysis works. Standard deviation can be calculated to a million a digits, but everything smaller than the leading number is irrelevant because that leading number represents an uncertainty at least an order of manitude greater than all the proceeding digits.

If you're interested, check out link1 link2 (check out the sig figs section)

edit: there are definitely more rigorous ways to treat sampling error, you're right, but what I'm saying is an easy way to clean up this data so it's easy to read and is statistically correct.

 

[Abu Antar]

agris Your'e probably right. These numbers also not accounting for all sales as there are many private Steam profile which SteamSpy does not have access to. I also think that any figures under 30k owners are likely to be very inaccurate if I understand the guy correctly.

If anyone wants to know how this works: http://steamspy.com/about

The bombing of Blackguards 2 will always be funny.

Bloodlines and AP selling roughly the same is also pretty appropriate.

It's not funny to me. I think they deserve crap sales. I was stupid enough to buy the game and it isn't nearly as good as the first one.


Because I have a personal interest in Pillars of Eternity, I checked, and 9 games released in 2015 have more owners than it.
It has 6 "rpgs" released in ahead of it when it comes to owners. 4 of those are F2P games. I say "rpgs" because the site is going by the Steam definition. The games that sold better are Dying Light and Darkest Dungeon.

For 2014, Talisman: Digital Edition has over 23,5M owners. Then we can begin(F2P games excluded): South Park was the best selling 2014 rpg on Steam, followed by Dark Souls II and third is Divinity: Original Sin. That means no major AAA rpgs were actually released on Steam.

 

[Abu Antar]

I hope it is kosher to post this.
Asked Darren Monahan about sales. It seems that we will have to wait for Feargod to come back from vacation.

I will also ask about how the Kickstarter achievement work. If it counts for the copies backers could buy for friends.

Spoiler

Hi Robert,

You'll probably not believe me, but we don't actually even know yet! Since Paradox is our publisher for the game, they have the access required to see the sales thus far on Steam, which we can't see, and I believe once our CEO is back from vacation next week, he's going to check in with them.

I'm curious if that 280k includes backer keys or not since it says the average price is now $44.99 (but backer keys have a price internally of 0.00).

Very interesting... should hopefully know soon!

Cheers,
Darren

 

[Infinitron]

lol

Yes, it includes backer keys.

 

[Abu Antar]

lol

Yes, it includes backer keys.

I know, I told him in my second PM. A little less than 17% have the achievement. The question is if those copies that backers could buy also count toward the achievement. It's still money in obsidian's pocket, but a fair chunk less than the Steam price.

 

[MRY]

I can't really verify the accuracy with respect to Primordia. The problem is that 40k "potential" Steam keys were sold in bundles, and 40k key were sold directly through Steam. (Using rough figures.) With the bundles, I know how many were sold, but not how many were activated. So the number could be right, it could be low (if >25k bundle keys were activated) or it could be high (if <25k bundle keys were activated).

 

[Abu Antar]

I can't really verify the accuracy with respect to Primordia. The problem is that 40k "potential" Steam keys were sold in bundles, and 40k key were sold directly through Steam. (Using rough figures.) With the bundles, I know how many were sold, but not how many were activated. So the number could be right, it could be low (if >25k bundle keys were activated) or it could be high (if <25k bundle keys were activated).

I guess bundles is the biggest hurdle here, then. Free weekends will also screw things up for proper data. I still think this thing can be fairly useful, even if the data won't be 100% accurate.

 

[MRY]

I think, in conjunction with Steam Charts, it can tell you, with some confidence, if a game is doing poorly. It is harder to tell whether a game is doing well. As Dave pointed out in his post at WEG, a game that sells 7k copies at $15 a piece makes more than a game that sells 100k copies at buck a key in a bundle. But the data would make you think that the former was a bomb compared to the latter. But even if you take out bundles, seasonal sales distort things; you can't really rely on the list price. Which is why really all you can do is look at a game and see if it has abysmal sales despite having already been bundled; if so, you can infer it's a bomb. Beyond that, dunno what more there is to tell.

 

[stony3k]

I think if you include median and average hours played along with the sales estimates, that should give you a pretty decent idea of the good games. Of course, this needs to be compared to the total length of a game, so 30 hour games that are only played for half an hour on average means a shitty game.

 

[LESS T_T]

Steam Spy now display "Owners of this game also own." It filters out very common best sellers like Half-Life, Garry's Mod and Skyrim.

Funny things:

Among D:OS owners: http://steamspy.com/app/230230
- 39% also owns F:NV,
- 25% also owns SR:R
- 19% also owns WL2
- 17% also owns PoE

Among PoE onwers: http://steamspy.com/app/291650
- 46% also owns D:OS
- 45% also owns F:NV
- 35% also owns WL2
- 39% also owns SR:R
- 23% also owns SR: DF
- 22% also owns BG:EE
- 21% also owns Dark Souls 2

Among WL2 owners: http://steamspy.com/app/240760
- 55% also owns F:NV
- 43% also owns Metro 2033
- 42% also owns SR:R
- 37% also owns D:OS
- 34% also owns The Bard's Tale
- 27% also owns SR: DF
- 24% also owns PoE
- 20% also owns FO2

(Note that SR:R was featured in a Humble Bundle.

And PoE has just released two weeks ago, so current PoE owner demographic might contains proportionately more RPG enthusiasts (who willing to buy RPGs without deep discounts) than other owner demographics.)

edit: nevermind

 

[Blackguard]

Pretty surprising how low the percentages are.

 

[Ovplain]

Beyond me why some here are so thrilled that certain games haven't sold that well.:S What's that about? We don't deserve the incline.=(

 

[Abu Antar]

Beyond me why some here are so thrilled that certain games haven't sold that well.:S What's that about? We don't deserve the incline.=(

The only game that tanked that makes me happy is Blackguards 2. I will finish it someday, but it doesn't even come close to the first one.

 

[Abu Antar]

Paradox announced that Cities Skylines has sold more than 1 million copies.

SteamSpy:
973,213 ± 31,473

Not bad.