Profile of an Average Player

What is the average game player like? We all have our ideas. Below, you can see a profile of an "Average Player", and some distinctions suggested between this "Average" player, a "Hardcore" player and a "Casual" player.

As I have mentioned before, I have the data from the DGD1 survey and am carefully working through it whenever I can find a spare moment. I'm expecting to get some interesting patterns out of this work some point soon, but in the meantime, I thought I'd share some of the more trivial high level data patterns.

What follows is a profile of the "Average" player in our sample. (Technically, it's the "Mean Player", but that implies something rather different!) As an obvious caveat, our sample consists of anyone interested in taking a play style survey - this is not a completely 'neutral' sample of all game players by any stretch of imagination. Still, it is still snapshot of a certain kind, and we can learn from it.

Preface

One of the conclusions of our original research was that play styles are not split into discrete Hardcore and Casual groups, but that there exists a instance of each play style for both the Hardcore and Casual clusters. This new data appears to broadly support this assertion.

Our system for identifying Hardcore and Casual players is self-identification. This means that those who fall in the 'Casual' cluster are at the very least sufficiently game literate to identify themselves as Casual palyers! I believe this Casual cluster represents Casual players on the border of Hardcore. There is also a third cluster of people who didn't know if they were Hardcore or Casual (we call these the "Unknowns"). I believe these may represent a Casual cluster closer to the "centre" of the Casual world.

The results are presented in terms of the percentage of people in each cluster (or in the entire sample) who answered yes to the sixteen questions in the DGD1 'test'.

• 319 respondents in the sample.
• 161 self-identified as Hardcore.
• 120 self-identified as Casual.
• 38 fell into the Unknown cluster.

All references to significant results refer to statistical significance at a threshold of 0.05 (i.e. 95% confidence). Any reference to the "Average Player' should be understood to mean 'the average result across the entire sample'.

The “Average Player” in our sample

1. “When I first start playing a game, I absolutely want and expect to beat the game.”

32% respond yes.

Hardcore: 34%
Casual: 34% 

2. “If I get stuck, I don't keep banging away at the puzzle. I go away, think about it, and come back with a new perspective.”

50% respond yes. 

Hardcore: 47%
Casual: 56% 

3. “I generally enjoy messing around with the game - it doesn't really matter if I'm not progressing.”

64% respond yes.

Hardcore: 65%
Casual: 64%

4. “The game I'm playing isn't as important as the people I'm playing with.”

34% respond yes. 

Hardcore: 34%
Casual: 38%

5. “When I'm working on a particular challenge, I'll try it over and over again until I beat it.”

55% respond yes. 

Hardcore: 60%
Casual: 57%

Note: Although there was no statistically significant differences between the Hardcore and Casual clusters for this question, the Unknown cluster is a different matter. These players responded in the affirmative to this statement only 32% of the time - considerably lower. I believe this reflects the fact that the further from the Hardcore end of the spectrum one goes, the less patient with repetition the players become.

6. “I want to feel challenged, and I don't mind the game adjusting to my level, as long as it doesn't become too easy.”

72% respond yes.

Hardcore: 80%
Casual: 64%

Note: The Casual cluster's response here is significantly different from the Hardcore cluster (but neither vary significantly from the complete sample). Clearly, becoming "too easy" is a concern for Hardcore players but not for those closer to the Casual end of the spectrum.

7. “When I face a challenge that feels too hard for me, I quickly lose interest.”

42% respond yes.

Hardcore: 33%
Casual: 53% 

Note: The Casual cluster's response here is significantly different from the complete sample and the Hardcore cluster. The Casual cluster has the highest response rate for this question. This isn't wholly a suprise! I view this as a minor flaw in the survey that this question (and several others) skew slightly towards Casual.

8. “Once I start looking after a game character, I feel bad if I don't take good care of them.”

50% respond yes.

Hardcore: 48%
Casual: 49%

9. “I love it when I beat a really tough challenge - that makes everything worthwhile.”

71% respond yes. 

Hardcore: 78%
Casual: 67% 

Note: Another difference between Hardcore and Casual clusters, with significantly lower rate of response from Casual players. This question is an informal test for fiero-seeking behaviour (i.e. the desire to achieve 'triumph over adversity'). Notice the number of people who respond yes! Clearly fiero is a significant driving force in the current audience for games. Also, it seems that fiero is less of a draw for Casual players.

10. “I like games with many different elements, so I can make diverse plans and strategies. I sometimes enjoy a game I lose if I feel I put up a good fight.”

74% respond yes (largest proportion in survey).

Hardcore: 74%
Casual: 73% 

Note: I think this question is rather weak, and the high rate of response coupled with the low incidence of deviation makes me wonder if this needs replacing with a better question.

11. “Sometimes I get swept up in the experience of the game and completely forget about the goals I've been given.”

63% respond yes.

Hardcore: 65%
Casual: 60%

12. “I'd much rather play with other people than play alone.”

36% respond yes. 

Hardcore: 38%
Casual: 43%

Note: The similar response rates for Hardcore and Casual players are unremarkable, but the response from the Unknown sample is highly unexpected.  If you'd asked me to make a prediction, I'd have said this would get higher into the Casual end of the spectrum. But the Unknown sample responds yes only 8% of the time to this question. This is a curious result! The Unknown players would much rather play alone, suggesting that whoever the people are in this sample, they are (collectively) highly Introverted by Myers-Briggs.

13. “Most of the time I won't stop playing until I know I've seen and beaten everything.”

25% respond yes. 

Hardcore: 30%
Casual: 21% 

Note: Although the difference between the Casual and Hardcore cluster responses is not quite statistically significant, if the Casual and Unknown samples are grouped together, the sample sizes become large enough to tip the balance. This combined group respond yes 20% of the time to this question with a sample size of 158.
Although at first sight this doesn’t seem much less than the 30% response rate in the Hardcore group, it is statistically significant. Unsurprisingly, the Hardcore group is therefore the most likely to respond in the affirmative to this question. We can conclude that seeing and doing everything in a game is a more important drive for Hardcore players.

14. “The way I play is more important than winning, because I want to master the games I play.”

40% respond yes.

Hardcore: 47%
Casual: 35% 

Note: Once again, the Casual cluster produces a result significantly different from the Hardcore (while neither is significantly different from the average). Mastery, it seems is a more common theme among Hardcore players than Casual. Not a surprise, but interesting to see the statistics supporting this. 

15. “I usually have more than one game on the go... I don't need to finish one game to start another - a new experience is more rewarding than mastering something familiar.”

64% respond yes.

Hardcore: 68%
Casual: 59%

16. “I prefer a small game world with lots of characters to interact with, rather than a vast world to explore.”

23% respond yes (smallest proportion in the survey). 

Hardcore: 24%
Casual: 21% 

The Hardcore Player (161 respondents)

It says something quite definite about the nature of our sample that there are no statistically significant differences between the "Hardcore" cluster and the complete sample, despite the fact only half of the sample self-identify as "Hardcore". On the one hand, this (weakly) supports our assertion about play styles being independent of Hardcore/Casual factors. On the other, it suggests our "Casual" cluster in this case is not greatly removed from the profile of a Hardcore player. 

The key conclusions based upon comparisons between the Hardcore and Casual clusters are that players in the Hardcore cluster are more likely to be concerned with games becoming "too easy", appear more  likely to be motivated by fiero (triumph over adversity), more interested in mastery and more concerned with doing and seeing everything a game has to offer.

The Casual Player (120 respondents)

The Casual player differs from both the Average Player and the Hardcore cluster in several significant ways, although the deviation in this particular sample is quite small. The only significant difference from the Average Player is a greater tendency to lose interest in the face of a tough challenge, whereas when comparing to the Hardcore cluster, we see less concern about the game becoming too easy, less focus on fiero (triumph over adversity), and less interest in mastery. 

The Unknown Player (38 respondants)

People in this sample did not know if they should be considered Hardcore or Casual… This probably implies a player who fits the conventional template of a Casual player, but is not familiar with the terminology. This is a small sample, consisting of only 38 respondants. For the most part, the Unknown cluster is close to the Casual cluster, but there are some interesting deviations (from both the Casual cluster and the "Average Player").

The peculiarities of this cluster are lower tolerance for repetition and greater desire to play alone. Perhaps this is an artefact of how the data was gatherered: Extroverted players are probably less likely to be interested in taking a Play Style survey. This suggests we may need to consider how we can gather data on more Extroverted players.

And still to come...

I will be grouping the results by people's favourite games in order to produce profiles of the players who enjoy certain games. There may not be enough data to process by individual titles (the largest single cluster for a particular game is for the Civlization games, at 26 respondents), and so I may be forced to cluster by game genre. Either way, it promises to be a new and intriguing perspective on the gaming audience!

Comments

[Duncan]

Question 12 seems a little broad to me. People who are familiar with gaming, and enough of the terminology to specify the type of gamer they were, tended to respond better to this question than those in the unknown group. Could this be due to a lack of understanding and experience in “gaming with others”?

I look at the question, and immediately I wonder if you mean playing via an online connection or in the same room, with people I know. Because these different situations would garner different answers. To those unfamiliar with a broad spectrum of gaming, perhaps they are unaware of games that can be played in local community. Couple this with an unfamiliarity with playing in an online setting (which is often perceived as hostile and “hardcore”), and their immediate response might be "No".

It might be better served to separate community play styles by online and local to get a better picture. Nintendo is trying to get to all those people who don't want to play online, but would play with their buddies in the same room. Have you seen the E3 demo of tennis? They are betting that there is a whole group of “true casual” gamers that exist, and would play if it was easy, fun and local.

[chico]

Just a curiosity: how did the "Unknown Cluster" behave? Like a median (or average) of the Casual and Hardcore gamers or something completely different altogether?

Any notes on their similarities and differences to the other groups?

(By the way, I would love to see a similar research including non-gamers...)

[chico]

oops.. sorry, there it was... "The peculiarities of this cluster are lower tolerance for repetition and greater desire to play alone".

Still curious about non-gamers, though ;)

[Chris]

Chico: I'd love to get the non-gamer data too - but where can I find a pool of (say) 100 non-gamers that I could get to take the survey? Any suggestions?

Duncan: you may have a point, but it's hard to be certain. My personal feeling is that a response rate as low as we got from this group is too polarised to suggest a fundamental misunderstanding of the question. Usually in a survey, misunderstood questions do not produce such one-sided results - because people do not tend to misunderstand a question in the same way. (This makes it hard to spot these problems, as they tend to produce 'invisible skews').

However, I do agree with your assessment that there are at least two different kinds of 'playing with people' to be covered here.
The bottom line for me is that if these people enjoyed playing with other people more than playing alone (in any context) I would expect them to be aware of this; in previous research, this tendency has been readily apparent, and unsuprisingly correlates with Extrovertion by Myers-Briggs.

I stand by my original assessment, but I freely admit that there are multiple possible interpretations. It was ever thus with such vague instruments! :)

I really appreciate you sharing view on this, though; the wording of the question could be improved. However, I am uncertain of how to do this: I expect (based on previous case studies) that a lot of the non-Hardcore positives for this question will be playing with other people in the same room, and a lot of the Hardcore-positives for this question will be for online. I don't want to draw people to one interpretation or another, per se, as I am hoping to detect the people who can confidently identify that they would rather play with other people. The context in which this occurs is less relevant at this stage. Perhaps a task for a different line of research. :)

[zenBen]

I don't remember if I completed the survey, so I don't know if you did this already.
But it would seem logical to ask respondents what game they played last. This seems to me to be something that would skew answers quite a lot.
For instance, if one has just played an FPS, it will probably contain the opportunity to play with style and aesthetic sensibility (a la Q.14), but definitely not allow one to identify with a character much (as in Q.8).
Having recently completed FEAR, this would be my mindset if I answered those questions now. Next month - who knows?

[Chris]

zenBen: We have to remember that survey data is about the lowest rung of the research ladder. We only use them because they're cheap. :) Trying to eliminate skews like the one you mention here would be practically an infinite task. We just have to count on getting enough data that small sources of skewing will average out in the final analysis.

[Brad]

There's another reason people may have selected "Unkown". I can't actually recall what I reported myself as when I took the survey, but I at least remember giving serious consideration to the "unknown" option. This isn't because I'm not familiar with the terms, but rather because they are a bit vague and I'm not sure which side of the fence I'm on. I play less and spend less on games than all the gamers I know, but I'm certainly more interested and familiar with games than the purely casual audience. I don't identify very strongly with either.

-Brad

[Chris]

I certainly wouldn't want to jump to any firm conclusions as to what the Unknown sample represents... They're a relatively small part of the total survey, anyway (about 12%).

I'm starting to wonder if we should stop using self-reporting to differentiate between Hardcore and Casual. Lazzaro uses amount of time committed to play as a differentiator, and our previous survey shows this is a perfectly reasonable indicator - but it is as prone to 'leakage' as our method. On the other hand, perhaps it's time to look at the entire issue from a different stance.

[Dave]

Have you considered doing a cluster analysis on the raw data to see if the respondents actually form into separate groups based on their answeres/

[Chris]

Dave: yes, I have, but I have some experience of cluster analysis from DGD1, and I can tell you now that the results we get will depend wholly upon the parameters we feed to the clustering algorithm. I am exceptionally doubtful that we would learn anything especially useful from applying this technique to this particular data - although I can't eliminate the possibility, and there certainly could be interesting patterns to be found.

The time I can spare for research is very limited, and right now I believe we will learn more by defining clusters based upon clear boundary conditions (e.g. favourite game responses) than by attempting cluster analysis. If I get time, I will probably perform some cluster analysis and see if there's anything interesting we can clearly pick out of the data, but at the moment I consider this to be a secondary concern.

Thanks for taking an interest! Do you have any personal experience with cluster analysis? I would welcome some tips for getting useful results. My standard method is to apply n-cluster models starting from 2 clusters and increasing up to (say) 8, then compare each resulting pattern to see if an underlying structure is implied. It's very time consuming, and the results are not often convincing! I'd certainly welcome guidance that would improve my clustering skills.

Take care!

[zenBen]

Have you thought about splitting the attitudinal approach of players across attributes of their play? Like, instead of just hardcore and casual, use indices like Regularity of Play, Game Literacy, Avg Length of Play Sessions, Investment in Games, etc? I suppose this isn't too original an idea, but it would seem almost as easy to implement as asking 'Hardcore or Casual?'

ps. How is the DGD2 research coming along? I'm quite curious as I'm trying to do something similar myself, using simple test games to measure experiential or typological attributes of players. Hard going!

[VagabondX]

The hardcore/casual question would be better as several questions, which would survey the real playstyle of a gamer. For example, when I filled the test first, I was in trouble. I knew exactly, what each category means, but I didn't fit in any. I used to be a hardcore player, and I have the knowledge, but recently I don't have enough time to play games, therefore I am casual.

There is an other problem with the test, btw. If the browser has a window blocking feature turned on, it might freeze without any advance warning. This just happened to 3 people I know. (Firefox and IE7) I checkhed the uderlying code, and found a serious problem with it, but I'm not expert enough to fix it.

Chris, I'll take some time to examine every question for possible improvements from the gamers point of view. I hope it will help you somewhat.

[Chris]

It's funny you should mention DGD2...

[Janus Rau]

Does anyone know where the casual/hardcore distinction stems from? When was it first used in the industry/litterature?

[Chris]

Great question Janus - although not one that I have an answer for. :) If anyone has any insight on this, please share... If I learn anything, I'll be sure to pass it on.

Best wishes!

[Ei]

I, too, think you should derive clusters from the data.

The lack of meaningful difference between hardcore and casual simply proves the failure of a question - players did not classify themselves correctly.

If possible, I would create clusters on variables which should differentiate hardcore/others (e.g. focus on beating the game), and then use them to compare other characteristics (e.g. attachment to the character).

[Chris]

Ei: this survey was not conducted with the same degree of rigour as our major studies, and was effectively a "side order" of data. We are currently gathering data for a new study (DGD2) and we have retained a statistician to help us derive clusters and find patterns in the results.

Thanks for commenting!

[Ei]

Maybe it would also be helpful to use some questions addressing "hard data", or rather addressing it more directly or gathering more than yes/no answer, because current questions/answers seem a bit hard to interpret (in fact the only one I found helpful for myself was multiplayer question).

Maybe it's also due to the fact that questions refer to player's perception rather than to the facts.

Examples of questions I have in mind: how many games you normally play at given moment (1,2,3, more), how often do you finish the game you are playing (almost never, rarely - ca 25%, often - ca 50%, almost always), how often you replay the game you had finished (..,..)

Not sure about scope of the survey, but could be cool to learn something about how they discover new games.

[Chris]

Ei: A basic problem with surveys is that they depend upon accurate self-reporting. Even when you ask questions such as 'how many games do you play a week', one does still not get factual information, but an answer filtered through the individual's perception.

If you take the new survey you can see the questions we're asking this time around. It's too late to make changes, though, but you might be interested to see the new content all the same.

Thanks for taking an interest!