RolStoppable said:
Salnax, I know you love to do extensive statistical work, but you should consider to ask yourself if your planned approach is going to net any viable results or if there are unaccounted variables that could potentially make the whole endeavor moot. If there's a good chance for the latter, then investing your time into something else wouldn't be a bad idea.
In this specific case, the first thing that came to my mind is marketing. While this renders your observations and conclusions mostly irrelevant, your statistics at least confirm something that you weren't going for. Games that are highly rated usually have higher development costs on average; in turn, games with high development costs have sizeable marketing budgets because the publisher wants the expensive-to-make game to sell well; plus high development costs have a larger probability of leading to a high quality game and that is what creates the correlation of well-reviewed games selling better than the average game. However, correlation does not equate causation; I would put down marketing as the main driving force if I had to pick between it and review scores.
The importance of marketing is highlighted by your findings regarding games that scored in the 71-75 range. You've got a bunch of well-marketed and well-known IPs there, so what you identified as an anomaly (at least on the surface) is actually a confirmation of the importance of marketing. Games that sit one range above (76-80) are often fan favorites that garner good review scores because they do well in catering to the tastes of their niche; however, they aren't advertised on a broad scale, but rather focus on running a few website ads in places where they know their audience resides. That's what created the outlier in an otherwise linear trend. Of course I am only guessing here, because I haven't looked up the actual games that scored in the 76-80 range.
Lastly, the reason why the curve becomes flatter the farther you go to the right in your graph is because the discrepancies between marketing budgets decrease. Games with bad scores were usually developed on low budgets which in turn also means that there wasn't much, if any, investment in marketing.
|
You're likely right. The role of marketing is likely the cause of the results.
As for this being a waste of time, I would disagree (though thank you for being so polite about it). To my knowledge, the topic of the relation between reviews and sales has not had a proper statistical analysis despite being often discussed. At very least, this gives me a resource to look back at and consider the future reviews/sales of games with.
In particular, this has taught me that although reviews do have some relation with sales, they are largely limited to higher scoring games. That confirmation alone made this worth the time for me.
