Tuesday, June 10, 2008

Collecting 101: Review Score Is Important Factor In Video Game Resale Prices

GTA IV Price DropJust like new cars drop in price when you drive them off the lot, the resale prices of video games drop in price the day you buy them. In fact the resale price of video games released in 2007 dropped 0.24% per day. But do well reviewed games drop as much as poorly reviewed ones? We set off to find out.

For our analysis we looked at every game released in 2007 for Xbox 360, Playstation 3, and Wii. A total of 322 games. We then gathered resale price data from VGPC.com and review scores from metacritic.com for each game. We then charted the review score vs the price change per day for every game to see if there is a trend (Every trendline show below is statistically significant with 99% confidence. See discussion below for more details). The chart below shows all the data points:

Click Chart For Larger Image

There is an obvious trend with the lower the review score the bigger the price drop per day. Based upon the data, a game with a review score of 90 points would be predicted to drop in price 0.19% per day, while a game with a 50 review score would drop 0.24% per day. This might not seem like a huge difference but after a year the great game would sell for $16.70 and the bad game would sell for $7.38. The same basic trend holds true for each console too:

Xbox 360 Games: Review Score vs Resale Price

Xbox 360 Review Scores vs Resale Prices
Click Chart For Larger Image

Xbox 360 games had an average review score of 68.7 and dropped in price 0.24% per day. A game with a 90 metacritic score drops 0.21%/Day and a 50 reviewed game drops 0.27%/Day. Also worth noting, in 2007 the 360 had the highest average review score of the three main consoles.

PS3 Games: Review Score vs Resale Price

Playstation 3 Review Scores vs Resale Prices
Click Chart For Larger Image

PS3 games had an average review score of 61.6 and dropped in price 0.24% per day also. A Playstation 3 game with a 90 metacritic score drops 0.18%/Day and a 50 reviewed game drops 0.27%/Day. The PS3 is statistically the same as the 360 in terms of the price drop per day, which makes sense because many games on the 360 are also available on the Playstation 3.

Wii Games: Review Score vs Resale Price

Wii Review Scores vs Resale Prices
Click Chart For Larger Image

The average Wii game had a review score of 45.7 and dropped in price 0.22% per day. A Wii game with a 90 metacritic score drops 0.16%/Day and a 50 reviewed game drops 0.22%/Day. The average Wii review score is below 50 so companies are making quite a few bad games for the Wii. But at the same time the Wii has the lowest percentage drop per day. Maybe all those casual gamers keep the resale prices up.

Why does this matter to the average gamer though? The video games you buy are a depreciating asset. If you are the sort who trades your games in to buy new ones or sells them online after you beat them, be sure you don't procrastinate selling your games.

Now developers and publishers take note too. If you make a game with a good review score it will sell at a higher price for longer. We only analyzed resale prices but it makes sense that the bigger the difference is between used prices and new prices, the more people are going to buy the used game. You will have to keep lowering prices to sell the game. Another good reason to STOP MAKING BAD GAMES.

Nerd Discussion Below. Warning! Here is some more info for math/stats people out there who want to know all the details about the regression. For the complete dataset (the top chart) the r-squared is 0.0539, so only about 5.4% of the price variation is predicted by the review score. The trend is statistically significant though with a p-value of 0.000025, way below the .01 needed for a 99% confidence level. Here are the other r-squared and p-value numbers:

Review Score vs Resale Price Video Game Regression
With this type of data we wouldn't expect one variable to predict a very large portion of the price changes so we think 5.4% is pretty good. In the future we would like to run an analysis with more variables in hopes of improving the r-squared. We are considering using these variables in future analysis:
  • Days since released - price drops usually slow down the longer a game has been available
  • Publisher - certain publishers like Atlus tend to publish games that keep their value
  • Serial game or not - games that come out every year like sports titles drop in price faster
  • Is it the last year of the console's life - games released in the last year of a console's support tend to become rare and don't drop in price
Hopefully with more variables we can increase the r-squared and be able to predict the changes in resale prices with more accuracy.

This is part of our video game collecting 101 series.


Travis Hendricks said...

It's great to know that I usually buy all of my games new [/sarcasm]. If I could only get a few months behind the newly released games I would likely save a good chunk of money every year.

Also of interest, Wii games have an average review score of 45.7 while the PS3 and 360 average in the 60's. I guess that shovelware talk isn't entirely false.

Cory said...

What is the game in the first graph that didn't depreciate?

JJ Hendricks said...

Cory. The game that didn't drop in price at all was Kororinpa: Marble Mania for Wii. It has increased since December 2007 to just about $40.

Anonymous said...

Cool article guys gotta love the regression analysis. The publisher analysis, sequel, movie based, amount of advertisement, current game revenue and game profit would all be great to see. Explanation of the outliers as well would be super cool.

Anonymous said...

Ok so what about good games over a long period of time. Say Metal Gear Solid, Legend of Zelda and Final Fantasy 7. Do they appreciate over a long period of time?

Anonymous said...

That sure explains why resistance fall of man is still 59.99 at best buy.

oh wait...

Unknown said...

What? Something is actually going DOWN in Price? No way, I wont believe it till I see it. It was getting so bad I was waiting for new games to show up used at Game Stop!


Anonymous said...

They obviously only went to locations that had multiple competitors.

Up here in Northern Ny. There is but one Gamestop. They have used games on the shelf that are over a year old and still selling for $45.

Anonymous said...

So, I guess with this information, I can find out how long I will have to wait if I want to buy a USED game for 10 dollars cheaper.

JJ Hendricks said...

Anonymous - the data is based upon internet sales and they are resale prices, not new prices. Just to clarify.

Anonymous - "what about good games over a long period?" Very good question. I will have to look into that. I would guess over a long period the difference between good and bad game prices would get even bigger. Look at how much the games you mentioned sell for compared to Spec Ops games. $20 vs $0.75 a huge difference. My guess is good games drop in price initially but then hit a point where they don't drop much anymore.

Anonymous said...

Could you please post your data?

It seems to me that you have to analyze your data according to a time series analysis to make sure that the price is indeed losing each value every day.

What makes me believe is that you only do a simple regression over the average depreciation value per day which in itself may be misleading.

Unknown said...

I think another important value would be first week sales and overall sales of new games. Also factor in a yes or no categorical variable for multiplayer support. For example, Halo 2 didn't decrease very much over a few years. With the single player vs multiplayer aspect, I see most blockbuster games like Bioshock drop to $40 new (2/3 of their open price) very quickly while games like Gears of War don't drop as fast. Common sense, but it would be an interesting question. My suggested regression would be:

used_price = time + multiplayer + log_opening_sales_volume

JJ Hendricks said...

Anonymous - The regression is based upon the average price drop per day for each game. If Game X's release price was $50 on Jan. 1 and then $20 on Dec. 31. The price drop per day for my analysis would be $0.082 per day. The prices do not drop exactly $0.08 per day, some days it drops more and some days it increases. This analysis assumed linear price drops over time. My guess is prices drop more initially and then slow down which is why I mentioned I want to analyze with the number of days since release as part of the equation to take this into account.

Dan - Multiplayer support is a good variable to test out in the future. I have noticed a relationship like that too so it would be interesting to test. Thanks for the suggestion.

Anonymous said...
This comment has been removed by a blog administrator.

Post a Comment



Login | Create Account