Intro

AAA vs Indie: Game analysis of 2020-2025 period

Background

Background

Steam’s stats - THE PC digital game store

• Launched in 2003
• Platforms: Windows, macOS, Linux
• 120K+ games
• Operates worldwide in over 230 countries and terrorities
• ~132 million monthly active users
• Sales revenue of $10.8 billion as of March 2026
• In US, 75% market share

Epic’s stats - who is 2nd place in PC digital game store

• Launched in 2018
• Plaforms: Windows, macOS
• ~6,000 titles
• Operates over 190 countries and over 30 terrorities
• 78 million average monthly active users
• ~$1.16 billion as of Feb 2026
• 8% market share

Global Model

Research Model

Research Problem and Question

Research Problem and Question

2020-2025 is a very atypical period of time with many world events

• Pandemic
• Cryptocurrency
• Logistics
• Tech Layoffs
• AI

How much does this affect gamer’s perceptions of their games?

Specifically AAA games ie Call of Duty, Cyberpunk 2077, GTA V and Indie games ie Among Us, Hi-Fi Rush, Hades, Balatro

Hypothesis

Using linear regression, H0 is AAA and Indie review scores didn’t change over time.

H1 is AAA review score drops from 2020 to 2025 and indie review score increases within the same time.

Reasons

• Local gamer sentiment on both AAA and Indies
• Economics

Dataset

Gamalytics

Why Gamalytics instead of Steam directly?

• Don’t need make publisher/developer account which requires a business tax number
• 81,107 games
• Has review data
• Has an estimated copies sold value
• Has public and free API to scrape to get game data

Variables

                                 column      type
steamId                         steamId   integer
id                                   id   integer
name                               name character
copiesSold                   copiesSold   integer
unreleased                   unreleased   logical
earlyAccess                 earlyAccess   logical
firstReleaseDate       firstReleaseDate   numeric
releaseDate                 releaseDate      Date
earlyAccessExitDate earlyAccessExitDate   numeric
EAReleaseDate             EAReleaseDate   numeric
price                             price   numeric
developers                   developers      list
publishers                   publishers      list
publisherClass           publisherClass character
reviewScore                 reviewScore   integer
genres                           genres      list

Dependent Variable

• Review score
• Release date

Independent Variable

• Publisher Class ie AAA or Indie

Intervening Variable

• Genre

Data Collection

Data Collection

Used Python to pulled the data from Gamalytics public API and format it and save to JSON format.

R for data cleaning, calculations, analysis and visualization.

Data cleaning

• Eliminate unneeded/empty columns
• Convert UTC release date values
• Isolating for AAA and Indie

Analysis

• Learned the number of indies games on Steam: 29,614 to AAA: 535
• In general, in copies sold, it’s not too much different between the two
• No notable differences in genre between the two publisher classes

Research Findings

The median review scores are similar but Indie games scores higher than AAA.

Research Findings #2

Research Findings #2.5

Research Findings #3

Research Findings #4

Hypothesis Test

Call:
lm(formula = reviewScore ~ releaseDate, data = aaa_data)

Residuals:
    Min      1Q  Median      3Q     Max 
-74.457  -8.682   3.340  11.890  22.543 

Coefficients:
              Estimate Std. Error t value Pr(>|t|)    
(Intercept) 132.428817  20.459274   6.473 2.18e-10 ***
releaseDate  -0.002877   0.001051  -2.736  0.00642 ** 
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 15.53 on 533 degrees of freedom
Multiple R-squared:  0.01385,   Adjusted R-squared:  0.012 
F-statistic: 7.487 on 1 and 533 DF,  p-value: 0.006421

Call:
lm(formula = reviewScore ~ releaseDate, data = indie_data)

Residuals:
    Min      1Q  Median      3Q     Max 
-79.504  -8.596   4.863  14.619  25.715 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)    
(Intercept) 3.079e+01  3.667e+00   8.395   <2e-16 ***
releaseDate 2.382e-03  1.881e-04  12.664   <2e-16 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 20.36 on 29612 degrees of freedom
Multiple R-squared:  0.005387,  Adjusted R-squared:  0.005353 
F-statistic: 160.4 on 1 and 29612 DF,  p-value: < 2.2e-16

Conclusion

Conclusion

H0 is no change between 2020-2025 for AAA and Indie games review scores.
Reject null hypothesis as the P-values of both are significant. BUT the amount of data and the scatter plot shows there is likely too many data points so there is only minor correlation between the review score and the release date and there is other factors that contributes to the AAA regression line decline and Indie’s to increase over the 2020-2025 period.

Further Research

Further Research

1. Narrow down the data set via smaller samples of Indie games or eliminating games that has less than 2000 reviews to get more relatable numbers for AAA comparsion.
2. Publisher Class Indie is very ambiguous now as there is different definitions ie solo developer ie Stardew Valley who was self funded, or studio with independent funding ie Clair Obscura who maybe closer to AA. So using budget size in consideration to help distinguish between indies


3. Genre wasn’t broken down enough. ie Action and Adventure is most common genre because it’s most used one. Also remove the “Casual” and “Indie” tags as that isn’t not descriptive and repeative. There is better genre labels and multi-genre considerations ie Action RPG or Strategy RPG, etc.


4. Epic games store comparison. See if the user base score games similarly.


5. Direct Steam Gamer Survey to get their opinions and thoughts on AAA vs Indie and world events has affected their opinions.

Q&A

Q&A

Questions?