A few years ago a smart friend of mine suggested using NES games as models to limit scope for contemporary commercial indie games.
The idea was, while studios in the '80s had larger budgets than indie teams do today, our tools are much more advanced so we may be able to do as much with less.
In this post I break down what games I looked at, why, and how. Then I present the main data table. Deeper analysis will follow in the next post!
tl;dr π
Here are the most important early take-aways:
- π₯ Team Size is most often 4 or 5.
- π Dev Time is almost always one year.
- β³ Most games are beatable in about 3 hours.
- 1οΈβ£ Most games are single player.
- π§° Less than 10 mechanics is typical.
- πΌ Soundtracks tend to have about 11 songs.
- π Most games have about 23 sound effects.
Motivation π
I really liked this idea, as overscoping is an everpresent problem. Once I started thinking about it, though, I realized that there's a lot of nuance here.
For most of us, our ideas about what games were like in the '80s come from gaming culture or from decades-old memories of playing. Our ideas don't come from actual '80s development experience.
This means few of us grasp how big β or small β these older games really are, from a development standpoint.

Shovel Knight looks and plays like a NES game, but it's not one. Not just in terms of technical limitations, but in terms of scope.
Shovel Knight clearly takes inspiration from the NES library, but it does many things that most '80s games would never have attempted. I think modern, non-homebrew retro games in general, from Towerfall to Delta Rune, are just as far away from the realities of '80s development.
So, I set about gathering data to build a sense of these games' ingredients.
NES vs The Rest πΉοΈ
First off, let's limit the scope of this discussion.
We'll be talking NES/FamiCom/Disk System/FamiClone games here, not games from home computers, arcades, or other consoles.
Second, let's limit the scope of this discussion much, much more....
Game vs Game π΄ππ
One area where collectors are perhaps a bit ahead of both players and devs is in recognizing that not all NES games were created equal.
This summary by Big Ole Words is a great introduction to the concept of NES Eras.
He breaks down the NES library into three parts:
- Black Box Era, pre-1988
- Mainstream Era, 1988-1990
- Castaway Era, 1991-1994
I have a couple issues with the way he handles this.
First, he totally ignores the fact that many games were developed in Japan years before being released in any overseas territory.
Second, I think the Black Box Era could be further subdivided, as the single screen arcade titles with which Nintendo began are far smaller and simpler than The Legend of Zelda or DΓ©jΓ Vu with which his first era ends.
However, his main point was a huge insight for me. When I say "NES game", am I talking about Donkey Kong, Final Fantasy, or Kirby's Adventure? These games are so different from one another that, as he says, they might almost have been released on different consoles.

Selecting Our Era π
I think we can exclude the very earliest, single screen arcade ports from consideration altogether, as my gut tells me that these games are just too basic to achieve commercial success today.
Because my focus is on solo developers, I think we can exclude anything from the Mainstream or Castaway Eras as well. Unless you have a multiperson team on the job, these games are just too big to produce fast enough to ever hope for profitability.
That means we'll be focusing on the more complex games from the Black Box Era ~~ by and large, the Black Box games that were not actually released in black boxes.
Also, even though I only read English and will therefore be focusing on North American NES releases, I will be using the dates Japanese games were released in Japan to better reflect actual development.
Ports, Tho π’
But now a thorny question.
Are we talking only about games developed specifically for the NES, or do we include ports?
Initially, I had thought to exclude ports. And I immediately ran into trouble.
Most obviously, there is the phantom of the FamiCom Disk System, which was never released outside Japan. For international release, many classic NES games were ported from disk to cartridge, sometimes undergoing significant changes along the way.
Then, too, localization itself often entailed major changes, particularly for story-heavy titles.
Excluding all these is artificial and very restrictive, so ports are in.
Filtering Even Further β
That still leaves a lot of games to look at.
I narrowed this down further in an extremely subjective fashion: which of these games am I as a dev personally interested in?
Actually, there were several games I wanted to include in my analysis that I just couldn't find enough information on, such as Valis 1, Uninvited, and The Wing of Madoola.
In the end, these are the games we'll be considering here:
| Game | Genre | Year | Studio |
|---|---|---|---|
| Elevator Action | Platformer | 1983 | Taito |
| Lode Runner | Platformer | 1983 | Doug Smith |
| The Tower of Druaga | Dungeon Crawl | 1984 | Namco |
| Urban Champion | Beat 'Em Up | 1984 | Nintendo |
| Hydlide | RPG | 1984 | T&E Soft |
| Castlequest | Metroidvania | 1985 | ASCII Corp |
| Ice Climber | Platformer | 1985 | Nintendo |
| Super Mario Bros. | Platformer | 1985 | Nintendo |
| DΓ©jΓ Vu | MacVenture | 1985 | ICOM Simulations |
| Castlevania | Platformer | 1986 | Konami |
| Dragon Quest | RPG | 1986 | Chunsoft |
| Deadly Towers | Action RPG | 1986 | Lenar & Tamtex |
| The Legend of Zelda | Action RPG | 1986 | Nintendo |
| Metroid | Metroidvania | 1986 | Nintendo |
| Castlevania 2: Simon's Quest | Metroidvania | 1987 | Konami |
| Faxanadu | Metroidvania | 1987 | Hudson |
| Super Mario Bros. 2 USA | Platformer | 1987 | Nintendo |
| Shadowgate | MacVenture | 1987 | ICOM Simulations |
| Mega Man | Platformer | 1987 | Capcom |
| Zelda 2: The Adventure of Link | Metroidvania | 1987 | Nintendo |

Notice that whole genres are missing (pure Puzzle games, Boardgamelikes, Shoot 'Em Ups), just because I'm not that into them.
For those unfamiliar with the MacVenture genre, it's a particular type of Point & Click Adventure that arose on the early Macintosh and eventually gave rise to the Mystlikes that were big business in the mid '90s.
What Do We Compare? π¬
I asked a while back for suggestions on how to compare the scope of finished products. Based on those answers I put together the following list of potential points of comparison:
I think it's worth capturing all of these for history, but the list is quite lengthy.
- Misc
- Title
- Year
- Studio
- Genre
- Launch Platforms
- Number
- Type
- Middleware/Tools
- Number
- Type
- Team
- Number of Developers
- Specialties Needed
- Time in Development
- Audience
- Size
- Friendliness
- Play Time
- Session Length
- Time to Beat
- Time to 100%
- Software Features
- Number
- Complexity
- Lines of Code
- Primary Availability
- Duration
- Post-Launch
- Length of Live Updates
- Length of Support
- Players
- Number
- Simultinaeity
- Modability
- Type
- Degree
- Mechanics (Primary and Mini)
- Number
- Components per System
- Degree of Nuance/Skill
- Dynamics per System
- Assets
- Number
- Size
- Type
- Characters
- Number of Playable Characters
- Number of Allies, if any
- Number of Neutral NPCs, if any
- Number of Enemies, if any
- Word Count
- In-Game World
- Area Number
- Area Size
- Accompanying Goods, If Any (e.g. printed manual)
- Number
- Type
- Risks
- Technical
- Have we used this technology before?
- Production
- Have we planned or developed this type of game before?
- External
- What does the market look like for this type of game?
- Is it likely to change?
- Organizational
- What resources and funding do we need?
- Technical
Now, some of these are not useful for our current purposes (Duration of Availability, Length of Support), and many are not available (Lines of Code, Middleware/Tools). Others need to be evaluated per-developer (Risks).
However, several remain that I think are worth digging into:
- Number of Developers
- Time in Development
- Time to Beat
- Number of Players
- Number and Type of Mechanics (Primary and Mini)
- Visual Assets
- Audio Assets
- Word Count
- Accompanying Goods
Data Sources π
In general, I drew information from the following:
- Playing the Games
- Watching Longplays
- HowLongToBeat.com
- MobyGames
- GameFAQs
- Wikipedia
- Other random pages found via DuckDuckGo
I took the sizes of ROM and RAM chips from this spreadsheet.
Data Caveats β
-
Team Size is not always reported and reports sometimes contradict one another. I've compared sources to try to find a reliable number.
In many cases publishing, localization, and QA workers are not reported. On the other hand, not all games required extensive localization, some publishers are hands-off, and QA was sometimes done by developers with other specialties.
In general, I've erred on the side of overcounting team members as we'd rather overestimate than underestimate the amount of work these games took.
-
Time spent in development is very seldom reported and sometimes depends on how you count, since many of these games are ports.
In many cases I've taken my best guess by finding individual developers who are known to have worked on specific games that were released serially. For example, I assume that work on Uninvited did not begin until work on DΓ©jΓ Vu ended.
This assumption does not always hold, as, for example, Super Mario Bros. and The Legend of Zelda were in development concurrently. However, it's the best I can do.
-
I have reported Visual Assets in terms of the size of the CHR-ROM chip that held them, measured in KiB.
This is because finding or making reliable counts of individual assets proved quite difficult.
Also, unlike audio or textual assets, visual assets are not easily directly comparable across genre. Like, how do we compare the monster graphics in Dragon Quest with the monster graphics in The Legend of Zelda?
Unfortunately, this didn't work well. The problem is, not every game actually fills up all the space it has available, at least not with tile data.
But I didn't realize how off this was until nearly the end of the project, so you can take these numbers as a vague guide or just ignore them! We'll dive into visual particulars for a subset of games as we go.
-
Mechanics are quite difficult for me to quantify, due both to their nature and my lack of game design training.
I came up with an informal rubric but it's not rigorous.
Here's what I came up with:
Mechanics Counted
| Game | Mechanics List |
|---|---|
| Elevator Action | Dodge, Shoot |
| Lode Runner | Platform, Dig, Collect, Solve Puzzles |
| The Tower of Druaga | Attack, Block, Unlock Doors, Solve Mazes, Break Walls, Randomization |
| Urban Champion | Dodge, Punch |
| Hydlide | Melee, AOE, Projectile, Collect, Leveling, Explore |
| Castlequest | Explore, Platform, Unlock Doors, Sokoban Puzzles, Melee, Map System |
| Ice Climber | Platform, Break Walls, Melee, Collect, Time Limit |
| Super Mario Bros. | Platform, Swim, Climb, Ranged, Power Up/Down, Invincibility, Stomp, Collect, Warp, Time Limit, Combos, Score, Find Secrets |
| DΓ©jΓ Vu | Explore, Examine, Talk, Collect, Unlock Doors, Melee, Move Limit, Complex Item Use, Solve Puzzles |
| Castlevania | Platform, Melee, Ranged, Power Up, Score, Time Limit, Ammo, Find Secrets, Invincibility |
| Dragon Quest | Explore, Melee, Spells, Items, Flee, Level, Talk, Shop, Fast Travel, Find Secrets, Light/Dark, Heal, Complex Foe AI, Search, Unlock Doors, Solve Mazes, Solve Puzzles |
| Deadly Towers | Explore, Melee, Ranged, Shop, Collect, Level/Power Up, Find Secrets |
| The Legend of Zelda | Explore, Melee, Ranged, AOE, Ammo, Power Up, Talk, Shop, Fast Travel, Find Secrets, Light/Dark, Collect, Heal, Block, Traverse Terrain, Solve Mazes, Solve Puzzles |
| Metroid | Explore, Platform, Ranged, AOE, Power Up, Ammo, Find Secrets, Break Walls, Unlock Doors, Freeze Foes, Climb Foes, Traverse Terrain, Complex Weapons |
| Castlevania 2 | Explore, Platform, Melee, Ranged, Power Up, Day/Night Cycle, Ammo, Find Secrets, Talk, Shop, Level, Equipment, Collect, Heal |
| Faxanadu | Explore, Platform, Melee, Ranged, Power Up, Ammo, Find Secrets, Talk, Shop, Level, Equipment, Collect, Heal, Break Walls |
| Super Mario Bros. 2 USA | Character Select, Platform, Power Up, Unlock Doors, Slot Machine, Heal, High Jump, Carry, Ride, Find Secrets, Collect, Build (Limited), Climb/Sink, Slip, Fly, Warp, Invincibility, AOE, Subspace |
| Shadowgate | Explore, Examine, Talk, Collect, Move Limit, Melee, Ranged, Spells, Complex Item Use, Solve Puzzles, Solve Mazes |
| Mega Man | Platform, Stage Select, Upgrade, Ammo, Complex Foe AI, Complex Weapons, Ranged, Ride, Heal |
| The Adventure of Link | Explore, Platform, Melee, Ranged, Defend, Spells, Power Up, Talk, Level, Collect, Find Secrets, Heal, Solve Mazes, Solve Puzzles, Traverse Terrain |
Also,
- I still don't know how to do statistics.
- There's lot of data and I could've made mistakes.
The Data π
As with my Queer Devs on Itch post, the primary table here is too large to digest.
So, I'm placing it under a spoiler block and will break it into sub-tables below.
The Full Data Table π
| Game | Team Size | Dev Time | Time to Beat | Play- ers | Mech- anics | Visual Assets | Songs | Sound Effects | Word Count | Goods |
|---|---|---|---|---|---|---|---|---|---|---|
| Elevator Action | 1 | 1? | 0.5 | 1 | 2 | 8 | 2 | 9 | Trivial | Manual |
| Lode Runner | 1+41 | 4 | 3 | 1 | 3 | 8 | 1 | 8 | Trivial | Manual |
| The Tower of Druaga | 4 | 1 | 2 | 1 | 6 | 8 | 5 | 10 | 500 | Manual |
| Urban Champion | 4 | 1 | 0.5 | 1 | 2 | 8 | 3 | 14 | Trivial | Manual |
| Hydlide | 5 | ? | 3 | 1 | 6 | 8 | 3 | 7 | Trivial | Manual |
| Castlequest | 4 | <1 | 2 | 1 | 6 | 32 | 5 | 5 | Trivial | Manual |
| Ice Climber | 7 | 1 | 2 | 2 | 5 | 8 | 3 | 17 | Trivial | Manual |
| Super Mario Bros. | 5 | 1 | 2 | 2 | 13 | 8 | 6 | 25 | Trivial | Manual |
| DΓ©jΓ Vu | 11 | 1 | 3 | 1 | 8 | 256 | 7 | 252 | About 11,0003 | Manual |
| Castlevania | 5+54 | 1? | 3 | 1 | 9 | 645 | 16 | 37 | Trivial | Manual |
| Dragon Quest | 13+76 | 1 | 10 | 1 | 17 | 16 | 25 | 23 | 7,000 | Manual, Map |
| Deadly Towers | 4 | 5 | 4 | 1 | 7 | 8 | 257 | 11 | 500 | Manual |
| The Legend of Zelda | 8 | 1 | 8 | 1 | 17 | 8 | 12 | 25 | 500 | Manual, Map, Hint Book |
| Metroid | 12 | 1? | 6 | 1 | 13 | 648 | 12 | 21 | Trivial | Manual |
| Castlevania 2 | 15 | 1 | 4 | 1 | 14 | 128 | 9 | 53 | 1,000 | Manual |
| Faxanadu | 8 | 1 | 6 | 1 | 14 | 1289 | 16 | 25 | 2,500 | Manual |
| Super Mario Bros. 2 USA | 20 | 1 | 5 | 1 | 18+110 | 128 | 14 | 19 | Trivial | Manual |
| Shadowgate | 11 | 1 | 3 | 1 | 10 | 128 | 1511 | 13 | 11,000 | Manual |
| Mega Man | 8 | 1? | 3 | 1 | 9 | 6412 | 16 | 35 | Trivial | Manual/ Poster |
| The Adventure of Link | 11 | 1 | 11 | 1 | 15 | 128 | 21 | 23 | 700 | Manual |
| Median | 8 | 1 | 3 | 1 | 9 | 32 | 12 | 21 | <500 | 1 |
| Mean | 9 | 1 | 4 | 1 | 9 | 63 | 11 | 21 | 1,880 | 1.2 |
Important Footnotes π
-
The original Lode Runner had one official developer; however, he worked with two friends for several years to develop the prototype, and an additional coder worked on the FamiCom port.
-
Of DΓ©jΓ Vu's 25 sound effects, only 9 are used regularly.
-
I could not dump DΓ©jΓ Vu's script, so I'm using Shadowgate's value here as they are in the same ballpark.
-
Akumajou Dracula had between 5 and 10 developers, depending on the source. I think the discrepancy may have arisen because Western commentators often overlook Vampire Killer and Haunted Castle.
-
Castlevania's graphical assets are crudely estimated by dividing the capacity of PRG-ROM chip in half: 128/2. This is because Castlevania uses CHR-RAM and I could not easily get a size for how much unique data is loaded into RAM over the course of the whole game.
-
Dragon Quest has 13 developers, more than I expected. But there are 7 more when you take into account the localization and redesign that went with it becoming Dragon Warrior.
-
Deadly Towers only has 14 songs if you exclude minor variations.
-
Metroid's Visual Assets estimated from PRG-ROM: 128/2. See footnote 5.
-
Faxanadu's Visual Assets estimated from PRG-ROM: 256/2. See footnote 5.
-
Super Mario Bros. 2 USA is the only game in the lot to have a minigame, and that minigame has only a single mechanic.
-
Shadowgate has one unused song in the ROM, so we know this additional asset was created.
-
Mega Man's Visual Assets estimated from PRG-ROM: 256/2. See footnote 5.
Immediate Observations π
- There are 12 studios represented, and one game, Lode Runner, that fits well with our current understanding of "indie".
- 7 of these games are by Nintendo, which makes sense given the way they managed their platform in those years.
- Konami and ICOM both show up twice, which really only speaks to my tastes.
- Every other company shows up only once.
- The list is about evenly split between (subjectively) large and small games with a few in between.
- It looks like '86 and '87 are way overrepresented, but this is due to three things:
- Many more games were released in these years than in the years prior. πΉ
- It's easier to find information in English about games as time goes on. π
- As games got more complex, they get more interesting to me. π€
- Related, dev team size, game length, and game complexity generally all increase over time.
- π₯ Team Size is most often 4 or 5.
- Elevator Action was essentially solo-developed (like a lot of Taito's earliest games).
- Many larger games have a team of 11.
- No team is larger than 20.
- These numbers fit well with current indie expectations.
- π Dev Time is almost always one year.
- This is the only value whose mean, median, and mode are all the same, so it seems a very reliable trend.
- However, it might not be, as this is also the only column with five uncertain entries.
- There is one game developed in less than a year, and two that took 4 or 5.
- Notably, Lode Runner, the game I labelled "indie", is one of the two games that took 4-5 years. But note that Lode Runner had a popular sequel, Lode Runner Championship Edition, which took only an additional year to make by reusing the engine and all assets.
- This is the only value whose mean, median, and mode are all the same, so it seems a very reliable trend.
- β³ Most games are beatable in about 3 hours.
- This does not take replayability into account, although most of these games are arguably highly replayable.
- Two early games, Elevator Action and Urban Champion, take only 30 minutes for a skilled player to finish.
- Dragon Quest and The Adventure of Link are way above the norm in requiring at least 10 hours to beat.
- 1οΈβ£ All but two games are single player, so I will omit this column in the future.
- Of the multiplayer games, both are two player and only Ice Climber is simultaneous two player.
- Mechanics counts vary widely. Breaking the data down by year or genre reduces this variance.
- π§° Yet, we can see that most games had 10 or fewer Mechanics, most often 6, and none had more than 20.
- As mentioned above, minigames are practically nonexistent.
- Half these games fit all their Visual Assets into only 8 KiB! That means 512 or fewer 8x8 tiles for all sprites, backgrounds, and UI elements!
- This is less surprising if we recall that mappers allowing for more graphics data weren't available in the earliest years.
- The next largest cohort of graphical sizes is 128 KiB.
- I was a little surprised to see so few games between 8 and 128 KiB.
- Only one game reached 256 KiB.
- Few NES games of this period focused on text.
- 10 games have less than 500 words.
- 4 more have less than 1000 words.
- Only 5 games have 1000 words or more.
- πΌ Soundtracks tend to have about 11 songs.
- Although, many of the later games have more and a couple have as many as 25
- The earliest games tend to have 3 to 5 songs while Lode Runner has only 1!
- π Sound effects are much more emphasized than music, with most games having between 21 and 25.
- Castlequest has the lowest number of sound effects, 5 in all, while Mega Man has the second-most at 35 and Castlevania 2 has a wopping 53!
- Keep in ming that both arcade games and early home computer games, from which many of these games derived, were typically music-free until 1984.
- Only three games shipped with anything more than a simple printed manual, so I will omit this column going forward.
- Overall, this is a small space to work in; yet, the largest games examined here (namely, the MacVentures along with the later Metroidvanias, RPGs, and Action RPGs) are quite a lot of work for a solo dev!
That's It for Part One β
In the next post, we'll dive into detailed examination! π
