Tool-assisted game movies
When human skills are just not enough

Submission #3155: Brandon Evans's NES Bugs Bunny Crazy Castle, The in 40:50.61

Console: Nintendo Entertainment System
Game name: Bugs Bunny Crazy Castle, The
Game version: USA
ROM filename: Bugs Bunny Crazy Castle, The (U).nes
Emulator: FCEUX 2.1.4a
Movie length: 40:50.61
FrameCount: 147279
Re-record count: (unknown)
Author's real name: Brandon Evans
Author's nickname: Brandon Evans
Submitter: Brandon
Submitted at: 2011-05-21 06:27:35
Text last edited at: 2012-11-25 08:31:02
Text last edited by: Brandon
Download: Download (20762 bytes)
Status: published
Click to view the actual publication
Submission instructions
Discuss this submission (also rating / voting)
List all submissions by this submitter
List pages on this site that refer to this submission
View submission text history
Back to the submission list
Author's comments and explanations:

The Bugs Bunny Crazy Castle is a frustrating action puzzler. In this run, I attempt to complete the game as fast as possible while trying to also maximize my score. I primarily use heavy luck manipulation to achieve this in around 40 minutes, which is impressive considering that the game is 60 levels long.


(Link to video)

With Input Display:

(Link to video)


Hello, and welcome to my first submission! I started this run during a break from my other WIP, Spider-Man/X-Men: Arcade's Revenge. Although many think this might not be the best game choice, I disagree for the following reasons:

  • The luck manipulation in this game makes it far from trivial.
  • There are ways to spice this up a little, and I do my best to use all of them.
  • Regardless of the AVGN's opinions, I liked this game a lot as a kid, and I bet a lot of other people here did too, so I thought a proper run should be made for it.

So, I went for it, and I think I was fairly successful. For more information on the run, please read watch the encode with subtitles enabled, as I document nearly everything I do in the run throughout.

Game Objectives

  • Emulator used: FCEUX 2.1.4a
  • Attempts to complete the game in the fastest time possible
  • Aims to maximize score / kills without losing time
  • Heavy luck manipulation


In the thread for this game, Mysterypea had already constructed a 33 level WIP. Using it as a model, I tried my best to obsolete it. Here is a table of our respective times. Note: I define a level's length as the difference between the frame in which Bugs' sprite shows up and the one in which the "1UP" bubble appears.

Level My Version Mysterypea's Version Difference
1 1682 1686 -4
2 2117 2209 -92
3 1628 1669 -41
4 1289 1329 -40
5 1981 2282 -301
6 1621 1829 -208
7 1262 1492 -230
8 2172 3166 -994
9 1903 2638 -735
10 1682 1974 -292
11 1460 1980 -520
12 1200 1400 -200
13 1710 2768 -1058
14 2487 2663 -176
15 2123 2519 -396
16 2070 2120 -50
17 1720 1879 -159
18 1969 2325 -356
19 2489 2618 -129
20 2484 3547 -1063
21 2299 2553 -254
22 2203 2372 -169
23 2019 2515 -496
24 1439 1579 -140
25 1783 1949 -166
26 1852 2223 -371
27 1720 1724 -4
28 1574 1894 -320
29 1939 2093 -154
30 1244 1589 -345
31 2317 2345 -28
32 1853 2149 -296
33 1337 1446 -109
34 1882
35 1670
36 1913
37 1692
38 2296
39 1614
40 1843
41 2863
42 1527
43 2801
44 1462
45 1606
46 1774
47 1371
48 2055
49 2145
50 1887
51 1364
52 2416
53 1674
54 2232
55 2656
56 1834
57 2247
58 1540
59 3799
60 2003
Total 94537 70524 9896
Average 1576 2137 300

In addition to this, I ended the game with 77000 points. It should be noted that the re-record count is unknown as it was lost somewhere in the editing process.

It's worth noting that Level 26 of could take a different path and get the exact same time. Here is a YouTube stream of that path, and here is the movie file.

Future Runs

If the reception to this run is positive, I might attempt a quad run of the Crazy Castle GB games. Considering that you can move forward using up or down as opposed to pressing left and right, and that there are many frames in which input is rendered moot, this would be the perfect candidate for such a run. UPDATE: Crazy Castle 4 apparently does not move forward when pressing up or down, so a triad run of the first three games would probably be a better option.

Possible Improvements

Although I have a fairly strong understanding of the AI for this game, I certainly don't have an exact formula. If I had such a thing, it might be possible to make every enemy irrelevant. It would also be helpful if I had a script that allowed me to see other segments of the screen so I can see how my input influences an enemy that is off-screen.

Known Improvements

nitrogenesis has reported that I transition from level to level incorrectly. Instead of pressing Start, I can press A and save one frame per level. If this was implemented for this run, then, it would be exactly a second shorter. Still, I don't think this improvement alone warrants a new submission, so this will have to wait until a bigger improvement is found.

Special Thanks

Thank you to Mysterypea for his original WIP, Randil for his not so accurate, yet helpful lost frame calculating script, andymac for his "AVI Heads Up Display" script, and feos for his work with andymac on the aforementioned script. I also appreciate all the feedback I've gotten so far.

Suggested Screenshot

Closing Statement

I put a lot of work into this run, so I hope you enjoyed it! However, if you didn't, all I can say is...

Haters gonna hate. Credit to NrgSpoon.

Nach: This was a tough one, tons of yes votes, but several strong dissenters as well. I have to agree, this isn't the most entertaining of games. However, this game is deceptively simple. Everyone who knows this game well (which seems to be the majority of viewers) knows how tricky it really can be. The lousy play control, unpredictable enemies, and confusing route chooses make this game tricky to play, let alone play well. Just paying close attention to level 27 for example shows how the game isn't so clear cut as it may seem.

Since this run requires TAS techniques such as route planning (which changed WIP to WIP), and luck manipulation, I feel this run deserves consideration to be published. The question is then down to: "Did it entertain?"

It definitely had its moments, and knowing the game well really enhances that. But it was rather long and repetitive. Although I sat through this run twice, most people won't want to watch this in a single sitting. Thinking about whether that is acceptable or not, I realized we have published several multi-hour RPG TASs in the past, which no one is going to be sitting through in a single sitting. Therefore I feel that this should not be a barrier to publication, especially since the levels in this game are pretty short, and can be watched individually without missing anything.

Those that don't know the game will probably skip this run, and those that do will probably enjoy it immensely. I decided to err on the side of caution and go with the majority of the audience, and accept this run as a first run for this game. I hope future runs will improve the routes further and find more exciting ways to manipulate the enemies and shave off two or three minutes.

Brandon: Publication underway! Thanks for allowing me to do the one thing I haven't done for TASVideos; make a published TAS. :)

Similar submissions (by title and categories where applicable):