Submission #3758: TrenchAce's NES Top Gun in 26:33.46

Console Nintendo Entertainment System Emulator FCEUX 2.1.5
Game Version USA Frame Count 95765
ROM Filename TopGun(U).nes Frame Rate 60.0988138974405
Branch Rerecord Count 854
Unknown Authors TrenchAce
Game Top Gun
Submitted by TrenchAce on 11/15/2012 4:53:57 AM

Submission Comments

I'm resubmitting this only because of the new Vault tier. It is 100% unchanged from my previously cancelled submission, which I only cancelled due to the 50/50 audience feedback (note: I had previously submitted as tag "Gamerskillsfull", and have since changed my tag to "TrenchAce"). An encode is already on YouTube:

Previous comments (with minor edits):


  • Another classic game AVGN game bites the dust.
  • No damage/death taken
  • The bigger missiles were chosen in levels 2, 3, and 4 in order to destroy the target faster. In level 1, there is no target, so it doesn't matter what missiles you choose. Choosing the first set of missiles saves 2 frames.
  • No frame-by-frame editing was used for the refueling (I found it easier to do those parts at normal speed, and just use savestates if necessary for increasing speed).
  • Frames were saved by hitting the lowest altitude during gameplay before landing. Otherwise, the game takes longer to go down to landing altitude automatically.
  • This game has a lot of downtime. In order to keep things visually interesting, I attempted to destroy enemies whenever possible for more "entertainment"
  • There weren't very many re-records since the game has so much downtime where you can't really do much anyways. Most were spent on the refueling sequences, and many on the enemy kills, too.
  • Luck manipulation is not an option, as all enemies always do the same thing, and sprite popping occurs on your screen no matter where you are and which way the cockpit is facing. The only changing factor is enemy gunfire, and sometimes your missiles will take a different path. Exploiting this was possible a few times.
  • FCEUX 2.1.5


  • Making sure missiles only were shot at the target in the end of levels (if there's a sprite in your sights, the game automatically will take out those with the missile instead of the target)
  • Increasing refueling speed
  • Hitting many of the sprites, since you most hold "up" or "down" to be facing up or down.
  • Avoiding all damage
For entertainment value I started this run thinking I could throw gunfire at an enemy and hit them just as they appear; unfortunately, as I found, sprites in this game appear at the SAME number of pixels from the center of the screen no matter which way you are turning, unless there are already sprites on screen as a frame of reference for the game.
EDIT 2: The game does not track lateral left-right movement in the X-Y plane. Basically, it's impossible to get the cross-hairs over to the appearing sprite. For example, if I know an enemy will appear to the right, so I go back a few frames to try to go right and line up a shot that hits them right as they appear, I can try all I want but will not be able to and it will still appear to the right.
EDIT 3: The game also does not track forward movement, and as a result, speed is irrelevant both in gameplay and in the landing sequence. It does, however, check altitude, which is why I need to be at 10,000 feet. For some reason, the game will go up 2,001 feet in altitude before coming back down to the landing sequence, and it does that no matter what altitude I am at, so beginning at the lowest altitude is vital to do.

The (Epic) Game Itself:

  • Top Gun is a slow and boring game (I just re-watched it at 6400% emulation speed), but it's also a misunderstood game. The game is not quite as bad as people say it is, and the concepts turn out to be pretty good. In fact, the more I played it, the more I liked it and appreciated what the developers were trying to do.
  • Landing and refueling - once you figure out what it is you're actually trying to do - is actually well done and gives you a sense of accomplishment when you do it. The problem is with the graphics, unclear and confusing instructions, and overall gameplay.
  • Remember, this game was ported onto the NES. It was originally made for the PC and Commodore 64. In terms of a flight simulation game, Top Gun actually does fairly well. As a video game (esp. for the NES), it hasn't been universally accepted nearly as well.
  • Some of the game's flaws, such as epic sprite popping, lack of level detail, too much downtime, and overall boring gameplay can be forgiven since the game was released in 1987!
  • The concept of the game is supposed to simulate how actual flight/fight training and combat missions would go in real life.
  • The landing sequences were meant to force the player to monitor his visual distance from the landing target, altitude (up vs. down), speed, and direction (left vs. right). The game has a very nice balance for this once you figure out how to monitor this data.
Unfortunately, due to the confusing gameplay, most people never ended up fully experiencing the game and the good things that it has to offer.
But I've said it before, and I'll say it again: this game takes a LOT of patience to master. But it can turn out to be very rewarding when you accomplish the harder aspects once you know what to do.
  • Why are there only 854 re-records? Can it really be fully optimized in such a long game with so few re-records?
Short answer: yes. Long answer: much of this game is downtime. Re-records were spent mainly on locating sprite-pops, and going back to kill them using only 1 or 2 records at a time. Also, I was able to make this almost exclusively on a gamepad, and as such it was actually easier to do boom/basket frame-perfect refueling sequences (and other parts of the game) in real-time; I was able to do them on the 1st time every time because once the boom lines up over the basket (the "X"), it will instantly connect. During landing sequences little re-recording also was needed, as speed/altitude are able to be monitored easily while playing in "slow-motion" (I tried frame advance to keep speed at 288 and altitude at 200, but it is impossible to stay exactly on these targets), and only had to re-record a few times if I strayed too far from the target numbers. This is of little importance anyways as it does not affect the speed of level-end. Even if I had to redo a whole section, re-records did not add up very high, yet the TAS was still able to be fully optimized.
  • Why bother re-submitting this?
I fully expect this submission to be rejected for official publication. I accept this, as I believe official publications should be entertaining to most people. On the other hand, the new vault tier is an excellent opportunity for this site to accept many games, such as Top Gun, that some people may still find entertaining because they know the game first-hand, and how frustrating it can be. This new tier is great for the site because it can now become a nearly comprehensive record-keeping site; this is the only reason why I am re-submitting this, even if some people may think this should be officially published.
  • Why didn't you use frame-advance in some parts?
As I said before, since I almost exclusively used a gamepad to create this, some sections of the TAS were easier to create in either real-time or simply slow-motion. I would use savestates often, but since enemy sprites cannot be manipulated prior to popping, I would not have to go far back into a savestate to begin attacking an enemy. Also, since it is easy to do frame-perfect refueling sequences (once you know what you're doing), and it still looks just as good as TAS, it was unnecessary to use frame-advance for the whole TAS. However, most of the game was played in either slow-motion or frame-advance.

adelikat: Accepting for publication to the Vault

Ilari: Processing

Last Edited by Ilari on 11/29/2012 10:54:12 PM
Page History Latest diff List Referrers