Submission #3352: MESHUGGAH & Randil's NES Indiana Jones and the Last Crusade (Taito) in 02:57.56

Since my 1st project (Indiana Jones and the Temple of Doom) has emulator issues, I thought I should give a try for another Indiana Jones game pulled from my SDARunsWithoutAPublishedTAS list.

When I completed my 1st WIP, Randil pointed out obvious mistakes. His WIP motivated me to learn the NES assembly and debug the RNG. If you want more details, check out the thread or the game resource page I wrote.

Also this TAS based on Randil's WIP that I optimized really good.

Regarding the game, there's an awful "4 frame rule". The game polls input at every frame, but the actions only "processed" if it's on the 4th frame.

The biggest challenge for this game was getting the best Holy Grail layout. There's 9 layout which ends on 6 different tiles. The fastest one needs only 4 sideways and ends on a bird. For this, you need 87, 116 or 164 RNG calls. These are the results of my lua scripts.

However, RNG cycles depends on which enemy you kill, how enemies you killed, what's your route in stage 2 and your health (cycles the RNG if you lose 66% of your health). That's why I have to go back and re-enter the last door at stage 2 to get the correct 116 RNG calls. It's impossible to get "only" 87 calls, or getting 164 RNG calls without killing both 3 guys at stage 2 (which results in a slower completion time. The closest one I had always (with faster time) was 163 calls. What a pity). The only possible improvement should be somehow get up with a clever idea that I never thought of, which I don't think it exists. I tried MANY things.

Randil's comments Back in 2010 (I think), after having watched the run on this game on SDA, I decided to make a TAS of this game, just for fun. The TASing went smoothly (the first level probably took the longest to make) until I reached the final level - I couldn't get a good route no matter no matter how much I tried. I speculated that the route you get is decided when you start the game, but without the motivation to analyze the game code, I abandoned the project.

In September this year, MESHUGGAH announced that he was working on this game. I posted my movie and my RAM addresses and wished him good luck with his run. It turned out that a lot of the input and findings I made for my run was used in this run, so I was put as co-author.

I think this is a short and nice TAS, and I hope you will enjoy watching it. }}

This is where luck manipulation starts. Starting the game at every 4th frames (I start the game as fast as possible: 0th frame) spawns the 1st enemy near to the ladder of stage 1.

I made use of the "bouncing off from walls kick", "very fast kill at doors" and "stopping as fast as possible".

This route is faster than Randil's and my old WIP's original route. As earlier mentioned, I have to go back and re-enter the last door to get the exact 116 RNG calls.

Randil made a very good effort on this place too, however I found a rather good trick to eliminate the top enemies as fast as I can. Notice that enemy spawns depends on where are you at the exact moment the game spawns an enemy.

The result of luck manipulation. This is the fastest map ending with the fastest tile. I added a variation of moving left for entertainment purposes. It doesn't changes completion time. The cup is always at the same place, it can't be manipulated.

I think the game is fast enough now, much faster then the SDA run. If you have any suggestions about adding more entertainment, feel free to say and I will hex edit in.

See you until my next TAS =)

Baxter: Well-done TAS. Accepting!