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

Adelikat / Movies Commentary

<< Adelikat

This is a listing of the games I have made a published TAS plus information & commentary on the making of the TAS(es).

Table of contents [expand all] [collapse all]


Zanac was my first submission to TASvideos (known as nesvideos then). I was new to the concept of TASing and wanted to do an easier game before I took on my main project which was Mike Tyson's Punch-Out!! Zanac didn't already have a TAS (I didn't feel qualified to try to improve anyone else's). I had played the game a lot as a child and thought it might be a good TAS. Even though it was an auto-scroller, there were many things to consider in terms of speed. It had many boss fights to optimize, fairies to instantly kill bosses (and the ability to choose to some degree which bosses they killed), and level warps. There was also a glitch to make some turrets of the boss remain closed to save some time. So I did an "all levels" movie as my first submission. I used Famtasia since it was the rerecording emulator I was familiar with. It was a little bit outdated at the time of the submission though. At the time, I thought Zanac made a good TAS, but looking back, it didn't. Bisqwit was very generous to publish it. It just is too long for what it is.

Later, I found out that FCEU could record soft and hard resets. This meant I could record the level select code (13 soft resets) and make just a 1 level run. This was controversial since it is a cheat code, but I felt that it was a better representation of the game to skip to the last (most interesting) level. This way the game does not get too repetitive. So I made a new submission. This eventually obsoleted the previous movie. As of date this is the largest margin of obsoletion in TASvideos history. ;P

The current published movie is an improvement based on a discovery I made later. I found that I could upgrade the rapid fire fully before the boss fight using a trick (discussed in detail in the submission text).

Overall, I like this movie and think it makes a good TAS but certainly not one of the best. The game itself is very pleasing to watch, and the run shows off some perfect reflexes. It also demonstrates many interesting features of the game in just 3 minutes of gameplay.

Mike Tyson's Punch-Out!!

This is THE movie by adelikat. :) I have more pride about it than any other.

The thing about this game that non-experts would not know is the insane amount of luck manipulation involved. Almost every hit and every action by the opponent was random and very unlikely. The Mike Tyson fight for instance is manipulated so he does his fastest pattern for every punch. The odds of this happening in any given fight are about 1 in a million. The Bald Bull 2 fight is even more random: every punch landed on Bald bull 2 for a star is extremely random - about a 1 in 32 chance - and there are about 14 of these punches thrown, not to mention his pattern must be manipulated. All fighters must be manipulated when they hit the canvas. The count on which the get up is random, and it must be manipulated so they get up on 2 (or 3 in some cases). In fact, there is only 1 fight which as no randomness at all - Great Tiger.

This game is what led me to tool assisted speedruns. This has been my favorite game of all time, and one that I competed heavily attempting to get console record times for each fight. With emulation (Nesticle at the time) came the ability to test strategies and find the theoretical perfect record times. Nesticle could record movies, but you couldn't use savestates, so I and others had to resort to making a savestate at the beginning of a fight and play it many times in order to get a good demo of a particular strategy. I wished that there was a way to use savestates while recording a movie so I went looking for such a thing. So TASes were the "holy grail" for me when I stumbled on the site. Later I found that it also used things like slowdown and other tools to achieve perfect times. So I began using Famtasia to make demos of each fighter and try to come up with different strategies. At the time I never had any desire to make my own TAS.

Sometime after, Phil and Genisto submitted the first ever full run of MTPO. There were several strategies I hadn't even considered before. I was blown away by this movie and it inspired me to make my own TAS. Theirs did not use the clock-stopping glitch, so I wanted to make a movie that used it. So I submitted my 2nd ever movie. My intent was to have it published along side Phil & Genisto's movie. This movie achieved some excellent clock times, including improving the Piston Honda time by 2 seconds. However, as a TAS newbie, I didn't consider trying to make sure it was as fast as possible in real time as a secondary goal. I didn't take time to manipulate it so the opponents don't try to get up and fall or try to manipulate them to get up on the earliest count. At the time I didn't care about this but after it was brought up, I saw that it would make the run look more polished, so I set out to improve it. This movie was eventually rejected.

Since I had an improved fight that wasn't due to the cs glitch, I saw that I could improve Phil & Genisto's movie too. So I made both a cs glitch and non cs glitch movie at the same time. This required a lot of TASing and a lot of testing. I wanted to make these movies absolutely perfect. I spent about 9 months making these movies and submitted both at the same time: non-cs, cs glitch

The non-cs glitch movie managed to improve 4 of the fights over Phil & Genisto's movie. I was very proud of this movie and felt I had found everything there was to find in this game. This was ironically a sad feeling though. I had been looking for new things in this game for 15 years at this point. It was sad to see it "done". (Fortunately I was wrong).

This movie was eventually published, and was my 2nd publication. It was also the first one to have an .avi file.

The cs glitch movie was very nice too, I improved several of the cs glitch times as well some other fighters. I also made it 30 seconds shorter in real time. Unfortunately, this movie was not interesting to the TAS community. That is sad, since I think it is a good movie and compliments the published movie well. I cancelled it when I saw that there wasn't any interest in it.

Later I stumbled on an improvement to Soda Popinski while messing around playing the game for fun, so made a new submission. I improved Soda Popinski from 0:47.00 to 0:45.25. I also found a way to improved Don Flamenco II from 58.97 to 58.61. I also corrected a real time (not clock time) error on Super Macho Man that no one noticed. At one point he did 4 super spins when it is possible to manipulate it so that he did only 3.

Later a console speedrunner of this game named Matt Turk found an improvement to Mr. Sandman so I made the current published run. This improves the Mr. Sandman time from 2:19.00 to 2:18.00. Only the last 3 fighters were redone, the rest is a copy of the previous.

Cobra Triangle

This is still a favorite of mine. I think this game is well suited for TASing because it has so many different types of levels. The racing levels are hard to optimize since they are 3D-ish. There are also traditional boss fights.

There are more linear levels and some odd ones. Like saving children from evil boats and UFO's (yeah this is a weird game). These are my favorite levels because I can play around. I let all but 1 child die (I am such a nice guy :P) in order to save some seconds on the points tally. I then perform some crazy antics as I try to prevent the boats from taking the remaining child. At one point I let a boat get the child and then ram it to the wall; keeping it from movies away. Anyone who has played this game would know who impossible that is to do. So that is a great TAS moment.

When I began my TAS Mike Tyson's Punch-out!! I thought it would be my only TAS. I had no interest in doing other games. But in the process of making my submissions of this game, I found how addicting the process is :) After my first MTPO movie was published I was hooked on making TASes. I started to look for other games to TAS as well. I was very familiar with Cobra Triangle and remember being disappointed when I watched the published movie at the time. There were many tricks I was familiar with not utilized in that movie. So I began and eventually submitted my first Cobra Triangle movie. It was a solid 44 seconds faster than Walker Boh's movie.

After this submission I found a new strategy for level 1 to get more speed upgrades. This was a big improvement so I cancelled it and later submitted what is still the currently published movie which improved the previous by 10 seconds.

Interestingly, I used FCEU .12 and didn't know anything about RAM addresses. Later I set out to improve it with the new tools of .16. I tried the first 4 levels and couldn't improve it by a single frame :O So I decided that this was indeed a high quality TAS. So I know of no way to improve it and therefore have no plans to redo it.

Marble Madness (NES)

This game has a 16 frame rule (the same type as the Super Mario Bros. 21-frame rule). This was a frustrating thing to discover. I really enjoy this movie. It is a 3D environment even though it is a 2D platform. This makes the TAS logic unique compared with other NES games. The route planning is tricky as a result and very tedious to TAS. The results are worth it, as it makes a great TAS.

After Cobra Triangle, I decided to give this game a try so this was my 4th game I completed a TAS for. Again, I was very familiar with the game and thought the published run could stand to be improved. So I made a submission. I found some improvements immediately (one suggested by JXQ who was a complete newbie at the time : and made a revised version. This one was published and stood for a long time. Later, with the aid of new tools like memory watch and made the previous published movie. Due to the 16 frame rule, I was pretty sure this movie would not be improved. Unfortunately I was wrong and was later pwned by Deign. He hexed in some of my levels, so I got to be co-author without lifting a finger; pretty sweet deal. The result is this movie.

Dragon Warrior 4

I began this TAS after having more published movies (Zanac, MTPO, Marble Madness, Cobra Triangle). I was starting to feel confident in my TASing abilities so I wanted to push my limits and do a larger project. This certainly fit the bill! I spent almost a year doing this project and it was a journey to say the least.

It was worth the effort. This is one of the movies I am most proud of. I think it is the most extreme example of luck manipulation on TASvideos. So many unlikely events happen like there are everyday occurences. Critical hits happen every time they are useful and each one has a 1/64 chances of success (some have a 1/256). All random encounters are avoided unless I need an item and want an enemy to drop one (which is often extremely unlikely) or I want to kill my characters to avoid backtracking. In this case I have my characters hit themselves and manipulate it for a critical hit, which I find hilarious. :)

Not only is luck manipulated, but I went through a tedious process of not only finding the desired outcome but minimizing the amount of frames it takes to manipulate it. I did this by advancing previous menus with different buttons, holding them for different amounts of frames etc. The result is that nothing takes more than 20 frames to be manipulated. The result in phenomenal, because at full speed you never see a noticeable wait period between actions. This makes the TAS look very polished.

Now that new tools are available (such as basicbot), luck manipulation could be taken much further than this. But due to poor emulation, movies of this game are terrible about desyncing. I attempted a v3, but couldn't handle losing my work constantly so I have no intentions of ever improving this.

Mega Man 6

This was my first team project. I am a fan of the NES Megaman series but this game was the one I was least familiar with. But AngerFist's enthusiasm inspired me to work with him on this project. He did all the levels and did all of the boss fights. This worked out very well since I don't care much to TAS the levels in megaman games but I love boss fights (in any game generally). AngerFist loved TASing the levels but struggled on boss fights. We submitted our first version of the game in January of 2006. It was a nice improvement to the previous movie by Kyrsimys. But there were some small minor improvements here and there. About 6 months later we submitted our improved version, which is the currently published movie.

Working with AngerFist on this project was a pleasure. He is a hard worker like me and we worked very well together. I would gladly work with again on some future project.

Currently this movie is being improved by Shinryuu.

The Flintstones

This was my first completed project with Brushy. Previously we worked on a movie of Kabuki Quantum Fighter. The movie was never completed because Nitsuja submitted his movie while I was TASing the final boss. D:

Anyway, afterwards we decided to team up on this game. Like Kabuki, Brushy did mostly levels, and I did mostly boss fights. Both Flintstones and Kabuki were games I knew nothing about prior to TASing them. I thought this would really help me grow as a TASer to have to start from scratch with a new game, and learn all its mechanics.

This movie came out pretty well, though it isn't one of my favorite runs. I think this game is good for TASing, but certainly not one of the best.

Interestingly, this movie was the first I made using FCEU .16 and the memory watch feature. This was the first game I used debugging to find important RAM values useful for TASing. Most of the improvements came from discovering the quirky nature of boss health in this game.

Wrath of the Black Manta

I was working on this movie while working on Flintstones around February of 2006. Originally I planned to team up with Brushy but we decided otherwise later. I don't know why, but I had always wanted to TAS this game. The game has very awkward controls and is a pain to TAS but for some reason it still appealed to me. So I set out to improve Walker Boh's movie.

In the process I discovered how to glitch myself in walls in order to save time. This was the first major glitch I had ever discovered in a game! I was quite excited about this.

I was becoming more competent with memory watch and I think this movie shows it. I am very happy with the level of optimization in this run and I find it entertaining to watch. It shows an elegant maneuvering in a very un-elegant game.

Duck Tales 2

This project started when Brushy made a WIP of level that he is so famous for doing. I decided to pick up the run and work with him. I ended up doing the rest of the game due to Brushies fading interest in it and my addiction to the game. I ended up working non-stop on the run. Duck Tales games are very fun to TAS. :)

In the end, we improved Genisto's version by 10 seconds. I was very happy with such a large improvement. Most of the time saved came from the "window of opportunity" at the boss scrolling. There is one frame where Scrooge can jump, and then 1 frame where he can jump again during the auto-scrolling. This gets him into the room much faster and farther than the auto-scrolling would have taken him.

I really enjoy watching this movie. The Duck Tales series is very rich for TAS possibilities. You have several movement options from which to consider optimal movement. They have a lot of level variety, and are fast paced.

Whomp Em

This game was unknown to me until I watched WalkerBoh's published movie. It seemed like a very interesting game for TASing. I thought I would try to improve it. I thought it would be an easy game to TAS, but it was actually one of the more difficult projects I have done. There is a lot of planning to do in terms of item drops and weapon use. There are numerous tricks and options so many things must be tested. Also the game is a pain to control during vertical sections (which are abundant). It was also a lot of fun to do (mostly), and I am really happy with the results. I managed to improve Walker's previous run by over 37 seconds. I consider this TAS a gem. Not many people know the game but it is an awesome game for TASing.

Marble Madness (GENS)

About a year after doing the NES version of this game (around October of 2006), I thought I would try out the genesis version. This was my first non-NES tas I ever made. I had to learn the new emulator (GENS) which was challenging. At the time, it was GENS9G which was crude by the emulator standards of the day. Nevertheless, I managed to finish this TAS and make it 2 seconds faster than the published movie at the time.

Nearly a year later SprintGod improved my movie by 10 seconds. He used a number of new shortcuts. This movie was my first official published obsoletion.

Duck Tales

I enjoyed the Duck Tales 2 TAS so much that I decided to work on this one as well. Again, Brushy began the TAS by doing the first 2 levels. I took it from there. The result was an 85 frame improvement over the at-the-time published movie by OmnipotentEntity. I small improvement was found in level 1 so this was cancelled. I restarted the movie myself. Along the way HHS found a huge glitch shortcut in the final level. This resulted in about a 40 second improvement!

I added this to my movie and the result was about a 52 second improvement. I think this is one of the more optimized movies of mine. The previous movie was very tightly optimized plus I did 2 version of my movie. So I feel just about every frame was scraped out of this one. Also, it was of my favorite to watch and TAS.


My motivation to do this TAS came mainly from the fact that Morimoto's at-the-time published movie was a) one of the oldest movies on the site b) a starred movie c) a famtasia movie (which I am trying to improve all of) and d) a movie that aims for entertainment yet I felt could be done much better.

I began this project in November of 2005. I thought this would be an interesting game to do because it would allow me to do something more artistic and unique compared to my other movies at the time.

Due to some desync issues I put this project away until late 2006. At this time, I worked with Miau with a special tool to help me realize my ideas for this game. The tool is a "Macro editor" which allows me to preprogram button presses into varies macros for easily implementation into the TAS. Miau added this to a modified FCEU. I could, for instance, set up a macro that moved the ship in a small circle. I could then simply hit this macro over and over to set up a constant circular pattern. I could inject an extra "up" press in the macro to make a ship move in a circle but this circle gradually move up.

In this way I could take a larger object (such as a circle) rather than the ship itself, and move the object around the screen. By doing this I start to morph it into some sort of visual art rather than a ship moving about a game. I tried to transcend the game itself and make it about how these patterns move and evolve over time. The use of the options (orange discs that echo your movements), these patterns are even more apparent. For instance, a 1st time viewer will usually catch me spelling out my name "adelikat" in this method. (I also have a few other messages along the way ;P)

This is easily one of my favorite TASes that I have done. It shows how art can be made out of a simplistic and otherwise uninteresting game.

Teenage Mutant Ninja Turtles

I began this project because Baxter convinced me to do so. At the time there were some new tricks and glitches that could improve Walkerboh's movie significantly. I started the project in November of 2006. Many improvements were found and the result was my 1st version of this game. Right after submitting it, I found a few improvements and really felt the run could be improved a good bit. I quickly made a 14 second improved version 2 that was published only a few weeks later.

This version was a lot better, but unfortunately had a couple minor mistakes in it. About 6 months later I decided that I should fix it up. I expected only a few seconds of improvement but to my surprise I managed to improve version 3 by nearly 8 seconds! I feel really good about this version. Only a few frames here or there could be scraped off it.

Contra - pacifist

The contra 2-player run was a very frustrating experience. At one point I had to quit for awhile. I decided that improving the 1 player pacifist run during that time would be a good idea. It would help me understand more about the game and hopefully pay off in the eventual 2 player run. While I didn't much care for the pacifist concept, I thought it would be fun to make the TAS.

I submitted my 1st version in November of 2006. It was good but someone suggested that I should die rather than kill the tanks. The ultimate pacifism, to sacrifice one's life to not harm another :) I decided this was a good idea even though this would result in a longer movie. The result was version 2 which was eventually published. It was a 47 second improvement over krieg747's movie.

This movie is good but unfortunately can be improved even further. I have begun a WIP that finishes level 3 1 second faster. Also, the level 5 boss fight can be improved about 40 frames or so from a rather obvious strategy change. I will finish this WIP one day, but it isn't high on my priority list.

Contra - 2 player

This was my first two player run I ever attempted (though Double Dragon 2 was published before this one, I had been working on it long before). This turned out to be one of the most difficult TASes I have ever done. Optimizing the spread gun shots was extremely tedious, especially when I have two players + lag to consider. The results were well worth it. This turned out to be one of my favorite my TASes to watch.


I did this project because I heard about a console speed runner who beat the published TAS. I had no real interest in Jaws, except that I felt it was the weakest movie on TASvideos. I just wanted to take a weak movie and improve upon it. Making the TAS turned out to be quite difficult and not fun at all. My published movie is likely improvable but I have no intentions of improving it.

Super Punch-out!!

The punch-out!! series is my favorite of all time. It is a shame that there is only 2 of them :( They are also the only games I have seriously competed in real time.

I had planned to do this movie right after my first published Mike Tyson's Punch-out!! movie but held off for about a year due to burn out with punch-out games. I began the project

Double Dragon 2

I had the idea to do this movie for about a year before I actually started it. My idea was to use "Mode B" where the two players could damage each other. I could then use this to flying knee a player across the screen to save time. The idea worked, though not in as many places as I thought. I found many other improvements along the way. I know this movie is improvable in a few places (though just by a few frames) but I am very happy with how it came out. Beat em ups are problematic, but this one uses unorthodox shortcuts by having the players beat the crap out of each other. This helps keep it action packed and increases the entertainment value a lot (for me at least).

This was my first completed 2 player movie. Optimizing 2 players is tremendously more tedious and time consuming than just 1 player.

Super C

I worked on this project with Baxter. Originally I planned to do it because I thought I could use the spread gun optimization techniques of the original contra here. Upon tasing the game, we realized that the spread gun is a sub-optimal weapon for this game. This was a very difficult game to TAS but very rewarding. We found many time saving techniques all of which look phenomenal. It was one of my favorite TASes of mine to watch :)

I also think we coordinated 2 players well in this movie. Most of the time, each player has an individual role to play in the goal of aiming for fastest possible time. Also this allowed death shortcuts. As a result Level 5 "Massacre Mountain" lives up to its name. This is by far my favorite level to watch.

Working with Baxter was also a great experience. He is an extremely thorough TASer and creative with coming up with new trick. He kept me on my toes and forced me to optimize and work harder than I might have otherwise.

Operation C

After Contra 2-player, Contra pacifist, and Super C, I thought I would use my Contra knowledge to improve this version. It is the gameboy version of Contra that mixes elements of contra and super C. I found that the homing gun has potential to save time. Like the spread gun of contra, it is a real pain to optimize but the results are rewarding. It makes for a nice TAS, especially due to the nice glitches in the game. But I don't find it nearly as entertaining as the other contra movies.


Prior to this TAS, I had never played a game using the Mupen emulator. The only N64 game I played before was Mario Kart 64, and only for a few hours total in my life. I had never heard of this game before, let alone played it. Meepers suggested this game as a possible TASable N64 game to look into. This was my greatest TAS challenge since I was unfamiliar with the platform, emulator, and game. Even more challenging is that few know of the game and so there is little information about the game. TASing it was truly exploring new territory.

I think this game makes for a good TAS. It has potential for many unorthodox shortcuts, and a few glitches. The character has a rich variety of movements and power-ups leading to many possibilities to consider when searching for an optimal route. Therefore a TAS of it is surprising since you won't know what to expect. These factors also mean that it differs greatly from a regular console playthrough.

Once I got the hang of TASing the game, it became a very rewarding experience. I first made a v1 that was totally unoptimized and just was an attempt at a "proof of concept" to show the potential of the game. Then I made a v2 that was almost 20 minutes faster, which is the previously published run. This run is not very optimized either though - at least not by the standards of my other movies. Part of this is that I didn't know if the game would be well received. The other reason is that N64 TAS tools were rather crude (as of the making of this movie).

With the help of MHS (memory hacking software) I was able to use some ram address values for an improved movie (version 3). I found numerous optimizations, shortcuts, and a rather large glitch shortcut in 3-4. The result is a 4 minute improvement that was published in September of 2007. This is a MUCH better TAS and represents the game well. It is also one of the 1st published N64 movies that used ram addresses to aid in optimizing a TAS.


This is one of the few Sega Genesis games that I played as a child. I though the published TAS of the game at the time was greatly improvable so I decided to improve it. This turned out to be a fun game to TAS and I think my improved version is much more polished looking. It is pretty optimized and, while not the most entertaining game, is pretty fun to watch.

Holy Diver

Originally Vatchern set out to improve his movie on his own. Near the end of level 3, he asked if I was interested in working with him. The game seemed very interesting for TAS due to its great difficulty and complexity in terms of optimization. At that point he was up about 250 frames. I had no idea we would end up improving it so much!

This was the largest margin on obsoletion of a published movie that I ever participated in (1 min 36 sec). I am quite happy with the results though I feel it could probably be improved more.

It wasn't until the last level that we both started competing with each other heavily to improve each other's attempts. This kind of competition brought out the best in us and the TAS shows it. But it makes me wonder how improvable the rest of the levels are since we just traded levels mostly on those.

Midnight Resistance

Shakespeare submitted a run of this game around 2005. I remember watching it and thinking it was a good game for TASing but not quite optimal enough. I forgot about the game and sometime in the summer of 2007 I was digging through old submission and ran across it. It is a contra-like game, so it is definitely a game for me to TAS ;P It took about two weeks to finish my first version (the currently published movie). I think it came out well, as I know of only 1 small improvement. It isn't my favorite of my TASes to watch, but I do enjoy watching it. Also, I really enjoy the music of this game :)

Super Mario Bros. 2 USA "Princess Only"

I started this project because I saw a few improvements to Laughing Gas's movie and also because I am very interested and love TASes of this game. This also makes a since way to "learn the ropes" of the game in preparation for an eventual improvement of the warps and warpless movies.

Super Mario Bros. 2 USA "Warp & Warpless"

I worked with Cardboard for these two movies. We worked on them at the same time since almost all of the warps run is identical to the corresponding parts in the warpless. These projects were by far the most frustrating and tedious projects I have ever done. The results were worth it though. These two movies are highly optimized and entertaining to watch. The warpless movie unfortunately has a missed trick in 2-3 that could save precisely 30 frames. I don't intend to redo for such a small improvement though. I would be suprised at any other frames were shaved in either movie.

River City Ransom "All aspects"

I worked with JXQ on this one. This was easily one of the most fun TASes I have ever done. I remember laughing a lot at new wips that JXQ would send me and creating some really goofy moments.

Blah Blah technical stuff goes here.

Blah Blah overdone tired speed/entertainment comments here.

Beavis & Butt-head

After 3 smb2 movies, I needed a fun project so I decided to improve the then published movie. I used the 2-player option which allowed me to control both characters simultaneously rather than 1 at a time. This saved a lot of time because I could divide tasks among the 2. The down side is I had to take a lot more damage and could use the trick of rapidly switching characters to avoid hits, but the time saved was far greater than amount lost. The result was a 47 second improvement.

Tecmo Superbowl

Super Mario Bros. 3 "Warps" and "Warpless"

Bubble Bobble

With Bubble Bobble, Cardboard, Tompa, and I did some live commentary: http://www.ustream.tv/recorded/4399545

This run took 2 years to make (mainly because it was nobody's main priority). We started it because of the famtasia challenge. It ended up being far more of a challenge than I anticipated originally. Then again, no 2-player TAS is ever trivial.

Combined RSS Feed
Adelikat/MoviesCommentary last edited by Ilari on 2012-08-14 04:51:52
Page info and history | Latest diff | List referrers | View Source