Posts for Baxter

Post subject: Adjustments to the player point/forum rank system
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
Due to future site changes regarding the publishing of movies that previously would have been rejected due to being a "bad game choice", as discussed in this thread, it may be necessary to change the current formula that is used to determine the number of player points. If you are already familiar with this player point formula that I am talking about, please skip ahead to the part below the === line. Otherwise, keep on reading. You may have noticed the number in parenthesis right after certain people's forum rank. These are the number of "player points" this person has. If you have a published TAS at this site, the rating of this TAS determines the number of player points you get for this TAS. The total number of player points determines your forum rank. Details about the forum rank system can be found here (this page will be edited if the system changes) or [url= http://tasvideos.org/forum/viewtopic.php?p=179250#179250]here[/url] (this page won't change and describes the system used as of this post). Here is a short description of the outlined formula to determine player points, outlined on those pages: numberofplayerpoints = (movierating^2.5) * (numberofauthors^-0.5) This formula has two aspects: 1) The number 2.5 makes sure that for instance one movie that was rated an 8 provides more player points than 2 movies rated a 4. 2) The number -0.5 makes sure that you will get fewer player points if the TAS has more authors. For a two-player TAS, you get roughtly 0.71 times the points you would have gotten had it been a single author TAS. =================================================== There are two reasons (but feel free to add others) why we should consider adjusting the current formula: (1) The number of ratings is not a factor in the current formula. (A movie that has an average rating of a 9 with 30 ratings gives the same number of player points as a movie with an average rating of a 9 with only 2 ratings.) (2) The addition of published "vault" movies. (Should a vault movie and a "regular" movie provide the same number of player points if they have the same rating?) Note that (1) and (2) may be related. This will be discussed later when looking at a possible answer to (2), but lets first look at (1): How should the number of ratings affect the player points you get for that TAS? There are a few things that may sound reasonable: #1 Since an average rating of a 9 with 30 ratings should be more valuable than a 9 with 2 ratings, one would think that more ratings should mean a higher number of player points given. Especially runs with very few ratings should be penalized, as the rating may not be very representative. #2 If a TAS with a certain number of ratings gives a representative rating, it should get about the same number of player points as it gets now. #3 To avoid deflation of player points due to TASes with few ratings getting less player points, TASes with more than the representative number of ratings should get slightly more player points than currently. #1, #2 and #3 can be taken into account by changing this part of the current formula: (movierating)^2.5 to: (movierating)^(A- (B(A-2.5)/X) ) "X" is the number of ratings that the movie received "B" is the number or ratings that is deemed to provide a representative rating of the TAS. If X=B, the formula reduces to (movierating)^2.5 which means you will get the same score as before. If X is lower than B, you will get less player points than before for this movie. A good value of B may be between 10 and 15. "A" is a number that influences what fraction of the player points you get for X<B and how many player points you get when your TAS has a large number of ratings. If X is significantly higher than B, the formula reduces to (movierating)^A. "A" should be higher than 2.5, but probably not higher than 2.6. Now lets look at problem (2), the addition of vault movies. As pointed out in a post by goofydylan8, it may be easier to create a ton of vault movies and get more player points than spending your time trying to perfect a star movie. Effort and high rated movies should be awarded, which is why a single movie with an 8 rating provides more player points than two movies with a 4 rating, but should this scale be adjusted to account for vault movies? Should a different change be made to account for this? One option would be to just multiply the player points that a vault movie would give by a factor lower than 1. Nice and easy solution which prevents the potantial problem that goofydylan8 raised. On the other hand, if a solution to (1) is implemented, where TASes with more ratings get more points, does this not already solve the problem? Vault TASes are bound to get fewer ratings and lower ratings. Both have a significant effect on the player points... do we really need another factor lowering the points for vault movies? You can also wonder why a vault movie that gets the same number of ratings resulting in the same average rating as a non-vault movie deserves less player points. This discussion will hopefully generate some posts with opinions on: 1. When taking the number of ratings into account when calculating the player points, what features should such a formula have? Are the ones (#1, #2 and #3) listed in this post good? Any objections or additions? 2. What do you think about the concept alternate formula that I posted (in bold the bold font), which takes the number of ratings into account? Is is good, do you know some tweaks that could improve it, or do you have a completely different suggestion? 3. How should the player point formula deal with vault movies? Will taking number of ratings into account be enough, or are additional measures required to prevent the issue that goofydylan8 raised? Also, if you have some suggestions/ideas for the player points/forum ranks system in general, those are of course also welcome!
Post subject: Re: Easter Egg
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
Derakon wrote:
The more optimized multi-runs tends to have similar problems.
The more optimized multi-runs? Either way, it is indeed somewhat pointless to watch them side by side, you will see no correlation, except that it may explain a certain wait at defenders of the crown because you will notice that pressing the button needed to continue at defenders of the crown would mean that lolo shoots a projectile where he shouldn't.
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
Must have miscounted. What influences these RNGs? Are they just timers? If you have to wait 65536 frames to get to the 65536th initial garbage state, then there is no use for the vaste majority of initial states for such a short TAS. Every frame that is spent manipulating must be gained again due to the better randomness.
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
Err, your discussion is nice and all, but you cannot just choose how many line clear animations you will have. Some initial state of garbage blocks will allow a solution with less line clear animations than other initial states. My TAS has a total of 9 line clear animations. HappyLee's 19 line solution, as shown in his picture, already had 10 line clear animations. Getting an initial state where as few line clear animations as possible were needed was obviously something I tried to get. I think I found a pretty good initial state that I was able to clear with only 9 animations. Of course, theoretically you can clear 25 lines with only 7 animations, but if there are no initial states that allow this, or if it takes too long to manipulate such an initial state, then it is not really relevant that this is theoretically possible.
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
If you make a tetris at the start, not clearing any garbage blocks, then you will have more than the minimum number of line-clear animations, so it is not as simple as you are suggesting. It has its ups and downs.
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
So this version will actually get done? :) Any estimate on that? (Are we talking about this year, or is that not reasonable?)
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
Well, I think that when B-mode was created by the programmer(s) they thought of it like "are you able to surive the time it takes you to make 25 lines if you start at a certain height?". So the goal for B-mode is to clear 25 lines. This would result in a movie extremely similar to the published A-mode movie, and would never be published. Then someone submitted a B-mode movie that got a bravo at the final line, which means that you need to build down all the garbage blocks that you are provided with from the beginning. I think this extra requirement of the bravo is what made this movie interesting and different enough to be published next to the A-mode TAS. But yeah, in the end, you do need 25 lines, or B-mode is not completed. Your question is an interesting one though. Would it be faster if you would start out by clearing a tetris (4 lines at the same time) at the top, and then start working down. This would mean that the pieces you use at the start with don't need to be dropped down as far. On the other hand, this may mean that more line-clear animations are needed to clear the 25 lines, which costs time. It's an interesting idea that's certainly worth testing, it could be faster!
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
Yes, zaphod77 is correct. That's exactly the calculation I did. I already had a suspicion that Happylee's solution would clear 19 lines, because 34 pieces * 4 + 54 garbage blocks = 190 = 19 lines. The picture confirms this. (I also had 54 garbage blocks + 4 * 49 pieces = 250 blocks = 25 lines). The number of garbage blocks is random, if I remember correctly, it will vary between like 50 and 55. It is indeed very possible to do it with 34 pieces if you get 114 garbage blocks, but it is very impossible to get that many garbage blocks. I tried a lot of random initial states, and 54 was the highest I could ever get. Getting 58 garbage blocks (meaning you would need 1 tetris piece less) is very unlikely I think, given what I tested. Even if it is possible (which I doubt), I think manipulating it would cost more time than dropping that single block, given that I tested random initial states for a lot of frames. Something that could save time would be finding an initial setup that allows you to have fewer line-clear animations, by having on average more lines cleared per line-clear. I did try to optimize this in my TAS though, and finding an initial state that allows this may cost more time to manipulate than the line-clear animation takes. Long story short, HappyLee's solution aims for the fastest possible bravo in B-mode. His solution clears 19 lines in total, which does not complete B-mode, and if the movie is stopped at that point, you would see pieces falling downwards in a straight line until game over. I do believe that actually clearing B-mode is a requirement for this TAS.
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
I am still puzzled by how you can do it with only 34 pieces. Besides getting a bravo, it is also required that you get this bravo while clearing the 25th line, and thereby clearing B-mode. Just erasing all the garbage tiles and getting a bravo, but not completing B-mode is not enough. Maybe I'm mistaken, but what you claim sounds... impossible to me. I would love to be proven wrong though, so yeah, I would be very interested in that picture if it explains something.
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
Yeah, you will have to manipulate the garbage blocks that you start out with. The goal of the B-mode TAS is to clear all blocks (a so-called "bravo") with the final piece you drop. It is possible to get an odd number of garbage pieces to start out with, in which case it is impossible to meet the requirements of the TAS. Also, for all we know the current version of the TAS is perfect, and it would be unfair if the time was improved based on the version of the game that it used. The current TAS was done on the (U) version, which is the version that the site generally encourages to be used (imagine an (E) version of SMB being accepted on the site if it were slightly faster than the (U) version).
HappyLee wrote:
I think Baxter's run is much improvable, and I've already got solutions to complete the very same level with 34 blocks instead of 49 blocks.
I don't really know what you mean by these "blocks" you are talking about. If you mean "pieces dropped", then I would be very interested how you did this while still meeting the "bravo" requirement.
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
jlun2 wrote:
I find a TAS that only does the extra levels kind of pointless. Why not get 100% using the GBA version? After all, it has extra levels.
Yeah, why not!? I think you forgot to add the word "just" to your sentence.
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
Any reason why this isn't a 2 player TAS? That would definintely make it (even?) more interesting. And why didn't you pick up the flamethrower in the first stage?
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
It may be a bit too much if we posted progress updates every time we completed a level. NxCy's signature is generally up to date on our current progress though. Anyway, we completed level 5-2 today, so we'll be working on 5-3 now. (20 more levels to go!)
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
Made a new map again after quite some time: Two Islands Let me know what you think :)
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
FCEUX is the new version of FCEU. FCEUX uses .fm2 files and FCEU .fcm. Since I made those movies at a time when FCEUX did not exist yet. As you mentioned, FCEUX is able to convert the .fcm files to .fm2, and while this works pretty well most of the time, there are instances where, due to small differences, the movie desyncs, which is what happened here. The easiest way to watch the movie without any problems is probably to just play it back in FCEU, which can be downloaded here: http://code.google.com/p/fceu/downloads/list (0.98.28 is the latest version, it should work fine with that. Also note that if you ever make a TAS, be sure to use FCEUX, as that's the currently preferred movie format). I don't know about the door when you enter Level S. I do know that at the original LoZ game, when you enter, leave and re-enter Level 1, the locked door in the first room is opened. Could be something like that, but I'm only guessing here... You probably already found it, but in case you missed it; this post contains links to maps of the overworld and the levels: http://tasvideos.org/forum/viewtopic.php?p=201683#201683
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
Hey avengah, and welcome to TASVideos :). I did not consider that possibility when I made the TAS, so I never tested it. Your idea would indeed save some time: about 10 seconds actually. Very nice find!
Post subject: World 4 complete + WIP
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
It's been a few months, so time for a progress update! We just completed world 4, and are starting with world 5. Some of these upcoming levels are quite long, but we'll try to keep working on it at a steady pace! We won't be posting too many more levels to avoid spoilers, but here is a short but sweet WIP of level 4-7: .smv Youtube
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
Funnel Up I changed the method that should be used to get to the exit. This part is probably the most challenging and interesting of the entire puzzle now. Definately check it out! Airbridge I made some changes that should make the level better. It's probably still a bit complex for the masses and I don't feel like this is the final version, but I would love to hear some opinions on it :). Laser Field (normal) A completely new map where you have to create a field of lasers to connect all 41 laser relays in order to open the door. In this version, you are provided with 10 reflection cubes, but it is possible to complete it by only using 7. If you're up to that challenge, then try Laser Field (difficult).
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
Airbridge My third map. Not sure how difficult it is. Hope you enjoy it; suggestions are always welcome :).
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
It should start at 0 every time you start the game... at least on a real NES, but I would guess on virtual console as well. The counter moves up 1 by killing an enemy. The picture only shows the item that will be dropped IF the enemy drops an item. Whether or not it will drop that item is random, as far as I can tell. This means that everytime you start the game from scratch, I think you should get bombs if you kill a blue octorock first (although it may not drop anything). Also note that killing lots (like 10) enemies in a row without being hit means you will get a 'good' item such as a bomb, a fairy, or a blue rupee. If you want to know exactly how it works, I would suggest looking at how the address behaves when you kill enemies and checking how it corresponds to the picture.
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
Funnel Up v2 I made a small change near the end. Transferring both yourself and the cube to the top at the end by funneling may be too frustrating, so I changed it a bit after hearing how Atma dealt with it. It's a little more puzzling and a little less frustrating this way I hope. Let me know what you think :). I'll take the v1 down if this version is liked better.
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
Funnel Up My second map. Use the tractor beam to maneuver around the room and watch out for flipped walls.
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
Cubic jump My first map. Try to get the cube to the other side of the water. If you want a real challenge: try doing it without using the light bridge.
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
Nice little TAS, voting yes.
Experienced Forum User, Published Author, Skilled player (1403)
Joined: 5/31/2004
Posts: 1821
I don't think it would. In fact, I'm pretty sure that it wouldn't... but that wasn't the point. The point is that no TAS would then show the different, entertaining non-BLJ gameplay that this TAS provides.