Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
Q: How much input was done by each author?
A: I did the majority of the input, while Tompa did a few improvements, in particular the beginning of the reactor and a bit in the rancor pit. Tompa is listed first, though, because without him the route and most of the tricks in said route wouldn't exist whatsoever.
Two questions here:
1) It says Tompa did some input, and that he is "listed first". Is Tompa supposed to be included in the list of authors?
2) If Tompa did contribute to this run as the submission text suggest - shouldn't someone other than Tompa judge it?
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
Am I correct in understanding that this WAD is famous because of the nude geometry in the final level? The level design does not look to be anything special otherwise. In fact pretty bad in some places...
Akse wrote:
Fredrik wrote:
The Doom 2 map27 trick where the AV opens the door while being on the locked side is different.
What is it?
http://doom.wikia.com/wiki/Monsters_open_locked_doors
The check to open doors (or pick up items) is done before ejecting the object out of the wall. Therefore fast monsters can open thin doors from the wrong side. Same thing as player picking up keycards through thin walls.
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
How nice that you're doing this game! It's on my list of games I'd like to see. (Here.)
phi wrote:
I don't know the practices of this forum.
Do people tend to write a lot to describe their progression or is it preferable to work alone and post from time to time WIPs ?
That's really up to you. Some people give frequent updates, some prefer to work alone and in secrecy. I prefer to give updates now and then, because the feedback from other members can be very helpful.
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
Ready Steady Yeti wrote:
I haven't yet seen a linguistics topic here (correct me if I'm wrong), so I decided to make one now! (:
There's this old topic from 2006: Languages?
I've never studied linguistics per se, only a few different languages. Where are you from and which languages do you speak other than English and Danish?
If Wiktionary had a function to find words which needed translations in language X, I would be editing it a lot more. Now I only fix something when I happen to see it.
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
So, is anyone doing a new run currently? Me, TehSeven and thelegendarymudkip started running this game a while back, but we got stuck without recording too much.
I managed to figure out the RNG calculation, so I thought I'd share the info if anyone needs it.
The RNG is located in four bytes at
E4 E5 E6 E7
Every frame, the contents are shifted right one bit. For example
0000000 11111111 00000000 11111111
would become
1000000 01111111 10000000 01111111
The new leftmost bit of E4 is created by XORing the second rightmost bit of E4 and E5. Bolded above.
How is the RNG used? This has not been extensively tested, only for the birds in Pharao Man stage. For these, the frame when the egg drops start, the game reads E7 and uses it to determine the random action, in this case the number of eggs to drop. (1 egg 50%, 2 eggs 25%, 3 eggs 25%, in case you wondered.) It then writes to E6 and E7 in an unknown way. Feel free to figure it out. But it's dangerous to go alone, take this: http://nesdev.com/6502.txt
Can the player influence RNG? Yes. Both directly and indirectly.
- Indirectly, by waiting for one or more frames before scrolling an enemy into view.
- Directly, by shooting and sliding directly after each other. This locks E6 and E7 to a value dependent on which weapon you are using. For example, the mega buster locks to 0x05 0x05 and 0x85 0x05 (switching between them).
I have not seen E4 and E5 being written to. If this is indeed the case it means that all changes to the RNG are temporary, and get shifted out of the RNG in 16 frames.
Since the RNG is otherwise deterministic according to the formula above, it can be detected when the RNG was written to and didn't change as expected. Here's a lua script for FCEUX which does that, marking the addresses which didn't go to the expected value in red:
http://tasvideos.org/userfiles/info/24508776408376192
It's pretty crappy, but it gets the idea across. If you want to improve it you could add lag detection and load state handling. These give false positives at the moment.
Good luck, feel free to ask if something is unclear.
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
Akse wrote:
If you guys really want it. Entryway did a fantastic job! It's nearly perfect and there's practically no room for improvement. But since there is a new trick in e1m8 that has been discovered i can try it.
I would love to see that. E1 is the episode everyone knows, so it would probably be appreciated the most by the viewers. You could probably shave a few tics here and there even if it doesn't make a full second.
What's this about a new trick in E1M8, the boss level?
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
Truncated wrote:
Warp wrote:
I remember seeing a side-view 2D game for the Amiga where you controlled a bouncing ball. The only thing you could control was which direction the ball would rotate and how fast, and the ball would then bounce from the scenery. I do not remember what the goal was.
Surely this same idea has been used in countless games on many platforms, but it would be nice to see the original Amiga one.
Aw, damn. I have also played this game briefly, but the reason it stuck in my head was that a friend of mine made a clone of it for PC in Pascal back in 1996. Sadly I cannot remember the name either. If it means a lot to you I could try to get hold of this friend (via facebook or something) and ask him?
Four years later, I think I have the answer for you! I met the friend I talked about and he remembered exactly which game I was talking about.
The game is almost certainly Wizball!
https://en.wikipedia.org/wiki/Wizball
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
Alyosha wrote:
The support comes from command line instruction '-control', which as far as I understand it provides an input structure that the driver writes to and then is passed to the game.
The reason this allows strafe50+turns is that the game code won't call the routine that initiates the turn while in the strafe50 state. But the structure can be filled with the a turn value independently and sent to the game anyway, which fails to perform the same check on it. (I hope that is right)
So allowing strafe50+turns is allowing the call of the '-control' instruction, which does seem to mean that all available states would have to be allowed (going backward faster than 50 etc) otherwise it would be a case of saying "you can use '-control' but only to do strafe50+turns" which seems a bit arbitrary. However no runs do that so this is still hypothetical.
I guess my question here is when are command lines instructions such as that valid? Had the instruction simply been '-strafe50+turns' would that be allowed? Or since the game ordinarily doesn't support it would it be considered a cheat? A more obvious example would be '-run at speed 100' or some such thing.
Yes, this is mostly right.
The limitation to allow strafe50+turning but disallow speeds over 50 is somewhat arbitrary, but not completely:
- There is no check to remove turns while using strafe50. That this doesn't happen is due to a limitation in the control scheme used. (Specifically, pressing the Strafe On button is required to reach strafe speed 50, but that also removes the possibility to turn with the mouse.)
- There is however a check to reduce speeds in any direction when going above 50.
- There is also a warning message when the player exceeds speed 50 by using the -turbo parameter (but only in the forward direction due to lazy coding).
I guess it can be seen more as disallowing speeds over 50, rather than allowing strafe50+turning.
Warp wrote:
1) AFAIK that's not the case with strafe50+turning. You cannot build such a device. 2) The demo file contains data that cannot be physically inputted to the game without an unmodified system.
1) This is correct, strafe50+turning cannot be input via the mouse and keyboard ports.
2) This is incorrect, because mouse and keyboard is not the only way that Doom accepts input. Vanilla unmodified Doom also natively accepts input via the -control parameter.
Warp wrote:
Also, has it been actually confirmed that strafe50+turning can be supplied to Doom even with a modified joystick driver? Or is it just hypothetical?
It hasn't actually been done as far as I know, but it seems obvious from the Doom source code and the source code for the Doom External Control API. The API supplies a movement struct per tic, which is basically the same as the movie format, and we know that the movie format supports strafe50+turns.
If you (or anyone) can recompile the Borland C in the API, we could verify it by making some changes to the source. That would be pretty cool.
z1mb0bw4y wrote:
[snipping my post since the relevant junk got moved]
Will Quake or half-life style demos ever be acceptable for submission here, or is that likely not possible due to their nature?
No. Quake (and probably Half-life since it's based on the Quake engine, but I am not 100% sure) record demos in a way which allows tampering, which is unnoticeable afterwards. Therefore they can never be accepted. Quake-based games have to be recorded in an encapsulating PC emulator or similar.
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
feos wrote:
I mean, I added blank frames to the movie, but the ending just doesn't start. Just like in the author's encode.
Sorry, I forgot to reply to this in the middle of everything else.
The ending doesn't start because you are encoding it with PrBoom. PrBoom adds a status screen after the last level, which is not present in original Doom. In original Doom, the ending starts.
You can modify the movie to add an extra use command to start the ending in PrBoom, I think. Let me know if you want me to do it.
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
Hi.
MESHUGGAH wrote:
Reply to Truncated (too many quotes):
1. Congratulations on allowing players to use whatever arbitrary values that isn't replicable with real conditions. Yes, the game accepts GF90 too, can I submit it then as an "experimental exception"...?
Allowing turning and strafe50 doesn't mean allowing any input. I guess you could submit a movie which uses max turbo and see if people think it's entertaining. For other movies we will stick to max speed 50.
MESHUGGAH wrote:
2. Seriously, just proove me wrong, are you going to say that the judges' task is ask whatever TASer about his viewpoint and run a demo in a dos EXE? Can I finally get some response about what the hell are you doing?
Finally? Did you ask me before? Look, it's very unfortunate that I didn't see your thread in the General forum, and that you didn't post in any of the submission threads, but I don't think we need to be hostile.
To judge the movies, I did this:
- Searched for faster movies (for example on Compet-N and DSDA)
- Ran the movie in the original Doom on Dosbox to verify that it synched.
- Inspected the LMP file in LMPC. This detects movement >50 in any direction.
- Watched the movie to see if it seemed well played.
MESHUGGAH wrote:
Regarding the fantastic "reasons":
- No, it's not left+right, it has nothing to do with the possibilty to PRESS the buttons. It's about the game DOESN'T PROCESSES THE INPUT you give, meaning YOU CAN'T DO WHAT THE GAME DOESN'T DO. How the fuck do I need to emphasize this?
- No, the game doesn't says "Indigo: is turbo!" for quicker running, it's only activated if 2 conditions are met, which is absolutely obvious in the source code, I would like to ask you to read my General\DOOM rules thread and educate yourself.
Yes, I know that the game only actually checks forward movement, so it would be possible to run backwards/sideways faster than 50 and not get the turbo message. (None of the submitted movies do, though.) Elsewhere in the code (line 413 in g_game.c) it seems obvious to me that 50 is the intended max speed in all directions.
I also know that turning during strafe50 is not possible with mouse and keyboard. It mostly seems to be a limitation of the original control scheme, because
1) strafe50 is allowed and
2) turning while strafing is allowed.
The justification was that it is possible with an external driver, which Doom natively supports, so this is what a God player would use. That's what made me say that it was OK.
You could of course argue that the limitation was done intentionally, and turning during strafe50 should be banned. Or that allowing movement larger than 50 backwards and sideways by using a driver was intentional. I don't think so, so that's not how I judged, but I admit we don't know.
MESHUGGAH wrote:
edit: oh, and as I've already mentioned on a few places multiple times, none of these TASes are high quality (void glitch is possible on many places yet not used and totally suboptimal at e3m5 or e3m6 whatever. movement is totally sluggish (player waits for 20 tics and go forward if he magically gets into the teleport), do I need to fucking continue?
Where and when did you mention this multiple times? I can't find a single post by you in any of the Doom submissions, until now. That's where I looked for criticism of the submissions, obviously.
Like I said, I checked around if there was any faster movies, but couldn't find anything. The Doom community seemed to agree that these were the best there were. But if you know of some way to improve them, great for you! I look forward to your submission.
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
feos wrote:
Was not anyone going to publish this?
Noone isn't never going to not publish this! (You should do it.)
And yes, the ending sequence has to be added. Let me know if I should help out.
Warp wrote:
1) Does this use strafe50 and turning at the same time? 2) If yes, are we just going to ignore the question of whether that should be allowed?
3) I'd classify it in a similar category as abusing emulator bugs.
1) Yes.
2) It was discussed at length in our first Doom submission.
http://tasvideos.org/forum/viewtopic.php?t=16510
I judged that it would be allowed:
"I was asked to comment on Strafe50 while turning. Aqfaq's reasoning is sound, it can be compared to left+right. It is not possible to press normally, but it can be accomplished with a modified controller in both cases. The game accepts Strafe50 and turning as valid input when it receives it. This is not true for all input which can be sent, for example sending quicker running input than normally possible makes Doom go "Indigo: is turbo!" to warn viewers/other players that the player Indigo is cheating."
3) It can't be an emulator bug because it syncs on real hardware (or at least original Doom executable on Dosbox).
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
Nicos wrote:
why nobody posted it yet ???
Perhaps because "The copyright owner has not made this video available in your country"? Or maybe it's just in Sweden because it's available on state-owned television here. We're backwards like that.
nfq wrote:
Good stuff. Reminds me of Far Cry 3: Blood dragon.
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
Clumsydoomer wrote:
Doom speedrunners don't care about saving bits of a second because intermission screen time is more important for us. Local rules are different and we aren't used to them yet.
Yes, 2 shots seem to be faster because it takes more time to wait for the BFG traces to deal damage than to fire another shot. I'll ask Akse if it's worth fixing.
Yeah, I understand that intermission screen is historically more relevant, because it used to be the most exact you could get for Doom speedruns. When editing movies at a per frame/tic level though, it doesn't seem relevant any more...
According to Doom wikia: "The function M_ClearRandom resets both functions' indexes (P_Random and M_Random) to zero. It is called during initialization of each new level so that demos will be the same each time they are played, and so that multiplayer games are synchronised."
So if the RNG seeds are reset every new level, the final level input from the old TAS could be copy-pasted and still sync. Doesn't seem like too much trouble for someone who knows how.
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
I saw this on your youtube channel a while back and was wondering if you would submit it. Glad you did! I saw that you have done episode 4 as well, will you also be submitting that?
Two questions. When comparing to your previous TAS of episode 3, two levels are slower (though neither by a second):
- E3M2. 0:12.89 in new run, 0:12.74 in old. Probably because you pick up green armor, which is needed for rocket jump in E3M7?
- E3M8. 0:07.86 in new run, 0:07.66 in old. I don't understand why. You say that you finish the boss with one shot, and the old TAS used two - but two shots seem to be faster? You have the ammo for it in the new version as well.
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
Anty-Lemon wrote:
Speaking of the files, could someone replace the movie with this one? (trimmed to last input)
I can replace it if ClumsyDoomer thinks it's OK.
You didn't reply to the question if you've run into trouble or not... but it sounds like you haven't? Is it a busy schedule which keeps this from being published? Since it's taken a month, perhaps it is better to hand it over to someone else if this is the case.
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
Anty-Lemon wrote:
I don't know what to do with the game version, but I'll set it to v1.9 for now, since that's the Doom engine version used for Final Doom.
Hi Anty-Lemon. This has been claimed for three weeks now. Have you run into trouble?
If you need help editing the movie file to include the ending, feel free to ask me, and if there are other problems with capturing/encoding, you can probably ask feos who did the other Doom submission.
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
Thanks marzojr, that straightened things out. However, now I feel that there are two people better suited to put this together...
Here is a very rudimentary first version. Works for Gens, I don't know what the Bizhawk equivalent to register is.
http://tasvideos.org/userfiles/info/22528320965302561
What's immediately obvious is
- The bonus damage changes its whole part and its fractional part separately. Every time you enter a room, it picks a number to increase the whole part with, and a number to decrease the fractional part with.
- When an enemy is killed, its slot is taken by the gold bag object. The gold bag object uses the HP variable to decide when it should disappear.
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
I am trying to put together a lua script which shows relevant stuff about enemies HP and damage.
>You can see the enemies' HP by using a RAM viewer. Nigel's HP is stored as a 2-byte word at address 0xFF543E. Other enemies follow every 0x80 bytes (at addresses 0xFF54BE, 0xFF553E, 0xFF5BE, ...)
This checks out. How many enemies can the game handle at once? Are objects (chests, pots) in the same array?
>001174:xxxx
I don't get this. What RAM address is this? Neither 0xFF1174 nor 0xA01174 seem to be it.
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
Truncated wrote:
Anyway, even though the projectile is a lot bigger, the damage increase isn't that much from Green 6 to Green 7. A test I made yesterday showed that having both at Green 7 saves about 55 frames at the boss. I will have to test how much picking up the powerup costs. My guess is that it's similar or less. (This is mostly out of curiosity. I'm not remaking this stage.)
Tested now. It takes 37 frames to get the weapon powerup if you have the right RNG (which could probably be arranged in the previous stage). There could be some extra delays due to the characters being far apart during the stage, but nothing massive. So probably 15 frames or so to gain from changing this.
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
>that spread sheet is cool but does it count the sword modifier?
No. But you can divide the number of hits with the Sword modifier and round up. For example
8 hits / 1.75 (flame) = 4.57 -> 5
8 hits / 2.00 (thunder) = 4.00 -> 4
>nole dont take 200 hits
That was an error because of the irregular behavior of Nole's and Gola's health, which lordash explained above. It's fixed in the spread sheet now.
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
lordash wrote:
Most enemies will die immediately at 100HP, but due to the high defence score of the golems they will actually take more hits to kill than King Nole / Gola! Silver golems will die after around 11 hits at 100HP, whereas King Nole seems to take around 9 hits and Gola around 6. Even at 170HP (the highest you can normally get in the game), King Nole takes around 4 hits and Gola 3.
I updated the formulas with the correct rounding order.
But something is off here...
- You say that Silver golems die from 11 hits at 100 HP. According to the chart it's between 9-12, so that seems right.
- You say that King Nole dies from 9 hits at 100 HP. According to the chart it's between 16-23, which can't be right. I also checked at 84 HP, the TAS takes 15 hits, but the chart says 28-40.
My guess is that the listed Defense 31 and HP 200 for King Nole is incorrect. What should it be? Gola isn't listed at all, so his (her?) stats would also be nice.
You can override the random number generator by using these patch codes:
You read my mind. This is what I was going to ask next.
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
Hello lordash, and most welcome to TASVideos! I have to say, as far as first posts go, that one was pretty awesome. This is some very helpful enemy info, formulas and sword damages.
>I don't know Japaneese, but I worked out that the columns of numbers from left to right read the following: Attack Power, Defence Level, HP, Gold Drop.
I know some Japanese and can confirm that you are correct. (The last two are special item, and chance to drop special item.)
>BA (Base Attack) = Min(1,(HP-2*DS)*0.25)
>MA (Modified Attack) = BA + rnd*3/8*floor(BA)
How did you come by this info? Disassembling the ROM, testing and observing the RAM, something else?
Also, can BA and MA be fractional? (It looks like it.) Does the enemies keep track of fractional health, or is the whole damage floored to an integer before being applied?
>(If you want to always one-hit kill enemies, you will need to add one to the calculated HP level, as there is a 1/96 chance that there will be a roundoff error and your total attack will be ever-so-slightly less than the enemies' HP)
How can there be a 1/96 chance of roundoff error? Isn't the same calculation performed the same way every time you attack?
-----
EDIT: I made a chart of the number of hits it takes to kill an enemy given your HP. The top is the maximum number of hits, bottom is minimum (if you get 3/8 bonus damage on every strike).
https://docs.google.com/spreadsheets/d/1bvb_4qVh4YpNMb2CeMxfYUeg-LSfKsTATB-Xr4lbQto/edit?usp=sharing
It's a guess until I have answers to the questions above, in which stages of the calculation things are rounded down. It fits the numbers given for slimes and orcs at least. Names are according to the Japanese charts, I don't know the English names.
Also, this:
>BA (Base Attack) = Min(1,(HP-2*DS)*0.25)
should probably be
>BA (Base Attack) = Max(1,(HP-2*DS)*0.25)
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
Hi Archanfel! Some questions about this great run.
The movie has these tags:
-Uses death to save time
-Contains speed/entertainment tradeoffs
I didn't see any deaths to save time (doesn't that automatically lose the level?) and didn't find anything in the level comments about this or the entertainment trade-offs when scanning through it. Could you elaborate?
Also about this:
>Rerecord count 3 is correct. I wrote this movie in TAS Editor almost completely, used read-only savestates to get to the edited frame quickly.
Changing input after playing back the movie with other input, is basically re-recording. Doesn't TAS Editor add re-records when performing this action?
It's not terribly important, but if the re-record count is kept at 3, this will probably skew the Movie Statistics:
http://tasvideos.org/MovieStatistics.html
Editor, Experienced Forum User, Published Author, Reviewer, Experienced player
(969)
Joined: 4/17/2004
Posts: 3107
Location: Sweden
ars4326 wrote:
Now that I have the full run done to look back on, I do think a better money grinding route would be able to save at least a trip at the two points where I have to stay put for a while and grind (Verla and King Nole's Palace). It was certainly a challenging part of the run to figure out, during the making of.
One idea might be to defeat Fahl up to the 400 gold level. It takes around 1 minute 10 seconds in the encode, no idea how that compares to later money grinding.
I guess massive RNG cheating at the underground casino could also be tested.
