AtariCommunity forums : Carbon Copy page 28

Quote:

Originally Posted by Rubin (Post 11369916)

That is correct. The time difference is not more than 0.01 second, except 0.04 or 0.05 second difference at "Special by bzracer".

I already tried a lot to manipulate dat files, BUT I failed completely.

I mean the time difference between high_times and visible on finish. This was also 0.02 or 0.03 on different tracks.

Please look above as I edited my last message.

Quote:

Originally Posted by OlegM (Post 11369928)

I mean the time difference between high_times and visible on finish. This was also 0.02 or 0.03 on different tracks.

Please look above as I edited my last message.

Conclusion:

a) After having finished a race, the final time of the race can differ from the indicated (and saved in high times file) time by between -0.02 and +0.05 second. And so the time has to be corrected.

b) After having found the final time in the replay or demo mode, it can differ on some computers by ± 0.01 second, depending on the computer. And we take the higher time for the Hall of Fame.

Quote:

Originally Posted by mattrocks_slee (Post 11369691)

Hey Guys,

This is mostly directed at Oleg M:

You should never mess with what's not yours! 1.11.10!

The corrected time (on my computer) is 1.11.11.

But I am already faster. :D

The funny thing is: Matt made his world record at 7:22 p.m.. The time of my new world record is 5.47 p.m., though I made it many hours later! The reason is: Matt is in Australia, while I live in the center of Europe, 9 hours behind Australia. Which means: When he made his world record in the evening, I was still having breakfast! And I had to lot of time to prepare.

Now, if we only look at date and time of both races, it seems that 1.11.11 has never been a world record, because my new world record has been earlier! ;)

So, this is very good news for the future: We beat Matt within 9 hours, then his new world records have never been world records. This is simply great! :p

Matt, you really don't need that records. We will be faster anyway.

We have a consensus with Michael: I have 3 or 4 world records, he has all other records. Nobody else is allowed.

I have checked several records on official tracks :

Quote:

Races check results :

Not OK for :

AM1.22.40 (0.01)

FE1.04.98 (0.01)

FHFM0.58.15 (0.01)

VE1.20.30 (0.01)

OK for :

AH1.09.61

VH0.58.25

EDIT:Previous page backuped

Re:

Quote:

Originally Posted by OlegM (Post 11370977)

Matt, you really don't need that records. We will be faster anyway.

We have a consensus with Michael: I have 3 or 4 world records, he has all other records. Nobody else is allowed.

Oleg did not tell that he stole my beautiful 1.10.46 world record at Alpine Medium v2. 1.10.26 is his new world record. Matt, you are now an awesome number 3. Hope you can live with it.

Quote:

Originally Posted by xinfe (Post 11371101)

I have checked several records on official tracks

Thank you very much for testing!

Your race/s are not available.

Quote:

Originally Posted by mattrocks_slee (Post 11368616)

This .reg file is very annoying because I can't play certain tracks like Special by Housemarque and Mongul Slope. (to mk) Could you make an 'extra' Special Tracks #4 and just put duplicates where the blanks would be? Or maybe make .exe files that don't need the .reg file? Would help a lot.

I don't understand the part "put duplicates where the blanks would be".

However: I made an installer to be used at Linux. The installer is still a beta version, I have sent you the link.

mk

Attached is a nice picture from the Hall of Fame Special Tracks. You can see that the TOP 2 at Alpine Medium v2 are very close together. Then comes a huge lap, and then the last 2, Matt and Jake, are also close together:

http://ssb.bplaced.net/picture1.gif

Quote:

Originally Posted by Rubin (Post 11372477)

I don't understand the part "put duplicates where the blanks would be".mk

Maybe he means that there should be no spaces in file/folder names?

Quote:

Originally Posted by Rubin (Post 11372497)

You can see that the TOP 2 at Alpine Medium v2 are very closed together. Then comes a huge lap, and then the last 2, Matt and Jake, are also close together

:10sur10::mdr2:

Matt, your race on Mogul Slope v2 is good (no problem, we will beat it).

But this track is easy! Try usual Mogul Slope!

Quote:

Originally Posted by Rubin (Post 11372477)

I don't understand the part "put duplicates where the blanks would be".

Duplicates = Extra Tracks e.g. if you need three more tracks you could put Village Easy, Medium and Hard.

The game calculation misteries

New topic for discussion.

The Supreme Snowboarding is a strange game. Or the programmers are crazy.

I play the Mogul Slope v2 and have the race of Matt as a guide rider and the replay_0.dat also. Here is what I see:

http://angelam.ho.ua/Temp/CP1_time_d...high_times.jpg

But the high_times for this race says:

"Matt",34.1,66.54,109.89,"Day","2009/7/24",1,

The CP1 of Matt in high_times is 34.10. But when I play the (my version of) SSB Demo shows 34.09! :eek:

Where the SSB took 34.09?

No, I don't think Demo takes the time from the Replay because when we change high_times, the time is also changed. And there should be no checkpoints marks in the DAT file.

This is not the rounding error. It would be a calculation error or rounding error if the game took data from dat flle and rounded, for example. The high_times data is completely exact: 34,10 cannot be rounded to 34.09.

Can you (Xinfe, Matt...) check what your version of the Demo shows? Michael said that he seen this with finish time 5 years ago.

And maybe this also explains why the total time of the race is 0.01 less?

Quote:

Originally Posted by OlegM (Post 11372649)

Where the SSB took 34.09?

I haven't the faintest clue! It just simply isn't possible. There must be some bug in the demo causing damage. I've lost access to the demo for a while so I can't check. Let you know my answer as soon as I can.

This is a very confusing matter.

Re: mistery

Quote:

Originally Posted by mattrocks_slee (Post 11372738)

I haven't the faintest clue! It just simply isn't possible. There must be some bug in the demo causing damage. I've lost access to the demo for a while so I can't check. Let you know my answer as soon as I can.

Update: In FULL version I also see 34.09. Michael also sees 34.09 in Demo. If the time in high_times is changed to 34,11 the time on screen is also changed to 34,11. If changed to 34 the screen also changes to 34.00.

Next pair: 30.1 - 30.10

Next pair: 31.1 - 31.10

Next pair: 32.1 - 32.09

Next pair: 33.1 - 33.09

Next pair: 34.1 - 34.09

Next pair: 35.09 - 35.09

Next pair: 35.1 - 35.09

Next pair: 35.2 - 35.20

Next pair: 36.1 - 36.09

Next pair: 37.1 - 37.09

Next pair: 40.1 - 40.09

Next pair: 41.1 - 41.10

Next pair: 42.1 - 42.10

Next pair: 44.1 - 44.10

Next pair: 60.1 - 1:00:10

So the result should be 0,01 less if the time is = X.1 where X in between 32 and 40.

The result was the same on another track with another race file (Village Easy). It depends only on time in high_times.

(Thanks to mk for testing)

The part of the code should be there for showing the number, and it shows wrong number.

The only calculation there is: to break the number (34,1) to digital 34 (and extract minutes and seconds) and remaining 0,1 then multiply 0,1 by 100. So simple. But this calculation is buggy in SSB.

Interesting: the showed time can be less by 0.01 but not more. So the right time should be bigger time.

Well Done!

Well done for testing and finding.

On a totally different matter I have claimed #1 on Village Medium v2 in a wholesome THREE tries! 1.07.06!

Re:

Quote:

Originally Posted by mattrocks_slee (Post 11374368)

Village Medium v2 1.07.06!

The time shown in my demo version is 1.07.07, and so it is published. I do not like this route (though I was the one who made this route possible, by adding a checkpoint and a banner). Only because of this you can own this world record for a while, if Oleg does not want to have it. Or anyone else, for example Dom.

What I really like, is Dom's new personal best time at Alpine Hard: 1.12.43 coming from the Alpine Medium route. Except myself Dom is the only one who ever made it on this route. A video of the race is here:

http://www.youtube.com/watch?v=XoCg1FlV2d0&fmt=18

Dom was only 0.03 second slower than Oleg's 1.12.40 on the world record route three days ago. A video link of Oleg's race will follow as soon as possible.

About the time bugs. There are some strange effects:

1. If the checkpoint_1 time of the best race in the high times table is x.1, with x exactly between 32 and 40, you see x.09 in the game top right.

2. On some! computers some! races appear 0.01 second faster. For example: Matt's 1.22.41 world record at Alpine Medium appears on Xinfe's computer as 1.22.40.

Ad 1. checkpoint_1 time difference: Reason unknown, it must be a calculation error of the game. However, 0.01 time difference at a checkpoint does not really matter.

Ad 2. final time difference: Reason unknown. But now we have decided to always publish the worse of two alternative times. I never! had the better of two alternative times on my screen, but always the "worst" time, and so it is very easy for me: What I see on my computer, is what I publish in the official Hall of Fame.

Again the links:

Hall of Fame original tracks: www.supreme-snowboarding.tk / Hall of Fame

Hall of Fame Special tracks: ssb.bplaced.net

Time bug

I have created a high times file with exactly 100 races as follows:

{

"Michael",10,20,34.01,"Day","2009/5/15",1,

"Michael",10,20,34.02,"Day","2009/5/15",1,

"Michael",10,20,34.03,"Day","2009/5/15",1,

...

"Michael",10,20,34.98,"Day","2009/5/15",1,

"Michael",10,20,34.99,"Day","2009/5/15",1

"Michael",10,20,35.00,"Day","2009/5/15",1,

}

Then I started the game and the internal "Hall of Fame" to check this TOP 100. And I found this:

http://ssb.bplaced.net/picture2.jpg

Here is a full list of all times which are wrong in the internal Hall of Fame, first the correct time in the high times table, then the indicated wrong time in the internal Hall of Fame:

34.01 - 34.00

34.10 - 34.09

34.17 - 34.16

34.19 - 34.18

34.26 - 34.25

34.35 - 34.34

34.42 - 34.41

34.44 - 34.43

34.51 - 34.50

34.60 - 34.59

34.67 - 34.66

34.69 - 34.68

34.76 - 34.75

34.85 - 34.84

34.92 - 34.91

34.94 - 34.93

16 of 100 are wrong. Tell me if you can regognize any rule.

After that I changed the high table as follows:

{

"Michael",10,20,40.01,"Day","2009/5/15",1,

"Michael",10,20,40.02,"Day","2009/5/15",1,

...

"Michael",10,20,40.99,"Day","2009/5/15",1

"Michael",10,20,41,"Day","2009/5/15",1,

}

Result for 40.x instead of 34.x (x = 01 ... 00): Same errors as above.

Next:

Result for 99.x instead of 34.x (x = 01 ... 00): No errors!

Result for 60.x instead of 34.x (x = 01 ... 00): No errors!

Result for 41.x instead of 34.x (x = 01 ... 00): No errors!

Result for 31.x instead of 34.x (x = 01 ... 00): No errors!

Result for 32.x instead of 34.x (x = 01 ... 00): Same errors as above.

So. We are now talking about the final time. And the final time fault is exactly the same as the Check_Point_1 fault, and so probably exactly the same fault at Check_Point_2.

Here is the rule: For times between 32.00 and 40.99 16 percent of the indicated times are wrong. It is a heavy bug of the game.

Re: numbers bug

Just to add: Matt has also checked the 34.1 time and found 34.09. Only Xinfe didn't test it.

mk: very interesting. This should be the bug of extracting the parts of the time from the number (minutes, seconds and hundreds).

Do you see the difference between wrong numbers?

0.09, 0.07, 0.02, 0.07, again 0.09, 0.07, 0.02, 0.07 etc. Total 4 groups in 1 second.

I will try to identify what is wrong in algorithm, but maybe I will need more testing :alien:

Another question: does any one has ANY idea of the structure of the DAT file (replay)? The main questions: what is the length of the header, is the final time written in the header, where are the marks of the time stopping (or start and finish line marks), where are the X-Y-Z coordinates of the shadow. I have already know some details but not enought to extract time from the dat file.

Edit:Here is the result of testing with 94.xx :)

http://angelam.ho.ua/Temp/Wrong94.GIF

Wrong numbers:

94.02 - 1.34.01

94.03 - 1.34.02

94.06 - 1.34.05

94.07 - 1.34.06

94.09 - 1.34.08

94.10 - 1.34.09

...

94.99 - 1.34.98

Here the difference between wrong numbers is different: 0.01, 0.03, 0.01, 0.02 and repeat this. Total 12+ groups in 1 sec. And already 4 of 7 are wrong! 57%!

I tried to do a division of 34,1 by 60 to extract minutes. 34,1/60 = 0,56833333333. If cutted to 0,568 then 0,568*60 = 34,08 which is not what we seen. But 0,5683*60 = 34,098 (0,568333*60 = 34,09998 etc). If the game round any number down then it can show 34.09. Rounding is needed to show number (convert to text form). But rounding down would be very bad behaviour in such calculation.

But this theory is wrong if we do the same with 94.02 or so. 94,02 / 60 = 1.567. No extra digits.

Edit 2. Interesting combination found:

95.93 - 1.35.94 (+0.01!)

95.94 - 1.35.94

95.95 - 1.35.94

If so, we cannot say that we should always increase the time by 0.01 in case of difference!

Edit 3.Such errors were for time 45.xx, even in almost second pair.

And, Michael, you said that there were no errors for 41.xx but I see this:

http://angelam.ho.ua/Temp/Wrong41.GIF

To Xinfe

To Xinfe,

I want to make a special track. I've been looking at your tutorials but I can't see the streamed video, it just comes out blank. Also, where is this 'race' folder? I can't find what you mean, I only cause errors. One last thing, I can't download the editor edition of Supreme Snowboarding.

Thanks

Re: @ Oleg M

My tests show correct times for 31.x, 41.x, 45x. 60.x, 94.x and 95.x, e.g. 95.93 - 1.35.93! My tests show wrong times only for the rows 32.x to 40.x, as written before.

Shall we stop the testing? We have to accept that there are calculation errors within the game, and that the calculation can be different on different computers.

We should now decide:

a) The official CP 1 and CP2 time of a race is what is saved in the high times file, not what we see on the screen. We have no chance to check the CP (checkpoint) times of a single race in the Replay mode or in the demo mode. The CP times are NOT saved in the race, only in the high times table!

b) The official (final) time of a race is what we see on the screen in the Replay mode or in the demo mode. If it is different in the high times file, we have to change the high times file. If the shown time in the Replay is different from the time in the demo mode, we should take the time from the demo mode. If the shown times is different on different computers, we always publish the worst time. The final time of a race is saved in the race, but obviously not as absolute time, but as calculated time, and so there can be a time difference of 0.01 second on different computers. We have to accept that.

Re: time bugs

Quote:

Originally Posted by Rubin (Post 11376580)

We have to accept that there are calculation errors within the game, and that the calculation can be different on different computers.

...

The final time of a race is saved in the race, but obviously not as absolute time, but as calculated time, and so there can be a time difference of 0.01 second on different computers. We have to accept that.

Ah, you see, this is different on different computers? Any idea why?

The finish line mark IS in the DAT file. And the race time is absolutely correct: that is the digital number of records from start line (didn't find the mark yet) to finish line (already found), and divided by 100.

Don't know. Maybe in difficult cases (difference between positions in hall of fame is 0.01) we should decide by SSF ;)

Quote:

Originally Posted by OlegM (Post 11376602)

Ah, you see, this is different on different computers? Any idea why?

No.

Quote:

Originally Posted by OlegM (Post 11376602)

The finish line mark IS in the DAT file. And the race time is absolutely correct: that is the digital number of records from start line (didn't find the mark yet) to finish line (already found), and divided by 100.

The start line is defined in the dat file behind the player's name. And where is the Finish line?

Quote:

Originally Posted by OlegM (Post 11376602)

Maybe in difficult cases (difference between positions in hall of fame is 0.01) we should decide by SSF ;)

As already written before: If there are two different end times, I will always publish the worse of the two times, which is by chance what I see on my own screen.

@ Xinfe and Matt

Hi all,

Oleg M wrote a nice program which can extract start moment, checkpoint 1 time and final time from a race dat file. But before he can tell you more, please do us a favour and test four races in the demo version:

(1) Forest Easy 1.09.79 by Dom

(2) Village Medium 1.08.12 by Matt

(3) Village Medium 1.09.29 by Oleg M

(4) Mogul Slope v2 1.49.89 by Matt

What exactly do you, Matt and Xinfe, see on YOUR screens? Which times?

Thank you for your testing!

___

With Oleg's new program it is also possible to compare the Village Medium best times of myself (1.07.97) and Jupp (1.08.00).

a) I (mk, slow starter) passed the Start Line 4.21 seconds after the recording started and reached the Finish line 67.97 seconds later.

b) Jupp (fast starter) passed the Start Line 4.15 seconds after the recording started and reached the Finish line 68.00 seconds later.

In other words, Jupp has the worse time, but reached the Finish line 0.03 seconds earlier:

a) I (mk) reached the Finish line after 4.21 + 67.97 = 72.18 seconds.

b) Jupp reached the Finish line after 4.15 + 68 = 72.15 seconds.

This is also what you see when you play both races together in the demo mode of the Demo version.

However: I win! :10sur10:

My Special Track

Quote:

Originally Posted by Rubin (Post 11377889)

Hi all,

What exactly do you, Matt and Xinfe, see on YOUR screens? Which times?

Village Medium: 1.08.12 Matt

Village Medium: 1.09.29 Oleg M

Mogul Slope v2: 1.49.89 Matt

Forest Easy is unknown for the moment.

Two questions for my special track:

How do I design my map?

How do I record an AI Path or do I just drag the blue squares?

I'm creating a small tour of my track, then I'll send it to Michael.

Quote:

Originally Posted by Rubin (Post 11377889)

a) I (mk) reached the Finish line after 4.21 + 67.97 = 72.18 seconds.

b) Jupp reached the Finish line after 4.15 + 68 = 72.15 seconds.

However: I win! :10sur10:

Of course. This means you should do everything possible to start later. As in the cycle track sport :) :up:

Quote:

Originally Posted by mattrocks_slee (Post 11378979)

Village Medium: 1.08.12 Matt

Village Medium: 1.09.29 Oleg M

Mogul Slope v2: 1.49.89 Matt

Forest Easy is unknown for the moment

Is this what you see on your screen when you look demo replay of that races?

BTW, the CP2 info is not included in the DAT file. So it's not possible to make a tool to extract high_times line from DAT file.

But if you want the details: I wrote a program that takes the information from the DAT file and show it on screen. The DAT file contains some numbers which allow to show some parameters: start and finish time, checkpoint 1 time and full time of the recording. Here is the example of the program's output for "AM1.22.41 Matt.dat" (it also takes the name of the player from file):

Quote:

Results for the race of Matt:

Total time of recording: 89.95

Start line is crossed at: 4.52

CP1 time: 18.60

Finish time: 82.41

Lag after finish: 3.02

I have only DOS version now so I would upload the program when I will have Windows version also.

P.S. Whoever there, please check Dom's AH1.11.08 and AH1.12.43 races also.

Re:

Hi Matt, your test helps a lot. We will tell you more later.

About designing the landscape: It is still much too difficult for me, please check Xinfe's tutorial.

About designing a map = map.tga: I have my own, very difficult method. However, I created all maps for all Special tracks. If you cannot do it, I can.

About creating AI paths: see cheats page (aja + F8). So far I made all AI paths, except for 9 orginal tracks and Special by fjguarino.

Quote:

Originally Posted by OlegM (Post 11379002)

This means you should do everything possible to start later. As in the cycle track sport.

Not quite comparable, because in cycle sports the winner is the one who reaches the Finish line first. I remember when I was faster than the best time of the game at Village Medium for the first time. I reached the Finish without any doubt earlier than the best racer of the game. And then I was very surprised that my new best time was only the second best. Same here, but now I am the winner, though I reached the Finish line behind Jupp. However: What we do NOT need is some strange competition where you climb back on the track first, in order to pass the Start line with a higher speed.

Quote:

Originally Posted by OlegM (Post 11379002)

Is this what you see on your screen when you look demo replay of that races?

What else should it be. However, in this case is evident: Your best race at Village Medium is 1.09.29 (what we all see on our screen), and not what you see in your program (1.09.30).

Quote:

Originally Posted by OlegM (Post 11379002)

BTW, the CP2 info is not included in the DAT file.

Strange game.

Quote:

Originally Posted by Rubin (Post 11379013)

About creating AI paths: see cheats page (aja + F8). So far I made all AI paths, except for 9 orginal tracks and Special by fjguarino.

That is something I didn't know.

Anyway, I record it, what happens to it?

Quote:

Originally Posted by mattrocks_slee (Post 11379101)

That is something I didn't know.

Everyone should read my Tips & Tricks carefully! You can learn a lot.

Quote:

Originally Posted by mattrocks_slee (Post 11379101)

Anyway, I record it, what happens to it?

Don't understand. But I can do it for you. Or read fjguarino's earlier posts in this forum.

mk

Quote:

Originally Posted by Rubin (Post 11379033)

However: What we do NOT need is some strange competition where you climb back on the track first, in order to pass the Start line with a higher speed.

WOW! Great idea! :D

Cycle sports: if you start later you will be second in pair, so you will have less wind resistance and more strength for the final lap run. Of course, this is not for SSB but was to compare that you should start later.

Quote:

Originally Posted by Rubin (Post 11379033)

What else should it be.

I just want to know if Matt understood what you asked. Did he look the race as video or looks to the high_times or the hall of fame?

Of course I don't think Matt is so stupid as a blond but he could misunderstood you. This was the question for Matt.