Wilmunder's Star Raiders II released

[SinisterSoft]

Has anyone converted the tokenised source on theses new atr images into text files?

[SinisterSoft]

This is a clip from something I've been working on for Android devices (phone, tablet, TV)...

I think that if you like Star Raiders then you may also like this:

If anyone want to have a play with the alpha version then please let me know here... Edited July 23, 2017 by SinisterSoft

[SinisterSoft]

The game is out now on Android - if you liked Star Raiders then you may also like this:

https://play.google.com/store/apps/details?id=com.sinistersoft.retrostar

[ClausB]

This deserves its own thread.

[doctorclu]

Has anyone converted the tokenised source on theses new atr images into text files?

 

So has the code been compiled?

[luckybuck]

Does anyone found the ranking list according to version I?

[+Stephen]

Demoed this at VCFMW this weekend. It was well received. A kid there (probably 6 or 7 years old) immediately said "Look - a Tie-Fighter, this looks like Star Wars".

 

Also got to demo it to Claus B. Sure would be great if this could get finalized.

[doctorclu]

Does anyone found the ranking list according to version I?

ranking.jpg

 

The way the current SR2 code seemed to work was you got your rank based on the star bases saved. I got my Captain Class 2 rank from saving three star bases.

[gladiator]

This is the first in a series of posts where I will attempt to answer LuckyBuck's question about the Rating Scale for Star Raiders II.

First, let me say that I think this is a really great game and I wish it was completed. Unfortunately, I claim that in its current state
it is not possible to obtain a rank of Star Commander and in the following posts, I will explain how I came to this conclusion.

 

The source files provided a great deal of information about the game including exactly how the rating is computed.

However, the posted files are in ATASCII saved on disks and cannot be read correctly in a text editor. I did a little

research into the differences between ATASCII and current text file formats and came up with a C program to

convert an ATR file into a readable TXT file. I am attaching my C source code here so that others can see what I did and

improve on this if you know more about ATASCII than me.

 

NOTE: I was not able to upload the C file, so I added a txt extention to it. I did not see anything in the guidelines prohibiting this
type of attachment although I may have missed something. It just appears at a C file is not a recognized file format.
I posted the C source code so that it is transparent on what I did

srctotxt.c.txt

[gladiator]

The result of running my program on the source files provided earlier are the following files that can be read from a text editor

D1_11_5.txt

D1_SRC_4_4_85.txt

D3_A_11_5.txt

D3_B_11_5.txt

D3_SRC_4_4_85.txt

Edited November 2, 2017 by gladiator

[gladiator]

Opening D1_SRC_4_4_85.txt, we see the method of computing the rating in a routine called FRANK.

The scale (see RATSCORE) appears to be:

 

0-30 Fledgling

31-50 Rookie

51-70 Ensign

71-90 Pilot

91-110 Ace

111-130 Lieutenant

131-150 Warrior

151-180 Captain

181-210 Commander

211-255 Star Commander

 

The class is something like finding the difference between the upper bound of a rating
and a score and dividing by 8 and adding 1.

 

So a score of 160 would be Captain and a difference of 180-160 = 20 when divided by 8

would give a quotient of 2 so this would be a class of 3.

Edited November 2, 2017 by gladiator

[gladiator]

Looking at the beginning of FRANK, we see the factors that contribute to the rating.

 

The first factor is a Base Score like in Star Raiders. The table BASE contains the

values and are

 

50, 52, 40, 60 ; STARSHIP DESTR

60, 60, 40, 50 ; OUT OF ENERGY

132, 120, 110, 130 ; MISSION COMPLT

 

The four columns represent the scores from each of the four difficulty levels.

So the maximum score that can be achieved here is 132 and this is in the novice level.

In commander level, the base score is 130 if the mission is completed.

 

The second factor is the number of KILLS. However, unlike Star Raiders where 6 points

are awarded for each kill, in this game the player only gets 1 point.

 

I suspect that there is a disk missing for the 4_4_85 set, but in D3_B_11_5, one can search

for KILLS and see that KILLS cannot increase beyond 99. This matches the behavior of the

current game where the number of kills is capped at 99.

 

So adding the base score and kills, we cannot get a value higher than 231. Everything which follows

will subtract from this score.

 

Curiously, however, in FRANK, there is logic that updates a value called ARG1HI which would allow

the score to be higher than 255.

 

I also found interesting that D1_SRC_4_4_85 that there is a commented out line of code

JSR TRENCH. If you read the design notes of the game, there is discussion of a "planet killer"

which is supposed to be the climax of the game. Perhaps that destruction of the planet killer

would count as an additional 50-100 points to the KILL score which would have pushed the

sum of BASE score and KILL to over 255. But as it stands now, I don't see how this score

can ever go higher than 231. We start subtracting from there.

 

In FRANK, we see that a third factor to the rating is the number of starbases destroyed.

Each starbase destroyed costs 20 points. By going through the logic of the code, we

see that one loses 20 points regardless of where it was destroyed by the player or the

enemy. This differs from Star Raiders where it was better to destroy a base yourself

than let it fall into enemy hands. The 20 points appears to be lost even for bases

destroyed in the "Galactic History" phase at the beginning of the game.

 

The final factor is time elapsed. Unlike Star Raiders which deducts this score in a linear

manner, it uses some table called ACCEL. I have attempted to figure out how this

table is constructed in GINIT. I don't have for the contents of this table on hand at

the moment, but its early entries are 1's and then increase fairly significantly later

in the table. I believe it is used to compute fuel consumed for a hyperwarp.

In any event, if a player is not able to complete the game fairly early, then he or

she will lose lots of points due to time consumed.

 

Note the fuel consumed is not a factor in this game.

 

This is why I got ratings of Fledgling for completing the Warrior and Commander games

no matter how hard I tried.

 

The best rating that I could achieve myself was Captain Class 2. I got it on a

Novice game where all starbases are far enough on the left where none get

destroyed at the beginning of the game. I tried to increase my score by letting

the base generate more enemies, but it takes 5 time units to generate a

squadron of 5 enemies. So that does not help and hurts even more over time once

the nonlinear part of the score computation is encountered.

 

So as it stands now, I don't think Star Commander is possible.

I'd be interested to hear from the author how FRANK was supposed to work

in the final game if it were allowed to be completed. Although this is a really

great game in its current state, we need to be reminded that the author

told us from the outset that this was an incomplete game. I think this is

one area where additional work may likely have been done prior to release.

[luckybuck]

WOW! Thank you so much gladiator, I really appreciate your help in this case.

 

By the way, in post #259 you wrote:

'The source files provided a great deal of information...'

Did I miss something? Do we have the source code from Aric? Would be a pleasure to load them to the AtariWiki. :-)))

 

Thank you so much.

[_The Doctor__]

no trench points = no chance... beaming people off and destroying your own base instead of letting the enemy destroy everything should hurt you less.. but hey if someone is actually going to complete the work on this game I'm sure they'll come up with something

[gladiator]

Perhaps I misunderstood. I thought that the five files above were the source files where the game was programmed directly in assembly language. I came to this conclusion because such meaningful variable names and comments were included in the files. Was there a higher level language used to program the game where it somehow compiles into the assembly language with comments on a number of lines? There are also lines in the files of the file #include FILE. I concluded that the assembly language was the source code. Can someone correct me and tell me that higher level language was used to program Atari games?

[luckybuck]

Hi gladiator,

 

There is no need to correct you, the files are indeed in assembly language! :-))) And they are quite huge! So huge, to put the whole game in... :-)))

 

Please let ask you, the files are named: D1... D3... and with A and B, are there any D2-files? From what source are the files? The format looks a little bit different.

[gladiator]

My only source to produce the txt files above were the files in posting #222.
I don't know why there are no D2 files. Perhaps someone else has the answer to this question.

All of the #INCLUDE statements seems to refer to D1 and D3 files with the exception with

something from a D6.

[gladiator]

Here is the missing piece to the puzzle that I left out from my postings last night.

 

If you search through the file D3_A_11_5 for HYPER ACCEL TABLES, you will see the code which generates the table called ACCEL. It has 182 entries and is filled in at run time I believe. I wrote a C program attached below which I believe simulates what is happening here. People can check my work and make sure I didn't miss any carry flag side effects or something like that. Assuming that I have everything right, the final table looks like this:

 

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2
2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 4
4 4 4 4 4 4 4 5 5 5 5 5 5 5 6 6 6 6 6 7
7 7 7 7 8 8 8 8 9 9 9 10 10 10 10 11 11 11 12 12
13 13 13 14 14 15 15 16 16 17 17 18 18 19 19 20 21 21 22 23
23 24 25 26 26 27 28 29 30 31 32 33 34 35 36 37 38 39 41 42
43 45 46 47 49 50 52 54 55 57 59 61 62 64 66 69 71 73 75 77
80 82 85 88 90 93 96 99 102 105 109 112 116 119 123 127 131 135 139 143
148 0

 

In the rating computation, the time is determined by the computation SDAT3 * 100 + SDAT2 and stores the result in TEMP. The rating computation does a loop from 0 up to and including TEMP - 1 where Y is the loop variable. Observe that the value ACCEL[Y + 16] is subtracted from the rating score for every iteration through the loop. So the first 16 time units of the game 1 point is deducted, then 2 points per time unit, then 3 points, and so on. You can see how the points deducted for time spiral out of control very quickly which is why every "successful" Warrior or Commander game that I have played, I always end up with a rating of Fledgling. Maybe I need to get more experienced at the game, but I don't see how I can ever complete it at the harder levels before the larger time penalties go into effect.

 

accel.c.txt

[luckybuck]

Hi gladiator,

 

OMG, I am not getting old, I am still old. Now, I have total recall. It is nearly 2 years the 1st post was made, so I forgot...

 

Bob has uploaded the source files. Thank you again, Bob. It was on my agenda to investigate further at the Fujiama meeting in August this year. But I didn't found the time for. My 1st priority is to save everything still on magnetic layers, which is not yet digitized. A few weeks ago, the Netherlands, Belgium and Luxembourg got 100 % digitized thanks to atarimuseum.nl. Germany is very close to this, max 3, maybe just 1 program from that goal. GB is very, very close, too. When that is done, there will be time for me to further investigate. Dream would be here, to have a source code of the quality Alfred has published. Condensed, just a few files and can be executed on cross assemblers very fast. The then available source code in PD would be best, especially for further development. Therefore, sorry for now.

[Snowflake]

Okay, I am going to speak some blasphemy.

 

There really isn't that much difference between Star Raiders II and Last Starfighter.

 

Basically, the zylons invade, attack starbases, and attack planets. Your job is to save your cities and bomb zylon cities.

 

Yes, the AI in Last Starfighter is non-existent. The AI in SR2 isn't much better. The horrible reticle aiming is where the challenge is. Otherwise it's just swapping between forward and aft view.

 

Truthfully, if they would have done better with the Fly Fighters in LS, given us an aft view, and made planetary battles better, LS would have been a perfectly acceptable SR2. Certainly the bombing campaign in LS is better than SR2. I love the banking on planets. The whole thing is a little reminiscent of Rescue on Fractalas (Behind Jaggi Lines). But again, that frustrating reticle aiming makes me want to bash my monitor in.

[luckybuck]

Snowflake, quite easy, take source code, make a SR III for us, with machine learning & ai and you will be be satisfied. The copyright can take your name of course. For the ai, I suggest to do some bank switching, of course...

[doctorclu]

All true. Honestly SR2 has a fun experience with the targeting, flying along the planet's surface, etc etc.

 

In short, Last Starfighter is more polished, but I think SR2 if a bit more polished an refined (such as the controls, the scoring, etc) would be a better experience.

[Snowflake]

Snowflake, quite easy, take source code, make a SR III for us, with machine learning & ai and you will be be satisfied. The copyright can take your name of course. For the ai, I suggest to do some bank switching, of course...

 

What, because I can't paint means I can't have an opinion about a painting?

[Snowflake]

All true. Honestly SR2 has a fun experience with the targeting, flying along the planet's surface, etc etc.

 

In short, Last Starfighter is more polished, but I think SR2 if a bit more polished an refined (such as the controls, the scoring, etc) would be a better experience.

 

Agreed. As was mentioned by the author, it's an unfinished product. I think the planetary experience was a candidate for the most tweaking. Loved the black holes.

[luckybuck]

Sure, but you have to go back in time, Aric was under great stress, no time to do it the way, he wanted. Further, please take into account, that the game was never published, because it wasn't finished! Seeing it from today with all the knowledge and experience, sometimes becloudes an objective view on the performance the authors have made in the golden age.