-
Content Count
27,895 -
Joined
-
Days Won
3
Posts posted by Thomas Jentzsch
-
-
Just watched the show. Tanya did a lot of VROOM! AI plays without noticing.
I guess I will make the computer players more obvious.
The computer boost flashing is just a result of the AI. It randomly presses when the cars are over a boost.
-
5 hours ago, insertclevernamehere said:I'm hoping these games aren't going to be super similar. On a surface level they do.
My thoughts exactly, but I strongly doubt that they would harm their reputation that way.
-
IMO the problem is that all eight Rhinos will meet in the middle, and that will cause hefty flicker.
-
1
-
-
Looking forward to your play testing. I suppose you can win the normal (yellow) variations, but I doubt you can do the same in hard mode. Else I have to tweak the AI.
-
1
-
-
Cool finding.
Reminds me at hidden the reserve ships graphics in Asteroids.
-
7 hours ago, SvOlli said:So what? Defeating most (if not all) types of encryption is easy.
All you have to do is to implement a trigger for bankswitching, which is not too complicated after dumping and disassembling the first bank. Only way I know to defeat this is using a Harmony/Melody board with code running in the ARM microcontroller.
Yup, and I suppose even that can be dumped.
The questions is what you would want to achieve.
- Prevent dumping in general? Tough if not impossible. Svolli will know for sure.
-
Tracking down the person who dumped and released the ROM? Possible if you watermark the ROMs. Removing the watermark without destroying the code will require an in depth analysis.
E.g. if you watermark just random 100 bits and sell 1000 copies, you can (most likely) still identify the pirate even if ~90% of the watermarks got removed. And finding 100 unused bits is easy, even in a 4K ROM. Just think about the mirrored addresses and unused TIA bits. I once did an experiment with one of my own 4K games and easily got about 2500 bits in 600 bytes.
-
15 minutes ago, Dionoid said:Wow, this game is progressing fast!! I just tried it out, and the gameplay is really good (= IMO simple but addictive).
This is a really nice casual party/mini game.
When playing as single player, it would be nice if there were "stages". E.g. for each race you would need to finish in the top-3 to go the the next stage, in which the A.I. cars are a bit harder harder to beat.
Anyway, it's a fun game with refreshing gameplay!
Thanks.
Currently I am focusing on the party game mode, because I plan to submit the game to the Revision Demo Party Game competition.
-
3
-
-
12 hours ago, alex_79 said:It turned out quite an addicting game, even when played alone against the AI.
Thanks. Actually I am surprised too how well the rather simple AI works. And since I currently have to play (and test!) alone all the time, this would be totally boring without AI.
Next planned change is high
scoredistance recording (for human players only). So that you can challenge yourself. -
2 minutes ago, CPUWIZ said:Re-arrange a couple routines in memory and your watermark is gone, optimize some code while you are at it.
Sure, if you dissemble the code completely, then removing watermarks becomes easy. But disassembling and finding e.g. all pointers is quite a lot of work and understanding. Definitely too much for the usual pirates.
-
2 hours ago, CPUWIZ said:A floating ROM can have it's serial number check etc. removed in no time.
Bits are bits.
If you know the bits to remove, then yes. But you can use bits randomly spread all over the ROM to watermark it (without breaking the code).
-
And now to version 0.8 (as always, see first post).
Major changes:
- the distance is displayed after the race
- split old easy mode into kids (white) and easy (blue) modes (I hope kid mode is now easy enough for kids)
-
2
-
16 hours ago, D Train said:that's 205 mph which is pretty reasonable (if not a little slow) for an open wheel racer on a closed track; things a slower on open tracks.
without trying to, you managed to nail it!
nice job!
Just assume the car is 10% longer than 4.2m and you are automatically 10% faster.
-
1
-
-
I did some fun math today. From the car width (14 pixel) I assumed that 1 pixel is ~30cm (-> car length 4.20m). Using this I checked the distance the computer reached in normal mode within 30 seconds (~2750m). And that resulted into an average speed of ~330km/h.
-
3
-
-
-
10 minutes ago, Al_Nafuur said:blue is easy? My 5yo still has problems getting better than 3rd last in this mode, she is very upset and demands a "Kids Mode"!
OK, next version will have an even easier mode.
-
1
-
-
14 minutes ago, alex_79 said:I only played a few rounds of v0.7 so far, but I like the new difficulty settings.
Normal and hard modes are challenging against the AI.
Thanks! I can sometimes win in normal mode, but I have a hard time not to become last in the hard mode.
14 minutes ago, alex_79 said:I tend to do worse with longer race length, as I start to lose concentration and doing mistakes and I cannot catch up.
Same here, I suppose that's normal.
14 minutes ago, alex_79 said:Also the "superboost" arrows are an handicap for me, at least against the AI, as it is way better at catching them than I am.
Yes, that's something I noticed when adding them. I had to make them longer (better for human players) and less effective, else the AI would profit way too much.
How is the frequency of the Superboosts? Currently there is a 1/16 chance, maybe that's too high?
-
Next version 0.7 add to first post. Changes:
- implemented three difficulties
- added optional SuperBoosts
- lowered braking frequencies
- more detailed instructions in first post
Size now increased to 4K.
-
3
-
I would name the project after what all variations have in common (e.g. multi-cart, for the same consoles, using the same CPU...). Or make up a completely unrelated name which could describe the unified character (e.g. "Trinity").
-
1
-
-
12 minutes ago, Al_Nafuur said:3rd last was here best try, but she has just come back to try again.
IMO that's pretty good for a 5yo.
-
1
-
-
6 minutes ago, Al_Nafuur said:v0.6 is too easy to win for me, the AI in v0.5 was much better.
The AI is the same, but the game has become easier for human players. I will adjust the AI when the other parameters are finalized.
QuoteBut for my daughter (5 years) v0.6 is still to hard (and frustrating) she gave up after 3 tries.
You take your 5yo to parties?
Was she able to become not last?
QuoteSo maybe you should use the difficulty switches to control the strength of the AI.
Might do. Each switch could control 4 cars. But maybe two levels are not enough to cover all needs.
-
5 minutes ago, Al_Nafuur said:...maybe we should rename it again, because Vroom! is not a bB demo.
Yup, 100% Assembler.
-
12 minutes ago, alex_79 said:I need a stronger AI, anyway, as I always win now.
Yes, the latest changes made the game easier. I will wait for more feedback before I tweak the AI.
12 minutes ago, alex_79 said:I'm using the green car, as I find more convenient to keep a joystick in the left port and use paddles on the right one. Don't know if that makes any difference for the AI skills.
You are eliminating the 4th strongest computer player that way, so that shouldn't make much of a difference for you.
-
1
-
-
Just now, alex_79 said:True, but I guess that's the same for the braking sound. Difficult to say which car refers to with 8 players.
The difference is, that braking should happen much less frequently (and computer never brakes). So it should stick out better.
-
New version 0.6 in first post. Changes:
- the average boost length increased (less short ones)
- braking got 20% less strong
- changed font (still experimenting here, looking for pop art style)
- button presses change boost arrow color during start phase
As always, feedback welcome.
-
2

VROOM!
in Homebrew Discussion
Posted · Edited by Thomas Jentzsch
AI worked well during the show.
Those who watched the show will know this already, for the (few) others: I am planning to join the Game Competition at the Revision 2021 Demo Party (happening Eastern, live on Twitch). I don't expect any fame and glory, but since VROOM! is designed as a party game, I thought it would fit well into that demo party. Of course it would work much better, if this year's party wouldn't have to be completely virtual, but maybe people can imagine how it would be.
Following their rules, I must not release my work to public before. Therefore, until then there will be no more public releases (sorry). Instead I will continue development and testing with only very few people in private. Of course this will change back to normal immediately after the show.
Nevertheless any feed back from everyone else is of course still more than welcome.