• entries
53
535
• views
67,223

# Knight Jumper with 32 puzzles

1,590 views

(new binary with bug fixes and new puzzle #1)

cc2_cd_zip.zip

Overall, I thought the comments from the 2006 1K compo were fair. The addition of new puzzles should significantly improve gameplay.

Note that all puzzles can be solved from any starting square. Also, the first puzzle is just a warm up; if you keep moving, you can't lose.

If you are new to Knight Jumper, the object of the puzzle is to guide your chess knight across the board touching each square exactly once. First move the knight to the starting square of your choice and then press the FIRE button to begin. From then on, just move the knight to the next square. You can FIRE to undo the most recent move, except at the end of the game. When the game is over you can reset with the FIRE button. Use SELECT and RESET for their obvious purposes.

It's nearly finished, with just a couple details to take care of. I still need to double and triple check that all the puzzles have solutions, and I want to make the scanlines stable during a FIRE reset.

Have Fun!

Note that all puzzles can be solved from any starting square.

Really? In puzzle #3, I would think the two top squares in the center two columns must be the start/end. In #5, the start/end are the two topmost squares. In general, if a puzzle has zero squares that have an odd number of neighbors, it can be started/ended anywhere. If it has exactly two such squares, it must start at one and end at the other. Otherwise, it is not solvable at all.

In general, if a puzzle has zero squares that have an odd number of neighbors, it can be started/ended anywhere. If it has exactly two such squares, it must start at one and end at the other. Otherwise, it is not solvable at all.

I've been looking over my graph theory book. I think what you are refering to is a trail rather than a path. The difference is that in a trail, no edge may be repeated, whereas in a path no vertex is repeated. In the knight's tour problem, the squares represent the vertices and the jumps are edges, so the problem is to find a path. The classic example of a trail is the Königsberg Bridge Problem, which turned out to be unsolvable.

It is indeed true that for a Eulerian trail (one that includes every edge and every vertex) to exist, exactly two vertices must have an odd number of neighbors.

I'm going to include Knight Jumper as a bonus game with Four-Play. Since I'll need to explain how to play in the manual, it doesn't make sense to make it an Easter egg. However, I can still keep the means of starting it a secret. What do you think? Would you rather be told how to launch a bonus game, or figure it out for yourself?

I'm going to include Knight Jumper as a bonus game with Four-Play.

Good idea.

Since I'll need to explain how to play in the manual, it doesn't make sense to make it an Easter egg. However, I can still keep the means of starting it a secret. What do you think? Would you rather be told how to launch a bonus game, or figure it out for yourself?

Depends on how hard you make it to find out. It shouldn't be too easy, then it makes no sense. And it shouldn't be too hard, so that everybody is able to find it. And they should then be able to play Knight Jumper without always having to "waste" a long time with Four-Play before.

This limits the possibilities quite a lot. So probably, not hiding it at all is the better idea then.

I'm going to include Knight Jumper as a bonus game with Four-Play.

Good idea.

Since I'll need to explain how to play in the manual, it doesn't make sense to make it an Easter egg. However, I can still keep the means of starting it a secret. What do you think? Would you rather be told how to launch a bonus game, or figure it out for yourself?

Depends on how hard you make it to find out. It shouldn't be too easy, then it makes no sense. And it shouldn't be too hard, so that everybody is able to find it. And they should then be able to play Knight Jumper without always having to "waste" a long time with Four-Play before.

This limits the possibilities quite a lot. So probably, not hiding it at all is the better idea then.

I dunno. Might be cool. Things like winning in a certain number of moves unlocks it, or a certain joystick+button combo from the title screen unlocks it. Or even just holding down certain switches and/or joystick buttons when you power-on unlocks it. I agree with TJ that it shouldn't be too easy and that it shouldn't take like 30 minutes of fiddling with Four-Play to unlock it. It should be tricky and quick.

If it's too hard, just put some clues in the manual.

I'm going to include Knight Jumper as a bonus game with Four-Play.

Good idea.

Thanks to both of you for the suggestions. I've decided as a compromise to divulge how to start it in the manual only. It'll be a simple procedure, once you know how. I'm also going to keep the existence of a bonus game low-key. I'll only mention it in this thread and in the manual, and I'll ask Albert not to use it as a selling point. After all, I don't want Four-Play to be classified as a multicart.

Good idea.

Didn't you consider some more bonus games earlier? I thought Marble Jumper 2 was supposed to be there as well?

Thanks to both of you for the suggestions. I've decided as a compromise to divulge how to start it in the manual only.

Boo!

Where's the challenge, then?

Boo!

Where's the challenge, then?

Find the bonus game before you get a copy of the manual. Also, I think the 32 puzzles will be challenging enough.

Didn't you consider some more bonus games earlier? I thought Marble Jumper 2 was supposed to be there as well?

No comment.

...yet

I'm posting the source to the Four-Play version of Knight Jumper in the hopes that someone can see what is wrong with the vertical hold on a TV. The picture is rolling slowly, like the background of a vertically scrolling game.

Try enabling VBLANK before VSYNC (or maybe even directly after the kernel).

Try enabling VBLANK before VSYNC (or maybe even directly after the kernel).

Thanks. I should have asked for help earlier. I've learned a couple things from this experience.

1) VBLANK is probably necessary to ensure that a game will work on every TV.

2) The top part of the frame is called vblank, but the VBLANK register should be set at the end of the picture, at the beginning of what is called the overscan.

Thanks. I should have asked for help earlier. I've learned a couple things from this experience.

I had the same problems with Splatform 2600. Learned my lesson there.

BTW: I suppose VSYNC is the really critical part.

VBLANK forces the TIA to output black, irrespective of what the playfield and sprites are set to. If the playfield or sprites are non-black during VSYNC, then the TV will probably roll. HMOVE can also glitch VSYNC.

Other than VSYNC, VBLANK isn't required (and even then it's just strongly recommended). Just be aware that when you turn off VBLANK that everything becomes active that cycle - it doesn't wait until the end of the line.

VBLANK forces the TIA to output black, irrespective of what the playfield and sprites are set to. If the playfield or sprites are non-black during VSYNC, then the TV will probably roll. HMOVE can also glitch VSYNC.

Other than VSYNC, VBLANK isn't required (and even then it's just strongly recommended). Just be aware that when you turn off VBLANK that everything becomes active that cycle - it doesn't wait until the end of the line.

My TV may unusual, but when I've tried programming games without setting VBLANK, I get a rolling problem.

VBLANK forces the TIA to output black, irrespective of what the playfield and sprites are set to. If the playfield or sprites are non-black during VSYNC, then the TV will probably roll. HMOVE can also glitch VSYNC.

Other than VSYNC, VBLANK isn't required (and even then it's just strongly recommended). Just be aware that when you turn off VBLANK that everything becomes active that cycle - it doesn't wait until the end of the line.

VBLANK is blacker than black (color 0). I'd recommend VBLANK even if everything is black.

×   Pasted as rich text.   Paste as plain text instead

Only 75 emoji are allowed.