Chess

[+Andrew Davie]

Just now, Keatah said:

The game froze right after I made this opening move. Unfortunately I didn't do a save-state or anything. But, next time.

 

 

Yes, I specifically listed this as one bug.

I also said "just restart' and it will probably work.

 

[Keatah]

So excited to try the new version I didn't read everything.

 

Can you make the preview markers a single color like red or yellow or green? And perhaps a single pixel. I still can't get a feel-good vibe on how they are fitting into the overall design.

[+Andrew Davie]

14 minutes ago, Keatah said:

So excited to try the new version I didn't read everything.

 

Can you make the preview markers a single color like red or yellow or green? And perhaps a single pixel. I still can't get a feel-good vibe on how they are fitting into the overall design.

 

I didn't mean for my reply to be so curt; apologies.

 

I can make them any of the colours of the 3 colour lines used to make the pieces.

Depending on which colour, they use either 1 or 2 or all 3 of those lines.

Pretty much all the variations have been shown in the past 50 versions or so.  The 1-line ones look sparse, like the chess squares. The 2-line ones look OK. The 3-line are white.  I have tried many variations, but will try and do some alternates for you soon.

 

See post #14 for (roughly) the colours available.

https://atariage.com/forums/topic/299157-chess/?do=findComment&comment=4403708

 

This post explains the colour mixing...

https://atariage.com/forums/topic/299157-chess/?do=findComment&comment=4410261

 

Edited May 19, 2020 by Andrew Davie

[Keatah]

Given the colour choices, I think a simple yellow square looks best. Both the colour and the square shape combined together contrast nicely against the board and the pieces. Like so. This from:chess20200421_Aerlan_3ply_8quiesce_rand15.bin

 

 

 

[+Andrew Davie]

3 minutes ago, Keatah said:

Given the colour choices, I think a simple yellow square looks best. Both the colour and the square shape combined together contrast nicely against the board and the pieces. Like so. This from:chess20200421_Aerlan_3ply_8quiesce_rand15.bin

 

 

 

 

 

 

chess3E+20200519_3PQ6_SQ.bin

[Keatah]

I think that works good.

 

[+Andrew Davie]

There are quite a few variations I've tried. Here's one I like with slightly different piece colours, too...

 

[Keatah]

That looks aesthetically pleasing. I would be happy with it. The alternating pattern provides nice contrast.

 

The wife suggested green, is that a color choice? I didn't see it in any of the pieces, but it is the whole square that you move around with the controller. Curious to see how it looks.

 

[+Andrew Davie]

1 minute ago, Keatah said:

That looks aesthetically pleasing. I would be happy with it. The alternating pattern provides nice contrast.

 

The wife suggested green, is that a color choice? I didn't see it in any of the pieces, but it is the whole square that you move around with the controller. Curious to see how it looks.

 

The "cursor" you move around is the solitary sprite on the screen.

The "available move" markers are drawn with playfield. They can only be one of the 8 colour variants shown earlier.

The colour variants change slightly depending on what 3 "primary" colours I choose for the 3 scanlines.

 

[Keatah]

Got it. Your last screen shot looks good.

[+Andrew Davie]

Found an en passant bug -- the move is (incorrectly) allowed more than one move after the pawn moves +2.  In other words, rather than ONLY being allowed the very next move, it is currently always allowed.  Whoops. Will fix.

[AW127]

8 hours ago, Andrew Davie said:

There are quite a few variations I've tried. Here's one I like with slightly different piece colours, too...

 

 

Interesting. These two new colors (red and yellow) are very easy to tell apart:

 

 

Could be interesting to turn the colors around for a test and use these two (the yellow and the red) for the two different set of chess-figures and one of the previous colors (pink or orange) for the squares of the possible moves of a figure.

 

Just to see how this would look, because i guess the own and the enemy chess-figures would be very good to differentiate in yellow and red, maybe better than with orange and pink. Would be worthy a test version.   

 

 

 

 

[+Andrew Davie]

1 hour ago, AW127 said:

 

Interesting. These two new colors (red and yellow) are very easy to tell apart:

 

 

Could be interesting to turn the colors around for a test and use these two (the yellow and the red) for the two different set of chess-figures and one of the previous colors (pink or orange) for the squares of the possible moves of a figure.

 

Just to see how this would look, because i guess the own and the enemy chess-figures would be very good to differentiate in yellow and red, maybe better than with orange and pink. Would be worthy a test version.   

 

 

I understand your idea, but it doesn't work. The "red" and "yellow" are just one scanline "thick".  The pieces are made of 2 scanlines ("black" side) and 3 scanlines ("white" side).  In fact there are not "two new colours".  These colours are/were already on the screen.  The "white" is actually the "red" + "yellow" + the "blue" of the squares.  The "orange" is actually the "red" + "yellow".  Those colours were blended by placing them on successive scanlines so you don't actually see them individually unless you look closely.

 

 

Now, as I said - "red" and "yellow" pieces don't work because they use 1 of the 3 scanlines in each pixel. The pieces become too low-definition vertically and much less pleasing to my eye, at least. Here's a quick example for a couple of pieces.  Not only are the pieces now much more "stripey", but there is no suitable colour available to use for the shading on the right-side.  It just doesn't work.

 

 

Edited May 20, 2020 by Andrew Davie

[NostAlgae37]

On ‎5‎/‎18‎/‎2020 at 6:37 PM, Karl G said:

Here's the 5-15 version. Any differences are subtle. 

 

 

 

Thanks, that's how I hoped it would look.      I think that the effect is very nice on a CRT, I hope that Andrew will decide to stick with it.

 

By the way, I am working on a title screen for this game and am pretty far along with it, so hopefully after I submit it to Andrew for his approval, he will happy to display it here for all interested parties to see. 

 

[AW127]

19 hours ago, Andrew Davie said:

 

I understand your idea, but it doesn't work. The "red" and "yellow" are just one scanline "thick".  The pieces are made of 2 scanlines ("black" side) and 3 scanlines ("white" side).  In fact there are not "two new colours".  These colours are/were already on the screen.  The "white" is actually the "red" + "yellow" + the "blue" of the squares.  The "orange" is actually the "red" + "yellow".  Those colours were blended by placing them on successive scanlines so you don't actually see them individually unless you look closely.

 

 

Now, as I said - "red" and "yellow" pieces don't work because they use 1 of the 3 scanlines in each pixel. The pieces become too low-definition vertically and much less pleasing to my eye, at least. Here's a quick example for a couple of pieces.  Not only are the pieces now much more "stripey", but there is no suitable colour available to use for the shading on the right-side.  It just doesn't work.

 

 

 

 

Aah okay, i see the problem. Thanks for the explanation.

[SoundGammon]

Got a Flashback vol. 2 Blast Asteroids with 20 built-in games. Anyway, it has Video Chess on it and I played it last night and the screen doesn't do the random colors when the computer is thinking! The board stay on the screen while the computer is thinking and just flashes for a second when the computer shows his move!  Also, Maze Craze no longer makes flashing colors while a new maze is being generated, just blinks and new maze! Wonder how they did that?

[Keatah]

Maybe the blank non-flashing screens are easier to do. Or they are the natural state, meaning they don't cost any extra programming.

 

But back in the day it would've been acceptable to have a flashing screen. It's a little more crude and in your face. And things like that, not wholly unlike videomusic, were ok. Today not as much, it looks weird, unfinished, uncouth and downright buggy. And releasing something that looks buggy is a no-no. Key word here being looks. A cultural thing you know.

 

So they took the flashing out.

 

[Thomas Jentzsch]

1 hour ago, SoundGammon said:

Got a Flashback vol. 2 Blast Asteroids with 20 built-in games. Anyway, it has Video Chess on it and I played it last night and the screen doesn't do the random colors when the computer is thinking! The board stay on the screen while the computer is thinking and just flashes for a second when the computer shows his move!  Also, Maze Craze no longer makes flashing colors while a new maze is being generated, just blinks and new maze! Wonder how they did that?

All games are adapted for the hardware and emulated. Since the hardware already does it, the ROM doesn't have to take care of creating the TV signal anymore. And it also doesn't have to race the beam, but the data is written into a graphics buffer. So when a game is "thinking" the screen is kept stable and the current buffer content is displayed.

Edited May 22, 2020 by Thomas Jentzsch

[Mr SQL]

Had a chance to play a quick Chess game against the latest version of the 3 ply from 5/19!

 

didn't get a chance to test en-passante this game

 

Knight let me chase it around a bit again in the opening.

 

Loved the red end game screen that shakes when you win with this version!

 

Colors are looking even better - the board is gorgeous with extra artifact colors showing up on the peices on my set that I wish I could photograph properly. 

 

The CPU is faster than before, going to play the 4x next...

 

[+batari]

On 5/19/2020 at 5:34 AM, Andrew Davie said:

 

chess3E+20200519_4PQ6.bin 28 kB · 9 downloads chess3E+20200519_3PQ6.bin 28 kB · 8 downloads

IF understand the 3E+ spec correctly, it should not be hard to add this scheme to Harmony Encore so more people can help test this.

 

Clearly you aren't using all of the ROM. How much RAM are you using in these binaries? If you are using less than 8k, then it could be made to work in the standard Harmony as well. More than 8k and it will be Encore-only.

[+Andrew Davie]

6 minutes ago, batari said:

IF understand the 3E+ spec correctly, it should not be hard to add this scheme to Harmony Encore so more people can help test this.

 

Clearly you aren't using all of the ROM. How much RAM are you using in these binaries? If you are using less than 8k, then it could be made to work in the standard Harmony as well. More than 8k and it will be Encore-only.

 

It varies.  The organisation of the program is such that a single 512 byte RAM bank is used for each ply. Another way of saying that - the less RAM, the less deep the search can be.  Typically a search is anywhere from 3-ply to 5-ply, with an added quiescence search on top of that from about 4-ply to 8-ply. In the "worst case" scenarios the depth is getting to 11 or 12 ply - which corresponds to 11 or 12K of RAM just for the searching.

 

Although I could fit it into 8K, I don't really think that is a great idea, as the strength of gameplay would be significantly compromised.

 

Edit: My error here.  The RAM occupies a 1K *bank* but that is for read/write access. So it's the number of ply * 512 bytes, which means that 12 ply would only be using 6K.  So I'll change my view on this - it should be possible to have it working on the Harmony.

 

 

Edited June 2, 2020 by Andrew Davie

[+batari]

11 minutes ago, Andrew Davie said:

 

It varies.  The organisation of the program is such that a single 512 byte RAM bank is used for each ply. Another way of saying that - the less RAM, the less deep the search can be.  Typically a search is anywhere from 3-ply to 5-ply, with an added quiescence search on top of that from about 4-ply to 8-ply. In the "worst case" scenarios the depth is getting to 11 or 12 ply - which corresponds to 11 or 12K of RAM just for the searching.

 

Although I could fit it into 8K, I don't really think that is a great idea, as the strength of gameplay would be significantly compromised.

 

Edit: My error here.  The RAM occupies a 1K *bank* but that is for read/write access. So it's the number of ply * 512 bytes, which means that 12 ply would only be using 6K.  So I'll change my view on this - it should be possible to have it working on the Harmony.

 

 

That's good news. Yeah, don't limit yourself.

 

But if it happens to fit in under 8k (more accurately, 7-7.5k is more realistic for Harmony), then I can make sure this game is supported on all versions of Harmony, not just Harmony Encore

[+Andrew Davie]

11 minutes ago, batari said:

That's good news. Yeah, don't limit yourself.

 

But if it happens to fit in under 8k (more accurately, 7-7.5k is more realistic for Harmony), then I can make sure this game is supported on all versions of Harmony, not just Harmony Encore

 

I've had a closer look.

The board display itself takes RAM, and that's one bank per row. 

There's 8 banks right there.

 

Technically, each row only needs 144 bytes for the bitmap/graphics, so it could be reorganised a bit to, say, 3 rows per bank.

But that makes the code more complex. But it could be done. Current version, though, one bank per row.

 

So, then there's the previously mentioned one bank per ply in the search.  A minimum search is 3-ply, quiesce 3... give or take. That's 6 banks.

Then there are a few extra banks for other stuff.

 

We're at 8 (rows) + 6 (search) + extras.

 

Very tight, and at this stage let's jsut say Harmony is possible but unlikely.

 

 

 

[+batari]

1 minute ago, Andrew Davie said:

 

I've had a closer look.

The board display itself takes RAM, and that's one bank per row. 

There's 8 banks right there.

 

Technically, each row only needs 144 bytes for the bitmap/graphics, so it could be reorganised a bit to, say, 3 rows per bank.

But that makes the code more complex. But it could be done. Current version, though, one bank per row.

 

So, then there's the previously mentioned one bank per ply in the search.  A minimum search is 3-ply, quiesce 3... give or take. That's 6 banks.

Then there are a few extra banks for other stuff.

 

We're at 8 (rows) + 6 (search) + extras.

 

Very tight, and at this stage let's jsut say Harmony is possible but unlikely.

In that case, for now I will just add support to Harmony Encore. When the game is done, the actual requirements will be better known and I'll see what i can do for Harmony support, if it is possible.

[+Karl G]

The idea of a pure 6502 Atari game using that much RAM boggles my mind. I'm wondering if this game uses more than any other homebrew so far?