Jump to content
IGNORED

Othello


devwebcl

Recommended Posts

I like the Atari 2600 implementation of Othello. I play it now and then, as well as playing 'Reversi' on other platforms throughout the years. I'm no expert by any means, but I do enjoy the game.

 

On the lowest level, it's easy to beat. Higher levels do provide some serious challenge for me.

 

Of the Atari-produced board games, this is my favorite. 3D Tic-Tac-Toe is also good. Checkers, chess, and backgammon are, IMO, sub-par.

Link to comment
Share on other sites

There are actually two different ROM versions of Othello. One in the text label cart and one in the picture label cart. In addition to some graphical differences, the AI in the two versions behaves quite differently. The picture label version appears to contain a bug in its decision making algorithm that severely restricts its decision trees sometime after about the 16th move has been made. However, when playing the picture label version, most players won't notice anything out of the ordinary because the variability of their own decisions ends up masking this bug. I only discovered the bug when I conducted a match-up of Atari Othello versus INTV Reversi. I keep meaning to write up a blog entry about it... :ponder:

 

Fundamentally, the AI in game has no real understanding of Othello strategy, especially mobility. Nor does it appear to project its positions several moves into the future. It appears to me that it's just evaluating the current position and selecting its next move according to a limited set of criteria, like will its next move enable it to occupy a square on the edge, or how many total pieces will it control after the next move.

 

Anyway, the AI is easily crushed by a human player who possesses even a limited understanding of the concept of mobility.

Link to comment
Share on other sites

There are actually two different ROM versions of Othello. One in the text label cart and one in the picture label cart. In addition to some graphical differences, the AI in the two versions behaves quite differently. The picture label version appears to contain a bug in its decision making algorithm that severely restricts its decision trees sometime after about the 16th move has been made. However, when playing the picture label version, most players won't notice anything out of the ordinary because the variability of their own decisions ends up masking this bug. I only discovered the bug when I conducted a match-up of Atari Othello versus INTV Reversi. I keep meaning to write up a blog entry about it... :ponder:

 

Fundamentally, the AI in game has no real understanding of Othello strategy, especially mobility. Nor does it appear to project its positions several moves into the future. It appears to me that it's just evaluating the current position and selecting its next move according to a limited set of criteria, like will its next move enable it to occupy a square on the edge, or how many total pieces will it control after the next move.

 

Anyway, the AI is easily crushed by a human player who possesses even a limited understanding of the concept of mobility.

 

Great info! I didn't that bug and the difference between versions.

 

I'd be great if you finally write something about the bug and the match between those two versions.

Link to comment
Share on other sites

I'd be great if you finally write something about the bug and the match between those two versions.

I really should do something with it, but it was one of those things that ended up leading me down the proverbial rabbit hole. It's difficult to make sense of it all, let alone make something presentable out of it, especially with everything else making demands on my time.

 

However, as far as the INTV vs. Atari aspect is concerned, I might as well share an overview of the results.

 

I ended up running 64(you see what I was saying about time?) contests between INTV Reversi and each of the Atari Othello ROM variants. Half the games were played with the Atari going first, half with the INTV going first. Both games were set to their highest skill levels.

 

The picture label version of Othello was crushed by INTV Reversi, 49-15. However, one quarter of the games played out identically, ending in the configuration below, which is strong evidence for a bug being present. Interesting to note that the score was the same as the overall outcome of the 64 games.

 

post-2851-126712790527_thumb.png

 

Conversely, the text label version of Othello more than held it's own against the INTV. In 64 games, it actually managed a slight 35-28 advantage, including one tie.

Link to comment
Share on other sites

  • 5 months later...

Maybe these two Othello versions have different algorithms to play, besides the bug found and the graphics.

 

Does Anyone know something about it ?

 

Devwebcl

I don't believe that the two versions were programmed around fundamentally different algorithms. Both versions behave essentially identical up until about 16th move has been made. After that point, the decision trees of the picture label version become extremely restricted for some reason. Nevertheless, despite the introduction of these constraints, the picture label version still appears to be using the same algorithm to evaluate its positions.

Link to comment
Share on other sites

Sounds like a good project for the ROM Comparison team to take up. :)

There are obvious playfield differences between the two versions, so a simple ROM comparison should reflect that. But in order to really understand what's going on with the "bug" in picture label version, I suspect that you'd need to disassemble both ROMs and directly compare their codes.

Link to comment
Share on other sites

That's really interesting... I've never actually played the pic version, but I played the text-label one tons back in the day (and watched my Mom play it even more... one of a handful of 2600 games she went through a period of being really obsessed with). I always thought the AI, while not that hard to beat, was reasonable on level 3 for something on the 2600. It's really fun to play against it on level 1... fairly often it will let you get it into a position where early in the game you can completely eliminate it.

Link to comment
Share on other sites

  • 4 years later...

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...