Chess Challenger was a game originally programmed for the Colecovision in 1983 by Fidelity Electronics but was never commercially released, with only a few demo cartridges of the game ever being produced. In 2011, Collectorvision got hold of one of the rare demo carts, turned it into a finished game, and finally released it to the public.
Objectives
- Uses easiest difficulty
- Genre: Board
Difficulty
The easiest difficulty (Level 1) was chosen here. See the "Level 3" TAS I submitted alongside this one for a real challenge... maybe.
Routing
The goal is not necessarily to checkmate in the least number of moves but to deliver the checkmate that makes the computer "think" the least (of course, reducing the number of moves does help significantly). Shorter "thinking" times are achieved by keeping the position simple. Giving the computer few options and not attacking with a lot of pieces simultaneously is a good way to minimize thinking times.
The Computer's Weakness? IT HAS NO OPENING DATABASE!!
That's a real problem. No joke, if the computer determines it's safe to make an opening move, it could just make a random pawn move. And I mean random. g4 and f4, which are notoriously some of the worst pawn moves because they weaken king safety, are fair game as far as the computer is concerned. This isn't even unique to the Level 1 computer, as my Level 3 submission makes quite clear. We take full advantage of this by manipulating the computer to move h4, g4, f4—and the fate of the game is quickly sealed.
Playing as Black
I selected black for this TAS because the quickest game I TASed as white ended with the AI having 9 seconds on the clock, whereas my best game as black (this one) ended with the AI having only 1 second on the clock LOL. It turns out the computer is much more willing to make bad pawn moves as white than as black, essentially falling for some longer variation of the Fool's Mate.
RNG
Computer moves here are determined with a fairly wide degree of randomness compared to other board games on old consoles. Fortunately, RNG can be manipulated through the timing and type of both cursor movements and moves—meaning it is, like, really easy to manipulate. This is great because if the computer decides it's safe to make an opening move, there's a good chance we can get it to make a bad pawn move. RNG manipulation with the cursor is so versatile, we only sacrificed a single frame in the whole TAS for RNG manipulation.
Not only can the above actions manipulate what the computer moves, but also when, even if the resultant move is he same. This makes a little less sense to me how computer thinking times can be manipulated in this way but I guess it must call RNG during some point(s) in its search algorithm. In any cases, I tested this extensively to make sure the computer made each move in the least time possible.
Game Summary and Q/A
| Move | White | Black | White's Time | Commentary |
|---|
| 1 | h4 | e6 | 0:00 | White makes a horrible pawn move
|
|---|
| 2 | g4 | Bc5 | 0:00 | White makes a horrible pawn move
|
|---|
| 3 | f4 | Bxg1 | 0:01 | White makes a horrible pawn move
|
|---|
| 4 | Rxg1 | Qxh4+ | 0:01 | White is punished for being dumb
|
|---|
| 5 | Rg3 | Qxg3# | 0:01 | GG |
|---|
- Q: Why let the computer move h4 when allowing that just delays checkmate if the computer is so willing to play g4 and f4?
- A: The computer is not stupid enough to throw mate in 1, so the computer will not play g4 or f4 on the second move if it played g4 or f4 on move 1. But it will play those two moves if it has already played h4, which is why we must allow it. Now while the computer won't miss mate in 1, it can miss mate in 2, which is why we had to set up some sort of more complicated version of the Fool's Mate. By taking white's knight with our bishop, we set up a trap. The right move here for the computer is not to take back, as then the rook would stop defending the h pawn. The computer chooses incorrectly (probably because the engine is too material-focused), blundering mate in 2 as our queen can now come in and deliver the final blow.
- Q: Why not play bishop to e7 instead of to c5? Wouldn't that potentially lead to a quicker checkmate?
- A: White will not play f4, a crucially bad move (and is mate in 2), if we play Be7.
nymx: Claiming for judging.
nymx: As with your Intellivision, I see the same type of strategy here....exploiting the weakness of the game's ability. Because the game contains no database of "Book Openings", I see that it was quite clever to play as "Black" to force the computer to make bad moves for exploitation.
You have once again demonstrated what a "True" TAS should be like for 8-bit chess games. Using a strong chess AI, would have not have helped out...since the TASing is all about taking advantages of coding weaknesses.
Accepting to "Standard" for publication.
