Roll Eyes
Lips Sealed
1 page
List results:
Search options:
Use \ before commas in usernames
Game Page: Doesn't exist yet

Chess Titans () (pc) [Any %] [Single Segment]

Decision: Accept

Congratulations to 'AntonioPeremin'!
Thread title:  
Run Information

Chess Titans () (pc) [Any %] [Single Segment]

Verification Files

Please refer to the Verification Guidelines before posting.

Please post your opinions about the run and be certain to conclude your post with a verdict (Accept/Reject). If you wish to remain anonymous, you can also send a pm with your reply to 'sdaverification' (please state clearly in that case which run you have verified). This is not a contest where the majority wins - Each verification will be judged on its content.
Which version of the game were you playing? (which OS) What difficulty setting is being used? 1?
OS windows 7; difficulty 1.
How likely is it to get those two moves out of the computer?
Around 1 in 20000. Smiley
Edit history:
LotBlind: 2018-12-13 09:05:37 am
What is that based on? Has somebody looked at the game's code or something? Just a guess?

Just in general, I'm kind of in the dark here. Is there any forums or anything for speedrunning this game? Any resources of any kind?
(user is banned)
Edit history:
AntonioPeremin: 2018-12-13 10:09:33 am
AntonioPeremin: 2018-12-13 09:15:39 am
There are statistics in the game and it took me almost 20000 tries.      = leaderboard, guide and forum.
A/V looks okay though a bit smutty despite supposedly being XQ.

Strategy: This is the second-fastest mate in chess as far as I'm aware. It's called the Fool's Mate Reversed. I quickly downloaded a Win 10 version of the Win 7 games suite and I tested what kinds of opening moves the computer will play (on difficulty 1): it never seems to move the f or g pawns as the first move as white, which is a requirement for the fastest mate, the Fool's Mate. So I'm assuming this has to be the quickest mate in the game, requiring the white player to make three moves total. It isn't clear if the second move has to be Qf3 or if Qe2 would also allow the AI to play ...g5. If we assume that Qe2 also works equally well, then making the first two moves with the mouse and switching to keypad mouse for the third one is faster than this surely. The game can be keyboard-controlled so you'd hit SPACE to select the last piece that was moved, then hit up and right on the keypad to move the cursor where it needs to go, then simply hit SPACE again. This would certainly save several tenths AFAICT, also for the fact that making the second move is faster that way. Even if you do have to play Qf3, making the last move (at least) with keypad mouse would still be faster.

In case you're wondering, keypad mouse is a standard Windows input method that has been used with other speedruns, such as Where's an Egg. In fact, if anything, it should probably be used more in any game with repetitive mouse-clicking tasks.

There's also this video showing other moves could be considered as well. I tried to get the response ...f5 to playing e4 just now. It didn't do that even once in the first 50+ attempts. Getting g4 played after h4 makes equally little sense so I guess it's just another super-freak occurrence. In that case, I don't really see the merit to doing it this way.

A few of the other attempts use Scholar's Mate, where white generally plays e4, Bc4, Qh5 and Qf7# against suitable moves by black. This doesn't seem to require as much precision in black's moves. Of course I don't know how easy it is to do in Chess Titans but looked like more than one run that does this exists. This is one more move but if it's a lot more common, this could in theory offset the time lost to that since you could grind out better execution. Since the last two moves are made on the queen you could, again, use keypad mouse to input the last one. Looks like most of the time, black will play the Sicilian and then go ...e6, ruining your plan.

Another idea is you could even try to guess how long the computer will think for and give the next input in anticipation of gaining back control. In fact, with keypad mouse, you could even spam the inputs although moving the cursor back where it was would be required after each attempt. I don't think this would be necessary for an accept in this case.

In conclusion, the feeling I'm left with is your approach is not bad but would be even better if you played Qe2, then keypad mouse moved to h5.

Timing: I don't know why the first move isn't timed. It's from start of control to loss of control. You have control the moment the new game has started. Let's put it this way: why is making the first move not equally important compared to every other move? How is the first move special? This wouldn't be the first time you've assumed whatever anyone did on SRC will pass SDA verification too, when it didn't. Literally all you have to do is bring it up in advance. However, no decision has been made about this yet and there is some potential for arguing that this is a good way to time it. Israeli says he's going to look at it over the weekend or so.

1st move: 40 frames (0.66s)
2nd move: 43 frames (0.72s)
3rd move: 47 frames (0.78s)
total: 2.17s

The thing is, the first move could be a lot faster. Here is an attempt from me:

1st move: 20 frames (0.33s)
2nd move: 37 frames (0.62s)
3rd move: 36 frames (0.60s)
total: 1.55s

While I'm generally a bit faster on every move, if we time the first move, that's twice as fast. I obviously didn't get the same RNG. The first move timing starts when the board resets just before I've made the first move. I.e. I first click on the pawn to get the highlight to show, then hit F2 to start a new game (clearing the highlight), and quickly input the first move. This gives a clear indication of when that particular match started. It's also different from simply clicking on the pawn three times, because when hitting F2, it seems the in-game selection cursor (the frames around the square) always goes to e4.

So because of the big RNG factor (I believe that it would have taken lots of attempts), this definitely isn't easy to reproduce or beat. The one time it happens, you could misclick, esp. if you have to play Qf3 instead of Qe2. In that sense this attempt does moderately well to give the correct inputs in a reasonable length of time. Verdict will follow after we've discussed timing. I think we can probably leave the computer thinking time untimed BTW, but Israeli will confirm that.
(user is banned)
Edit history:
AntonioPeremin: 2018-12-14 07:15:23 am
AntonioPeremin: 2018-12-14 05:33:17 am
AntonioPeremin: 2018-12-13 11:39:31 am
AntonioPeremin: 2018-12-13 11:11:53 am
On timing starts when the 1st piece is moved and the PC starts to think.
So Israeli decided that timing will start when the first piece has been selected. This means your first move only takes 16 frames (0.27s). That doesn't seem too long. Also CPU thinking time will be left out.

To summarize, I don't think this is the best way to input the moves/last move, possibly 2. Qe2 is better than Qf3 (but who knows if the computer ever plays ...g5 to that move) but the overall strategy seems solid. With the usage of mouse keys several tenths could be shaved from the time, and possibly this would even make the last move more consistent. Several tenths doesn't sound like a lot but it obviously is a lot percentage-wise. However, I don't know if the idea was brought up with you in the past and I can't expect every runner to know about this feature, so I won't hold it against you. Still, you should think about it in the future.

The A/V quality is a bit smutty. I think you could have considered playing and recording the game full-screen in case that's why this is.

Thanks for the positive feedback.  Smiley
Decision posted.