2600 Words (Wordle clone)

[Thomas Jentzsch]

10 hours ago, Nathan Strum said:

Sometimes... you just can't win.

 

Probably you were just too greedy.  

[+Nathan Strum]

Shouldn't the first 'E' in 'WEAVE' and the second 'E' in 'EXCEL' be orange, since neither are in a correct position?

 

[CardboardBox]

 

9 minutes ago, Nathan Strum said:

Shouldn't the first 'E' in 'WEAVE' and the second 'E' in 'EXCEL' be orange, since neither are in a correct position?

 

It's working the way Wordle works. Since there's only one E in CYCLE, only one E gets colored.

[351coug]

if you saw both colored that would imply there was another one which there isn't. that would lead you to chase incorrect words and you only get so many correctly spelled word guesses. your not allowed to just fill the lines with random letters to exclude stuff. 

[+Karl G]

10 hours ago, Nathan Strum said:

Shouldn't the first 'E' in 'WEAVE' and the second 'E' in 'EXCEL' be orange, since neither are in a correct position?

As others have mentioned, this is how the reference game works. It took me a bit to wrap my head around when I was developing it, but the easiest way of thinking of it is that each letter in the solution is only matched once.

[+Nathan Strum]

v. 9 is still not displaying stats correctly:

 

The number of guesses doesn't add up to the number of wins:

 

I'm no longer sure how many games I've actually lost - but it's at least 10 games: 

 

 

But the closer the Games Played and/or Games won get to increments of 100, the less accurate the totals get:

 

[+Nathan Strum]

7 games later, the number won has gone down, but the total aligns with the number of guesses:

 

 

8 games after that, it looks back to normal, but the guesses now add up to 311, because the "5" guesses added an extra 10 in there:

 

[+Karl G]

Thanks @Nathan Strum. I had figured before that it must have been a bizarre fluke as I could think of no way the code could have messed up the stats and then later corrected them. I will think on this one some more and see what I can come up with. 

[DeafAtariFromKansas]

I saw Nathan Strum post above, I decide to reset status and do test few games then I noticed, there is a problem between games played and guess dist.  I counted numbers of guess dist to 28 while Games played showed 29.

 

[DeafAtariFromKansas]

1 hour ago, DeafAtariFromKansas said:

I saw Nathan Strum post above, I decide to reset status and do test few games then I noticed, there is a problem between games played and guess dist.  I counted numbers of guess dist to 28 while Games played showed 29.

 

Try one more game and enter incorrect words, see what happened with status now.. Total wins - 28 and loses - 1

 

Edited March 22, 2022 by DeafAtariFromKansas
add win and lose counts

[+Karl G]

@DeafAtariFromKansas - are you playing on Stella as well? I had another possible theory for the stats issue. I'm wondering if you or @Nathan Strum are quitting Stella via the escape key before stats get fully written. It would have to be almost immediately after the game was won or lost, so it's probably a longshot.

[+Nathan Strum]

I didn't quit Stella at any time during the stretch of games covering the screenshots in my last two posts. I usually exit out of Stella from the main 2600 Words title screen several seconds after viewing the stats, specifically to try and avoid that issue.

[DeafAtariFromKansas]

32 minutes ago, Karl G said:

@DeafAtariFromKansas - are you playing on Stella as well? I had another possible theory for the stats issue. I'm wondering if you or @Nathan Strum are quitting Stella via the escape key before stats get fully written. It would have to be almost immediately after the game was won or lost, so it's probably a longshot.

First Q: Yes I playing on Stella.. Second Q: No, I quit after atariovox/savekey been written (Show UI message enabled in emulation setting). thanks for asking.. 

Edited March 22, 2022 by DeafAtariFromKansas

[+Nathan Strum]

Some more (helpful?) screenshots:

 

Guesses add up to 328; I've lost 10 games, and there are 99 "5 guess" wins:

 

3 games later, without losing a game, I've now lost 11, but the guesses total 321, and "5 guess" wins are down to 90:

 

A few games later, the numbers seem to be holding:

 

It seems as when a number (any number) first approaches 100 (around 90, I think) it stops displaying it correctly, and rolls back to 90. But I don't know how or why (or even if) it corrects itself.

[+Karl G]

10 minutes ago, Nathan Strum said:

Some more (helpful?) screenshots:

Thanks. Could you (re)send me your savekey_eeprom.dat file, please? You will need to exit Stella for it to write the current version in RAM to that file, I believe. Last time the data matched the screenshots, but just to be sure.

 

Edit: Also, the writing of data happens right after you win or lose, so by the time you are off that screen you don't have to worry about interrupting stats. being written. And it shouldn't let you go to the title page or start a new game until stats are written. Something else for me to confirm I guess.

[+Nathan Strum]

This screenshot:

 

Should match this data:

savekey_eeprom.dat

 

And this Save State:

2600words-9-NTSC.st0

 

So right now, the numbers add up, because it's already crossed the "Threshold of Wonkiness™" for the number of 5-Guess wins. It probably won't happen again until one of the numbers approaches some multiple of 100.

[+Nathan Strum]

Really hoping to see the 99 bug resolved...

 

I was closing in on a streak of 100 wins in a row...

 

 

Then I won one more game! But...

 

 

Fine. I knew I just had to keep playing more games until I passed the point where the scores started working again.

 

But then...

 

 

 

(sigh) So, I get to start all over again.  

[+Karl G]

6 hours ago, Nathan Strum said:

Really hoping to see the 99 bug resolved...

Sorry that I've not updated in a while. I could not find the cause of this issue, and then I got frustrated and worked on other stuff for a while. It seems like an obvious case of not using decimal mode when I should, but I can't find any evidence of that. If I can't figure it out, I may just share the source and see if anyone else has any idea.

 

6 hours ago, Nathan Strum said:

(sigh) So, I get to start all over again.  

At first I thought you were reporting another bug, until I looked more closely and saw that it was "just" a loss. A relief for me, but my condolences to your streak!

[john_q_atari]

I just wanted to stop by and say my 8 year old son and I have been playing this game this week and we both love it! Well done! He actually comes to me and asks me if I can load it up and we both take turns with guesses and work together trying to solve it. Thanks for creating this awesome version for the 2600.

[+Karl G]

Whelp ... I had said that if I didn't make progress on resolving the stats bug I would share the source in case anyone else is willing and able to take a look and see something that I missed. So, I'm attaching the source to this post.

 

The function in question is UpdateStats. It gets called multiple times over the course of several frames so as not to disrupt the display while the stats are being written.

 

If you are assembling and don't need to change/rebuild the dictionaries, you can just assemble with the usual "dasm 2600words.asm -l2600words.lst -s2600words.sym -f3 -v2 -o2600words.bin". If you want to rebuild the whole thing, you will need python3 in the path.

 

This issue looks exactly like a what would happen if decimal mode were not enabled when incrementing stats, but I can find no evidence of that in my code. Does anyone have any guesses as to what could be causing the stats issues that Nathan is seeing?

 

2600words.zip

[Thomas Jentzsch]

I may have a look after my vacation. If nobody is faster. ?

[+Pat Brady]

On 3/16/2022 at 7:20 AM, Karl G said:

As others have mentioned, this is how the reference game works. It took me a bit to wrap my head around when I was developing it, but the easiest way of thinking of it is that each letter in the solution is only matched once.

 

I think your implementation is not quite correct. I think the correct way is a 2-step process:

1. Loop over the solution; if the guess contains the same letter in the same position, color it green

2. Loop over the solution again; if the corresponding letter is not green and the guess contains at least one of the same letter in a different position that has not been colored, color one yellow

 

(There are other algorithms that will get the same result.)

 

For example, if the solution is FIFTH and the guess is FLUFF, the first F should be green, one of the other Fs should be yellow, and the remaining F should be gray.

 

For another example, which just happened to me, if the solution is SPASM and the guess is SPEWS, the first S and the P should be green, and the other S should be yellow. Instead the other S was gray. See screenshot. I couldn't find any other words that fit the available information so I guessed SPASM anyway.

 

Otherwise, this is really nice!

[Jamtex]

Read the whole thread but not seen this mentioned. If you put in a nonsense word of random letters and hammer the fire button, you can see the letters turn Orange or Green as it flashes, so the turn is not counted so you can find the letters and where the letters go. So it's not too difficult to get the solution in one step... One assumes it's checking the letters first, changing them, checking if it is valid and reseting it if it is not.

 

[Thomas Jentzsch]

Good catch. Here is a snapshot showing the bug:

[+Karl G]

56 minutes ago, Jamtex said:

Read the whole thread but not seen this mentioned. If you put in a nonsense word of random letters and hammer the fire button, you can see the letters turn Orange or Green as it flashes, so the turn is not counted so you can find the letters and where the letters go. So it's not too difficult to get the solution in one step... One assumes it's checking the letters first, changing them, checking if it is valid and reseting it if it is not.

 

Thanks for the bug report. I was unaware of that one. If I pick this one back up, I may well rewrite a large portion of it since there are a number of bugs that I think ultimately stem from RAM starvation. My kernel uses too much RAM leading me to reuse the same RAM for many different things.

 

The good news for me at least is that I have the source, since I posted it on here seeing help with the stats issue that Nathan experienced, so I wouldn't have to start completely from scratch if I pick it up again.