Subject: "winnable" coming

Date: Tue May 3 19:50:40 2011
User: Denny
Message:
All you folks who've complained because you lost a Lucky 13 streak due to one of the unwinnable games, that will be changing shortly. free@last has worked thru the majority of the games here and developed data showing whether a game can be proven winnable, and I've got that loaded into the database. It's still going to be a good bit of work to make the necessary changes to the server, clients, and site, but it's in progress.

Date: Tue May 3 19:59:51 2011
User: Catailong2
Message:
That's great news! Thanks for all of your hard work Denny. Your dedication to making this site better and listening to the players is sincerely appreciated. I'm excited to see what you have in store.

Date: Tue May 3 20:17:49 2011
User: firenze
Message:
Adding or substituting, Denny?

Date: Tue May 3 20:30:11 2011
User: Punster
Message:
THANKS a BUNCH Denny & free@last & anyone else who helped. I like knowing if I can't win a game, that it's just because I didn't make the RIGHT moves somewhere along the way, rather than being an "unwinnable" game !!! Thanks again !!

Date: Tue May 3 22:00:01 2011
User: luxfero
Message:
Thanks Denny, your efforts are greatly appreciated.

Date: Tue May 3 22:14:32 2011
User: luxfero
Message:
Rats. Just hit an unwinnable in Lucky 13, killed a medium nice streak. I am really, really, looking forward to the new version.

Date: Wed May 4 02:41:01 2011
User: joeygray
Message:
To each his own. As long as you not getting rid of the game set we have now, I got no problem. Did I hear you say you were just doing 13x0 for the time being?

Date: Wed May 4 06:31:48 2011
User: BackToBack
Message:
I love it!

Date: Wed May 4 10:46:31 2011
User: oonwayvos
Message:
Wonderful! You have a great site and one that makes my job bearable. fwiw: I would love to see you use the win rates of games to filter them into difficulty classes so that level 5 were easy, 6 a little harder etc. Thanks for all the work you put into the games. They truly are fun.

Date: Wed May 4 12:11:12 2011
User: anicca
Message:
I also am most grateful for all your consistent care for us all, Denny! One thing i have wondered, have you ever given any thought to scoring relative to "Number of moves"? Thanks, Anicca

Date: Wed May 4 13:35:44 2011
User: free@last
Message:
I think the named variants should all have winnable versions (already the case for 8x4, 9x5 and 10x6, I'd add 13x0, 12x1 and 7x4). On a selfish note, I'd like to see a winnable 4x10 as that is my new favorite. If I could only pick one (not already winnable) from each width, I'd go with: 4x10, 5x8, 6x5, 7x4, 8x3, 9x2, 10x1, 11x1, 12x1, 13x0

Date: Wed May 4 14:28:28 2011
User: Buffy
Message:
Some good choices there, free@last. Would definitely be interested to see a 10x1 winnable, although I think I'd be up for most of the 11-sum variants or lower, to be honest.

Date: Wed May 4 15:10:21 2011
User: dogtuba
Message:
My preference is for a game that is driven by random presets, nothing guaranteed, nothing fixed. That's how the universe seems to work.

Date: Wed May 4 15:18:20 2011
User: free@last
Message:
I guess I should explain my reasoning. Here's a look at the number of unwinnable games (as per the solver) out of 32768 possible. These are for the -10 difficulty. (Note: 4x and 5x are incomplete.) I tried to pick the "sweet spot" for each width where it would still include most of the games (>95% including the lesser difficulties) but not just eliminate a few "bad deals". 20236 4x8-10.txt 9562 4x9-10.txt 2773 4x10-10.tx 544 5x8-10.txt 55 5x9-10.txt 9 5x10-10.tx 32490 6x2-10.txt 27720 6x3-10.txt 13371 6x4-10.txt 2402 6x5-10.txt 174 6x6-10.txt 19 6x7-10.txt 1 6x8-10.txt 0 6x9-10.txt 0 6x10-10.tx 32502 7x1-10.txt 24964 7x2-10.txt 7330 7x3-10.txt 532 7x4-10.txt 30 7x5-10.txt 0 7x6-10.txt 26717 8x1-10.txt 6871 8x2-10.txt 273 8x3-10.txt 30817 9x0-10.txt 11481 9x1-10.txt 519 9x2-10.txt 9 9x3-10.txt 4 9x4-10.txt 2 9x5-10.txt 0 9x6-10.txt 22746 10x0-10.tx 2275 10x1-10.tx 18 10x2-10.tx 0 10x3-10.tx 0 10x4-10.tx 11930 11x0-10.tx 200 11x1-10.tx 0 11x2-10.tx 0 11x3-10.tx 4769 12x0-10.tx 16 12x1-10.tx 0 12x2-10.tx 1654 13x0-10.tx 0 13x1-10.tx

Date: Wed May 4 15:37:35 2011
User: huffy
Message:
2 level 10 9*5's unwinnable??

Date: Wed May 4 15:40:22 2011
User: firenze
Message:
I found that hard to believe, too. Do you know the numbers, free@last?

Date: Wed May 4 15:43:56 2011
User: free@last
Message:
They are not unwinnable. They were unsolvable by the solver. On some games it runs out of memory. There are too many for me to double check. Example in link.

Link: solver output sample

Date: Wed May 4 17:56:49 2011
User: huffy
Message:
Given I have put zero time into this maybe I shouldnt question it but "unwinnable" and games where the solver runs out of memory are not necessarily the same thing. From the game stats page I beleive that all level 10 9*4 have been played and won. As is clearly evidenced there are games that have been deemed "unwinnable" that are winnable and in the 9*5 variant no level 10 game has less than a 69% win rate. Based on that one would should assume that there are going to be many other winnable games (other than 9*4 & 9*5) that are also removed. In taking out the games based on this method are we not taking out more than just unwinnable games but the harder games (well I assume they are harder because they have caused the solver to run out of memory). Maybe this can be the called "childrens version". No unwinnable and no really hard games.

Date: Wed May 4 18:16:10 2011
User: free@last
Message:
huffy, you assume incorrectly that running out of memory equates to harder games. I have played dozens of these and most of them were, dare I say, on the easier side. Why does the solver get hung up on them? I don't know. (I didn't write the solver.) I can tell you that the solver is not as good as it could be. In some solutions it will do something like move a card from a freecell down to an empty column, then move a different card to a foundation, then move the first card back to a freecell! It is also not smart enough to immediately put cards to the foundation like Denny's auto-play algorithm. Anyway, I can assure you that if Denny puts up winnable 7x4, you will not find the games childish. :)

Date: Thu May 5 08:20:12 2011
User: joeygray
Message:
Yikes. Basing a 'winnable' version of Freecell on the results of a solver already known to be unable to solve a solvable game doesn't seem like the 'right thing to do' at all.