Monday, December 31, 2007

Spider Solitaire

In response to a comment by black dog barking:

Spider Solitaire is the Devil’s game.

I’m writing about the Windows version, though I’ve tried several other computer variants, and they’re pretty much as bad. Computer solitaire games are in an entirely different universe from the hard copy versions, not least being that they can keep statistics.

The skill-to-luck ratios of solitaires are pretty obvious. “Regular” solitaire, Klondike, in other words, is pretty much a matter of luck, except when you miss a play. There are a few choices during the game, but most often, there is no way to determine whether or not which choice is best. In the Windows (and most other) implementations, the “undo” is severely limited.

Freecell is much more a matter of skill, and most games are winnable. But it’s a full information game, which means that luck mostly doesn’t come in to it, and there’s a ceiling. You can only get so good at Freecell, and most people can get that good, or pretty nearly. You can also replay the same hand, over and over, and they’re numbered in the Windows version, so you can tell your friends about specific deals. At least one of them is unwinnable, I forget the number, but it can be found if you do a web search for it.

Spider has a huge amount of “top,” room to get better and better. It took me probably 20 or 30 games at the beginning before I won my first game at the highest difficulty level, where all four suits are used. Of course, that was before I began to really use the “undo” function. Without the undo, I eventually got a win statistic of slightly less than 10%.

The undo changes that entirely. The Windows version has an undo function that is limited only by each deal from the stock (which happens 5 times per game), or when a suit is formed and moved off the tableau to the foundation. Sometimes forming a full suit can cause you to lose by leaving you in an untenable position, one that you could have gotten out of if you still had the lost suit to play with.

With liberal use of “undo” I can win about 45% of Spider Solitaire hands. But a friend told me that you can Save a game at the beginning, then keep playing the same game over and over again without it counting as a loss by just going to the Open Last Saved Game (“Replay Game” counts as a loss). You can use similar Save tricks to never enter a loss in the statistics, but that’s pointless. Replaying the same game until you either win, get tired of it, or convince yourself that it’s unwinnable, though, that’s a challenge.

I once played a game that had a bottleneck (no possible plays under ordinary conditions) on the third deal; I played it for an entire weekend and found a way to create a play on the third deal that won the game. Amy plays the two suit game and once hit a game that gave no plays on the final deal, which looks like an automatic loser. But it’s possible to put cards underneath the deal that put a dealt card in play (the cards underneath have to be the same suit in sequence so they can be moved as a whole). So I noted the final deal, then played to that final layout and beat it. I’ve done similar things in the four suited game.

I did compile a statistical run of over 400 games with 21 losses; the win statistics were 94%. And there were at least a couple of games in that run that I lost because I’d forgotten to Save the game at first. More recently, easing off my OCD a little, I win about 90%, because sometimes I just get tired and go to bed.

So, we have a really difficult game that rewards pattern recognition and logical thinking by giving a statistical score that other players can find mind boggling.

The Devil’s Game, I’m sure of it. I really should give it up.

20 comments:

black dog barking said...

So Save is far more valuable than I imagined. I've never thought about Save except when prompted for it at the end of a game. As I'm then at the point of abandoning a lost cause I've always thought the notion a bit strange. Now, I see, one should Save before losing the cause.

It took me a while to get past the "cheating" aspect of compulsive undoing -- kinda like the chess "J'adoube" touch-it move-it ethic. But I eventually reframed it as shifting from visualizing the solution to recognizing it.

I saw something a while back about a form of chess where the players can use chess programs to check the tactical wisdom of the various strategies they're comtemplating, a hybrid man-machine approach playing with the strengths of silicon and carbon-based problem solving skills. Raw over the board human chess strength is not the determining factor in these matches. Evolution? Devo?

BTW, WinXP Spider Solitaire doesn't count a deal until you make a play. So you don't have to play that layout that opens with 8 kings and 2 aces.

Oh, and winning 379 of 400 SS games is insane.

James Killus said...

I think that "insane" pretty well covers it.

Anonymous said...

Did anyone notice the score system does not really work?
case in point: A game where 200 moves is the Least number of moves you can make and win, gives you the same score as a game that you could have one in 150 moves but still took 200 moves to win.
see what i mean?

James Killus said...

I have to admit, I've never paid much attention to the scoring system. I count it as a win even when my score is zero.

Anonymous said...

You guys are my kind of OCD. But enough flattery. I am under the impression that through use of the save & undo functions it is possible to win an even higher % than 379/400. I've got one on the work computer now that I am really struggling with...but I like it.
I don't look at the stats, rather finishing all suits is a win.

Anonymous said...

the windows vista version of ss doesn't limit the undo at all (unless there's no moves to undo, of course :) )
it was quite a shocker to me at first when i switched over, but it raised my win percentage up a little. i'm not nearly as hardcore player as 379 out of 400 though...

of course, undoing instead of saving means you'll get a lower score, but if all your after is the win percentage, then that doesn't really matter.

bty this post is like top hit on google for "spider solitaire strategy"

Anonymous said...

whoops... i meant it's top hit on google for "spider solitaire statistics"

Anonymous said...

Best Solitaire hit I've made!! I have a scoring question but first, Spider Solitaire played, let’s say, enthusiastically is simply a diversion. For me it takes my mind off scary issues in my life that I have no control over. This is better than dwelling on things without having to think too much. Which is better than, let’s say, enthusiastic drinking, drugging, etc. Perhaps not as much fun but ultimately better so don’t be so hard on your selves. I know; justify, justify.
Done to business. It is easy to get caught up in the game and ignore the scoring once you realize it means very little. The challenge is your percentage won in the statistic box. At the intermediate level the scoring allows only up to 1300 points as far as I can tell. You begin with 500 points and points are deducted for each move or undo you play. You are awarded 100 points for each suit stack” that is completed (8 are available) for a total of 800 added to your score at the bottom to reach your final score in a successfully played game. At any point your moves and your score will equal 500 points plus 100 point for any suit stack completed. A good score seems to be anything around 1250 meaning it took 50 moves to win the game. Once you have played over 500 moves things are not so straight forward. Beyond 500 moves you get into negative points that are, for some reason, reset to zero when you place your next undo move. You also are given your beginning score of 500 points back at some point in all this mess. I have never played past 1300 moves that I have made a note of but I am sure I have done so at some point. Somehow I have a high score of 1910 which I did not catch when I scored it and I don’t see how this can be. I have played some very, very long games and have noticed at some point you get your beginning score-count of 500 back. Could that be it? Is it possible I was awarded extra points for an unusually “short move” game won? I would really like to understand this. Can anyone tell me why or a present a maybe why?
Using the unlimited undo function I have found I can win 98% of up to around 400 games. If you are willing to play any game for quite a while practically every game can be won if you don’t have a life. I've averaged 100% of 100 games played just for grins with no re-deals. Took a picture of my statistics box with my phone and sent it to my friends who could not have cared less and pondered an intervention. Screw them. Right now I am 563 out of 610 for a 92% win ratio but I have not been trying very hard for the last 75 games or so. I do not select the hands I play. I play every deal. A good way to increase your average is to be selective about which deals you play. I love the information about using the “save” function. I will give it a go.
Another note: the cards do not seem to show the statistical spread expected. If you deal a hand with a lot of 3s and 7s you would not expect to see any more for a while but the opposite seems to be the case. Play a hand with a lot of any card in the beginning deal and you will see an unusual number of that same card in the first quarter of the game. I have not done the math but this does not make sense to me. Is there anyone else who has noticed this? If so, any insight?

Anonymous said...

In Vista spider solitaire you can save a difficult game, then go into the folder with saved games and delete it and it will not count as a loss on your statistics.

I'm trying to find a way to hack the statistics but have been unsuccessful so far.

elisa_guiga said...

oi

loopus said...

using the undo feature and the save i win EVERY single possible game.
i.e 100% of the games....
most of the times i win it at the first try sometimes i can run into very difficult games and it takes me up to 5-6 tries....anyway until now i never ran unto a game that it was impossible to solve...i guess it could be possible to demonstrate mathematically that such game doesnt exist...
such a pity that spider windows doesnt have numbers like freecell!!!

mens dress boots said...

Playing Spider solitaire suck! Never I enjoy playing this game.

zettel said...

Sorry, Loopus, but you're wrong. Not every game of SS is possible to win, even using save. A simple example shows this: on one of the deals (it doesn't matter which) every card is a king. Game over. That's very unlikely, of course, but it is possible and proves that mathematically the game isn't 100% winnable.

Here's a question: you can't deal a new set of cards while there's still a space. But what happens if you get rid of so many cards early on that you have fewer than 10 left and so cannot possibly fill all the spaces?

Has this ever happened to anyone?

Anonymous said...

Seriously dude? U play over 400 games of solitaire and keep statistics on it? Wow...

TerryFW said...

@Zettel If you have less than 10 cards you can't deal, full stop. We had been thinking about this so when it actually happened it was a small celebration rather than utter frustration.

Anonymous said...

i've played 4000 games of spider solitaire difficult version and my win rate is 40% my band's free music is at TheFezz.com

Kent Madin said...

A different topic: Is there a downloadable version of Spider that numbers the games, so that two players can compete playing the exact same game?

Kent Madin said...

A different topic: Is there a downloadable version of Spider that numbers the games, so that two players can compete playing the exact same game?

Jeff said...

@ zettel: a deal of all kings does not necessarily make the game not winnable. If it is not the last deal, you could be dealt the cards on top of those kings to complete a suit, and hence get back "under" all those kings. However, if a deal of all kings is the final deal, there would be no moves.

Just a technicality. You're still right that there exist games that are impossible to win.

vintagehotdog said...

ALL Spider games are winnable. That is what I believe and have found to be true.

Happy to prove it. If someone has a game that can not be won, just save it, send me the saved game file, and I will win it. I will send you back the saved game file with one move left to win. You can click on replay the game to see it is the same game.

Originally I noticed that you never are able to fly away all but one stack. I then decided that the program would be written this way and there would be no unwinnable games.

IMHO