Page 8 of 9 FirstFirst ... 3456789 LastLast
Results 141 to 160 of 177
  1. TRgEOff's Avatar
    Posts
    589 Posts
    Global Posts
    591 Global Posts
    #141  
    Quote Originally Posted by boberang View Post
    hey, if going the route of non-point for wins & 1/2 points for draw what about:

    win: 2 * op win %
    draw: 1 * op win %
    loss: 0 pts

    this way quality of win comes into play. player 1 is 22-4 (0.846) & player 2 is 10-11 (0.476) so:

    hehe, I think hteasthope's scheme was motivated by a desire to eliminate 1/2 points. Of course the three decimal points of your scheme could be eliminated by just multiplying everything by 1000.

    but a deeper "disadvantage" of all of the systems suggested to date, compared to an Elo-style system, is that they don't account for the difference in ratings of the two
    players. Thus a weak player who wins against a very strong player gets the same reward as a strong player.

    Equally problematic is that they are one-sided - there is no change in ratings of the losing player. Also, because they are cumulative systems, and without potential to have the equivalent of "negative" ratings, it would mean that "members" who might be weak, but have played a lot of games, could retain much higher ratings than very strong players who were new to the "club". Eventually a strong player might catch up, but one needs a system that is much more responsive to playing strength. (a "negative" rating in a ratings schemes is achieved by having a pre-defined "average rating" which is the equivalent of a zero point (e.g. 1600 on ICC)).

    These might seem minor points at first, but unless a system is set up sensibly, you will eventually find:

    * that the ratings can be meaningless. This defeats the primary purpose of a ratings system which is to allow people to identify players whose playing strength is close enough to their own to make playing enjoyable.

    * weird patterns of opponent choice are being made as ratings-oriented players manipulate the idiosyncrasies of the system. The "meta-game" of ratings starts to distort the real game!

    A fair and accurate ratings system must reward and penalise wins and losses, and do so in way that is sensitive to the ratings difference between the two players. The Elo system, used in one form or another in many chess ratings schemes, is spelled out here in Wikipedia and here on the ICC site, The basic idea is not as complicated as it looks at first. The most difficult aspect to implement, which is not described in sufficient detail there, nor anywhere that I know of as yet, are the "special rules" that need to apply within a new member's "provisional" period . The provisional period is a necessary feature when using the basic Elo system. FICS uses a slightly more advanced system based on work by Glickman, the "Glicko" scheme, which removes the need for a provisional period by introducing a Ratings Deviation variable (see FICS help file). All the algorithms used by FICS are very clearly covered in that file.

    Ratings schemes are an art form, but any system set up for chess would also work for other wifight games (Elo or the Glicko system are used in many online games, eg World of Warcraft). I don't think what I'm suggesting here is a fancy and sophisticated approach, basically I think it's the bare minimum that would actually work.

    I know I was enthusastic about a Win-Loss scheme in an earlier post, but I hadnt really thought it out. The traditional 1-1/2-0 system works fine for tournaments where you have a closed system of players/games, but it's not applicable to open server-based game play.

    The only way I see for a simple win-loss system to be informative is if there were a game results matrix on the web somewhere, with all players listed along the vertical and horizontal axes. It would report the record of each player's games against each other player (e.g. the BSpooky- TrGeoff intersection would now report 5-1). Once any new player has played a few games, and thus has a sense of the playing strength of the opponents they have played against, they can then check who else those players have won-lost against, and thus pick suitably strong opponents for their next games. We sort of do this already in an informal way, based on the results of games reported in the ChessChumps thread.

    It would be informative - until the number of players got so high as to make it visually unmanageable!

    cheers
    Last edited by TRgEOff; 09/22/2007 at 05:50 AM.
    geoff

    Treo 650 1.20 LAP.
  2. #142  
    Quote Originally Posted by TRgEOff View Post
    .....

    Ratings schemes are an art form, but any system set up for chess would also work for other wifight games (Elo or the Glicko system are used in many online games, eg World of Warcraft). I don't think what I'm suggesting here is a fancy and sophisticated approach, basically I think it's the bare minimum that would actually work.

    ......
    Well, after reading this post and poking around on the web I am in full agreement with this......going with Glicko for simplicity of not needing a provisional rating period sounds best to me, but elo would be good too. The great thing is it does appear to be able to work in any game scenario with the Glicko.

    checkers, hey, what about othello?, chess, etc. etc. etc.

    Though it would be also nice to keep track of W-L just for an interesting sake. So maybe a player would be 1340 rating, but say 72-55. Rating is interesting measuer of strenght, but the W-L is an interesting measurement of how active, etc. a player is.

    I would vote to move this way up on the priority scale if Basu is willing to do it!!!
  3.    #143  
    Well, I *think* I fixed chess, finally. This time I'm pretty sure I found the problem. I'll keep on testing though.
  4.    #144  
    Quote Originally Posted by boberang View Post
    I would vote to move this way up on the priority scale if Basu is willing to do it!!!
    I definitely want to do rankings. Although per-opponent win/loss stats will probably come first, since that's a bit easier to implement.
  5. TRgEOff's Avatar
    Posts
    589 Posts
    Global Posts
    591 Global Posts
    #145  
    Quote Originally Posted by Brennan View Post
    Well, I *think* I fixed chess, finally. This time I'm pretty sure I found the problem. I'll keep on testing though.
    yes castling is working again now, great---- can back to playing!! Thanks.

    one quirky thing Ive noticed since 0.10 is that the Treo's auto time-out (well actually I use Auto-off) wont work if wifight is left on. After the time-out period, the screen flashes off-and-on again rapidly for 1-5 times, then stabilises to stay on until the time-out period expires again, at which point the cycle repeats. Nothing of real concern, and could be some idiosyncrasy of my own highly modified Treo. Just thought you'd like to know.

    good idea about per-player W-L record.

    if you can think of any development tasks you could farm out, I for one would be happy to help. Ive no programming background, but maybe there other things that I could do that would help the project along.

    cheers
    geoff

    Treo 650 1.20 LAP.
  6. #146  
    Quote Originally Posted by TRgEOff View Post
    yes castling is working again now, great---- can back to playing!! Thanks.

    one quirky thing Ive noticed since 0.10 is that the Treo's auto time-out (well actually I use Auto-off) wont work if wifight is left on. After the time-out period, the screen flashes off-and-on again rapidly for 1-5 times, then stabilises to stay on until the time-out period expires again, at which point the cycle repeats. Nothing of real concern, and could be some idiosyncrasy of my own highly modified Treo. Just thought you'd like to know.

    good idea about per-player W-L record.

    if you can think of any development tasks you could farm out, I for one would be happy to help. Ive no programming background, but maybe there other things that I could do that would help the project along.

    cheers
    Im actually not getting this issue on my 755P....do u happen to have mTools installed? I had this happen when I installed and was trying that program out. once I change some of the light control settings it stopped.
    Sprint Palm Pre
  7. TRgEOff's Avatar
    Posts
    589 Posts
    Global Posts
    591 Global Posts
    #147  
    Quote Originally Posted by hteasthope View Post
    Im actually not getting this issue on my 755P....do u happen to have mTools installed? I had this happen when I installed and was trying that program out. once I change some of the light control settings it stopped.
    Thanks for that advice, hteasthope. No mTools, but I did install another Tools app, PowerHero recently, so I tried uninstalling that. That had some effect- the screen now flickers on-and-off repeatedly without ever stabilising (Note: this only occurs with wifight, not other apps).

    Does make you wonder how much of the PowerHero program is (designed to be?) NOT removed by Uninstall Manager. Obviously some aspects of it arent because, when I re-installed it, UIM didnt recognise it as a new app (but neither did this return the flickering pattern to the previous pattern).

    cheers
    geoff

    Treo 650 1.20 LAP.
  8.    #148  
    Quote Originally Posted by TRgEOff View Post
    yes castling is working again now, great---- can back to playing!! Thanks.

    one quirky thing Ive noticed since 0.10 is that the Treo's auto time-out (well actually I use Auto-off) wont work if wifight is left on. After the time-out period, the screen flashes off-and-on again rapidly for 1-5 times, then stabilises to stay on until the time-out period expires again, at which point the cycle repeats. Nothing of real concern, and could be some idiosyncrasy of my own highly modified Treo. Just thought you'd like to know.

    good idea about per-player W-L record.

    if you can think of any development tasks you could farm out, I for one would be happy to help. Ive no programming background, but maybe there other things that I could do that would help the project along.

    cheers
    Hi again, sorry it took me so long to respond.

    I'm not too sure about what's causing that weird behavior. I'm trying to just follow the SDK docs on going into REM sleep mode. I'll try to figure it out.

    Something that would be very helpful would be (if you feel like it) to improve the wifight wiki. The Chess page in particular could probably use some love.

    BTW I set the wiki edit password to "wifight" but only to deter botspam. Please feel free to edit as much as you like.
  9.    #149  
    So to keep you guys posted, the web site has been getting some work. Pretty soon you'll be able to log in and change some settings and get some stats.

    I also started on Checkers a bit. The pieces move around but jumping is still unimplemented. In fact it looks like it will be a bit of a UI challenge since you can have two different jumping sequences for the same piece that can end up on the same square.

    Also, I managed to dunk my 680 into the ocean yesterday. It didn't get too wet but the hotsync cable button circuit must be shorted. It just registers continuously. So I'm leaving it to dry by the window. Luckily I still have my 650 as a (hopefully temporary) backup.
  10. #150  
    I like the update status screen on the ver 0.10 but it also does it when you have just made a move. Maybe it's just me but why does it have to go to that screen?
  11. TRgEOff's Avatar
    Posts
    589 Posts
    Global Posts
    591 Global Posts
    #151  
    Quote Originally Posted by Brennan View Post
    So to keep you guys posted, the web site has been getting some work. Pretty soon you'll be able to log in and change some settings and get some stats.

    I also started on Checkers a bit. The pieces move around but jumping is still unimplemented. In fact it looks like it will be a bit of a UI challenge since you can have two different jumping sequences for the same piece that can end up on the same square.

    Also, I managed to dunk my 680 into the ocean yesterday. It didn't get too wet but the hotsync cable button circuit must be shorted. It just registers continuously. So I'm leaving it to dry by the window. Luckily I still have my 650 as a (hopefully temporary) backup.
    whoa, bad luck with the 680. any water is bad enough, but marine immersion is almost the worst! I hope it recovers.

    Ive been able to logon to the wiki, and will get around to some editing soon. I will start by migrating over a version of what Ive written at the top of the chumps list.

    am intrigued by the upcoming web page developments!

    like ronbo, I also have the update thing happen when making moves but just thought it was part of making the whole thing more stable. I like the whirling atoms, very cool.

    good luck with checkers

    cheers
    geoff

    Treo 650 1.20 LAP.
  12. #152  
    Brennan,

    I go back to the games that were just completed and it would not me review the game.
  13.    #153  
    Quote Originally Posted by ronbo2000 View Post
    Brennan,

    I go back to the games that were just completed and it would not me review the game.
    I'm not sure I understand, ronbo2000. Can you give me some more details please?
  14.    #154  
    Quote Originally Posted by ronbo2000 View Post
    I like the update status screen on the ver 0.10 but it also does it when you have just made a move. Maybe it's just me but why does it have to go to that screen?
    Well, I guess it doesn't absolutely have to but it seemed more consistent/reliable this way. This way you know whether your turn successfully submitted or not right away. Did you prefer the mini "Updating..." popup over the gameboard then? The problem with doing it that way is that if you exit the app before the turn submits completely you might think it submitted ok when it hadn't yet.

    What do you think?
  15.    #155  
    Quote Originally Posted by TRgEOff View Post
    whoa, bad luck with the 680. any water is bad enough, but marine immersion is almost the worst! I hope it recovers.

    Ive been able to logon to the wiki, and will get around to some editing soon. I will start by migrating over a version of what Ive written at the top of the chumps list.

    am intrigued by the upcoming web page developments!

    like ronbo, I also have the update thing happen when making moves but just thought it was part of making the whole thing more stable. I like the whirling atoms, very cool.

    good luck with checkers

    cheers
    It (the 680) seems to be working OK now. Whew! I'm gonna give it another couple days to be sure, though.

    Please do edit the wiki Chess page! That would be awesome. I look forward to seeing the changes.

    I should probably hurry up and just launch the login feature on the web page. Hopefully before the weekend.

    Re: checkers, I actually prefer playing chess to checkers! So chess is still getting more of my attention. But checkers is coming along too.

    Speaking of chess, I remember some time back you posted about the blue primary selection squares obscuring the light/dark board squares, so I was experimenting with lightening some of the selection squares based on the tile color:
    http://shup.com/byaa!/5832/107941323...ase.rom%5D.png

    Also, I tried having primary selection squares hidden during secondary selection:
    http://shup.com/byaa!/5833/107941326...ase.rom%5D.png

    Let me know what you guys think!
  16. TRgEOff's Avatar
    Posts
    589 Posts
    Global Posts
    591 Global Posts
    #156  
    Quote Originally Posted by Brennan View Post
    let me know what you guys think!

    Good news re 680, I'm sure it was planned as part of your rigorous alpha testing process

    Yes checkers, not a fan either, im afraid

    wiki: will get onto it next week, this one has been flat-out.

    improvements: I particularly like the second (2ry selection) change, because that really helps with seeing the dark squares. I think even in the primary selection phase I would select some piece that didnt control (and thus partly obscure) many squares, just so i can think all possible moves more easily.

    the first change helps a bit but I think i would really need to see it on the treo to give a proper judgement. it definitely doesnt hurt.

    looking forward to the login feature!

    also: do you have any idea of the total # of games that have been played on wifight to date?

    also: do you want more testers at this stage? you or I could post something over at mytreonet to increase the coverage.

    cheers
    geoff

    Treo 650 1.20 LAP.
  17. wilsonb's Avatar
    Posts
    434 Posts
    Global Posts
    439 Global Posts
    #157  
    A nice Feature would be to send an SMS message to the opponent to let them know it's their move.

    Thanks for an awesome product..!
    ============,,,================
    ===========(o o)===============
    =======oOO==(_)==OOo===========
    Treo 755p ;-)
    Tre 700p UPdated..
    Treo650- Sprint custom rom
    Tungsten T3 creative design.
    http://www.WilsonBradley.com
    http://www.myspace.com/wilsonbradley
    AOL screen name; WilsonIsON
    MSN; WilsonIsOn
    Yahoo; Wilson2bw
    ========ooooO==Ooooo==========
    =========( )===( )==========
    ==========\ (====) /===========
    ===========\_)==(_/============
  18. #158  
    Brennan,

    Just finished playng the Tic Tac Toe game with you. I had all my X lined up vertically on the first column. When it said I won, the line crossed on the last row instead of the first column of X's.
  19. #159  
    Quote Originally Posted by ronbo2000 View Post
    Brennan,

    I go back to the games that were just completed and it would not me review the game.
    Quote Originally Posted by Brennan View Post
    I'm not sure I understand, ronbo2000. Can you give me some more details please?
    I looked at a completed game and hit the < button to review the moves and nothing changes. I did this on a number of times on different days. However, it's working ok now.
  20. #160  
    Quote Originally Posted by Brennan View Post
    Well, I guess it doesn't absolutely have to but it seemed more consistent/reliable this way. This way you know whether your turn successfully submitted or not right away. Did you prefer the mini "Updating..." popup over the gameboard then? The problem with doing it that way is that if you exit the app before the turn submits completely you might think it submitted ok when it hadn't yet.

    What do you think?
    It is nice with the mini popup but am I right in understanding that with the way you have it now that if you exit the application before the turn submits completely this version works better? Its just aesthetics preference, you can leave it as it is if reliability is a better.
Page 8 of 9 FirstFirst ... 3456789 LastLast

Posting Permissions