222410 Bullies, 4779 online  
  • Register
Our Sponsors:

Results 11 to 20 of 43
Page 2 of 5 FirstFirst 12 345 LastLast
Sponsored Links Spacer Image
  1. Fighting Cephalopod is offline
    Fighting Cephalopod's Avatar

    Submitting 1d6 Investigators per round

    Join Date
    Jun 2005
    Location
    new jersey
    Posts
    2,981

    Posted On:
    1/27/2009 11:29am

    supporting member
     Style: ZHOO ZHITSU

    --
    Hell yeah! Hell no!
    Quote Originally Posted by Frenzal
    Are you sure?
    Am I sure that a system that gives beating Shannon Ritch and BJ Penn equal weight isn't a good predictive tool? Pretty sure, yeah.

    I don't, but my friends system which is similar worked great for the last half of the season. He went a layer deeper and used strength of oppositions opposition for the first half of the season but ran into big problems accounting for offseason moves, preseason games etc... hmmm, this isn't a league so that isn't a problem, but i think for fighters with more than half a dozen bouts he should be fine. Except, that is, for the odd circumstace like the one you mentioned or the one coming up with with Machida and Silva where one has fought world class opposition with a good record lately and another hs fought mediocre competition with a similar record in the same time frame.
    Except that this isn't an odd circumstance; it happens all the time in mma.
    Undisputed KING OF ASSHOLES.
  2. Lampa is offline
    Lampa's Avatar

    Middleweight

    Join Date
    May 2004
    Location
    Oceanside, NY
    Posts
    1,288

    Posted On:
    1/27/2009 1:42pm

    Business Class Supporting Member
     New York MMA Examiner Style: magic FUCKING powers!

    --
    Hell yeah! Hell no!
    Quote Originally Posted by Fighting Cephalopod
    Am I sure that a system that gives beating Shannon Ritch and BJ Penn equal weight isn't a good predictive tool? Pretty sure, yeah.



    Except that this isn't an odd circumstance; it happens all the time in mma.
    In defense of the system, even its creator admits it's just a jumping off point with some obvious flaws and we're taking time to work on it. We're not claiming it's the real predictor.

    Also, it's handicapping. Even really thurough systems have no absolute certainty. The Detroit Lions occassionally win a ga... oh wait.
  3. J_Treez is offline

    Registered Member

    Join Date
    Jul 2008
    Location
    Texas
    Posts
    149

    Posted On:
    1/27/2009 7:05pm


     Style: Freestyle

    --
    Hell yeah! Hell no!
    Quote Originally Posted by Fighting Cephalopod
    Given that your system would assign equal "power ranking" weight to having beaten BJ Penn and Shannon Ritch (both 7-3 in their last 10 bouts), I'm betting on "won't work at all".
    You might want to reread the OP, because you don't seem to understand the process. You've taken into acount about 40% of the work involved in determining basic power rankings. You still have to determine strength of schedule and generate power rankings as shown by your above incomplete work.

    As a side note, BJ would have a different power ranking at 155 and 170. I've been doing seperate power rankings for different weight classes.
    Last edited by J_Treez; 1/27/2009 7:24pm at .
  4. J_Treez is offline

    Registered Member

    Join Date
    Jul 2008
    Location
    Texas
    Posts
    149

    Posted On:
    1/27/2009 7:09pm


     Style: Freestyle

    --
    Hell yeah! Hell no!
    Quote Originally Posted by Fighting Cephalopod
    I compare the strengths of both fighters, then watch their previous fights and see how those strengths have stacked up against other fighters with similar skillsets or levels of ability to the person they're supposed to be fighting.
    What method do you use to generate what you feel are accurate odds, so you'll know if a bet is a good value or not? How off does a line have to be before you'll bet on it?
  5. J_Treez is offline

    Registered Member

    Join Date
    Jul 2008
    Location
    Texas
    Posts
    149

    Posted On:
    1/27/2009 7:24pm


     Style: Freestyle

    --
    Hell yeah! Hell no!
    Things I'm thinking about:

    (1) Incorporating "Strength of Schedule's Opposition". The reason I don't now is that I'm a non-spreadsheet savvy barbarian that 'caps with a pencil, a pad, and a calculator. But the more I ponder Machida vs. Silva the more it seems necessary in order to generate as accurate of a power ranking as possible. How much of a pain in the ass would this be?

    (2) Style Modifiers: Based on the charts I make of a fighters last 10 fights, we have a percentage break down of how the fighter won or lost each fight. I'm working on a way of incorporating this info so we can use it to modify our basic power rankings. Maybe this will give us a more accurate forecast as well.

    (3) What the **** do I do about people with records of less than 10 fights? Clearly having a 1.000 win % and a good power ranking because you beat 3 guys that were 1-0 is not an accurate reflection. (ie. Fujita at 4-0 from the Mark Coleman example) I'm thinking that maybe in cases where they have more fights later we could use that data as well to give a clearer picture.
    Last edited by J_Treez; 1/27/2009 7:32pm at .
  6. Matt Phillips is online now
    Matt Phillips's Avatar

    NOTE TO SELF - MOAR GRAPPLE - GET A NORMAL HAIR CUT - REPEAT

    Join Date
    May 2002
    Location
    Bahstun
    Posts
    9,466

    Posted On:
    1/27/2009 7:36pm

    supporting member
     Style: Submission Grappling

    --
    Hell yeah! Hell no!
    Quote Originally Posted by J_Treez
    ...Ok, that's the basics. There's still a lot of work to do, of course, but a lot of the fights I've been looking at have very similar odds to what this basic method comes up with, so I don't think I'm entirely off base. I still need to figure out the effect of styles vs. styles on odds and a few other things I'm sure you all will point out. I'm looking forward to discussions, questions, answers, and of course being called a stupid fucking n00b bastard. I'll post my Odds conversion Chart later today. I hope I was as concise as possible and made sense.
    Since your system is based on averages, there is an implicit appeal to the Law of Large Numbers. You should expect to see a return on your betting only in the long term. A method of this sort has very limited predictive power when applied to an individual match.
    I hope you have deep pockets.
    Now darkness comes; you don't know if the whales are coming. - Royce Gracie


    KosherKickboxer has t3h r34l chi sao

    In De Janerio, in blackest night,
    Luta Livre flees the fight,
    Behold Maeda's sacred tights;
    Beware my power... Blue Lantern's light!
  7. J_Treez is offline

    Registered Member

    Join Date
    Jul 2008
    Location
    Texas
    Posts
    149

    Posted On:
    1/27/2009 7:55pm


     Style: Freestyle

    --
    Hell yeah! Hell no!
    Quote Originally Posted by PeopleSoft
    Since your system is based on averages, there is an implicit appeal to the Law of Large Numbers. You should expect to see a return on your betting only in the long term. A method of this sort has very limited predictive power when applied to an individual match.
    I hope you have deep pockets.
    That's a good point for anyone wanting to put this (or any other) system into practical use. Since that's the point, I should point out how to use it in conjunction with proper bankroll management. The only thing you're wrong about is the "deep pockets" thing. You can make $5 bets on Sportsbook.com, which would only require around a $200 bankroll to get started.

    It's pretty much Gambling 101 that if you seriously want to bet and win, set aside a certain amount of money that you can afford to lose and earmark it strictly for betting on MMA. That's called your bankroll. What we want to do when determining how much to wager is to bet enough so that when we win, we increase our bankroll by 3%. That way, if you you truly have an edge, you should be relatively safe from bad luck, without going into a drawn out discussion of standard deviation, regression to the mean, and other such mathematical topics.

    Example: I've got $1000 in my bankroll and I think GSP is gonna whoop B.J.'s ass. 3% of my $1000 is $30 bucks and GSP is a -165 favorite, so I'll bet $49.50 to win $30. Or instead of being a tight ass, just bet $50.
    Last edited by J_Treez; 1/27/2009 8:31pm at .
  8. J_Treez is offline

    Registered Member

    Join Date
    Jul 2008
    Location
    Texas
    Posts
    149

    Posted On:
    1/28/2009 9:49am


     Style: Freestyle

    --
    Hell yeah! Hell no!
    Odds watch:

    GSP -200
    BJ Penn +155

    I hope you got GSP at -165.

    Machida -285
    Silva +225

    Looks like money's going down on Thiago as he drops from +250. Machida is a MUCH better value here than at his previous price of -325.

    Bonnar -170
    Jones +140

    Line remains static. I haven't had time to analyze this fight. I wonder if Bonnar is a value?



    Boxing:

    Cotto -2500
    Jennings +1250

    Both fighters have only one career defeat. Cotto has fought DRASTICALLY tougher competition. (.936 vs. .781 difficulty over last 10 fights) Still, you don't see a +1250 dog every day. $10 will get you $125. Just something to think about.
  9. Fighting Cephalopod is offline
    Fighting Cephalopod's Avatar

    Submitting 1d6 Investigators per round

    Join Date
    Jun 2005
    Location
    new jersey
    Posts
    2,981

    Posted On:
    1/28/2009 1:08pm

    supporting member
     Style: ZHOO ZHITSU

    --
    Hell yeah! Hell no!
    Quote Originally Posted by J_Treez
    You might want to reread the OP, because you don't seem to understand the process. You've taken into acount about 40% of the work involved in determining basic power rankings. You still have to determine strength of schedule and generate power rankings as shown by your above incomplete work.
    I was using only part of the process to demonstrate why your power rankings are inaccurate. As long as they only take into account the opposition's win percentage and not "strength of opposition's opposition", you have the incredibly common circumstance of fighters whose win percentages are higher than others, but based on records that are padded with tomato cans. There are literally scores of examples of people who are drastically superior fighters to people with higher win percentages.

    Quote Originally Posted by J_Treez
    What method do you use to generate what you feel are accurate odds, so you'll know if a bet is a good value or not?
    I find out as much about the fighters as I can, and then make a judgement call.
    Undisputed KING OF ASSHOLES.
  10. Fighting Cephalopod is offline
    Fighting Cephalopod's Avatar

    Submitting 1d6 Investigators per round

    Join Date
    Jun 2005
    Location
    new jersey
    Posts
    2,981

    Posted On:
    1/28/2009 1:10pm

    supporting member
     Style: ZHOO ZHITSU

    --
    Hell yeah! Hell no!
    Quote Originally Posted by J_Treez
    Things I'm thinking about:

    (1) Incorporating "Strength of Schedule's Opposition". The reason I don't now is that I'm a non-spreadsheet savvy barbarian that 'caps with a pencil, a pad, and a calculator. But the more I ponder Machida vs. Silva the more it seems necessary in order to generate as accurate of a power ranking as possible. How much of a pain in the ass would this be?
    This would be a good start.

    As far as how big a pain in the ass it would be, the major fighter databases are available online. Find someone who can do simple number crunching with some database queries and it's not too difficult.
    Undisputed KING OF ASSHOLES.
Page 2 of 5 FirstFirst 12 345 LastLast

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  

Powered by vBulletin™© contact@vbulletin.com vBulletin Solutions, Inc. 2011 All rights reserved.