....a new way of thinking

Logged in as: Guest

Users viewing this topic: none


Login  

....a new way of thinking  11/27/2004 11:42:59 AM



RPXL
Posts: 58
Joined: 2/14/2004 From: ardsley Status: offline

I use winpicks for nfl and nba and am constantly attempting to find a way of winning. Bilbal and the rest have made tremendous contributions to all of us. This is my attempt...
It's amazing how we can tune any formula to be >90% for 10 or 20 dates, but what I have realized is that NONE of them stand the test of time..that is, they all approach 50% eventually. This is a statistical expectation...so why not use it to our favor??
This is what I've tried the past few weeks and so far the results are encouraging to say the least....
I took one formula (any formula) and set it up in 11 different ways (rolling averages of 2,3,6,9,12,15,18,21,24,27, and 30) then I tuned them maximally so that all of them are 90% or more using 3 sets of dates..for nba I use 2,3,and 5 dates, for nfl I use 1 sunday, 2sundays, and 3 sundays. Then I place all the formulas in a consensus.
When a consensus of 151, or 142, or 80 for example arises, I play the OPPOSITE of the prediction. (Let's use statistics to our advantage)
My thinking is this..If all of these formulas are OVER 90%, what is the chance that 15, 14, or 8 of the formulas will continue to win and increase their winning % over the starting 90%?????????
The answer so far has been this...NOT MANY.
I would appreciate any input..
For those who understand and read this till the end I will post some plays using this technique in a separate thread
Good luck to all and lets continue to find WINNING ways...
_____________________________
Excel



RE: ....a new way of thinking  11/27/2004 6:54:13 PM



hellbringer507
Posts: 228
Joined: 8/8/2002 From: las vegas Status: offline

Is there a quick way of testing the formulas with different dates used ?
When I change the data base, from seson to date to say : rolling average last 5 game,and then analize all formalua, then go to "my formala's" and look at the list it genatates, it never changes from the default season to date. What am I doing wrong ?
Thnaks,
JEFF



RE: ....a new way of thinking  11/27/2004 7:07:38 PM



hellbringer507
Posts: 228
Joined: 8/8/2002 From: las vegas Status: offline

The more I think about your idea the more I seems to make sense to me.
The longer and stronger a formaula goes the more proability if will fail. If I can just get a grasp on this data base thing I will be ready to rock. Do you go in and change the default test method of each formaula ?
Thanks,
Jeff



RE: ....a new way of thinking  11/27/2004 7:54:49 PM



RPXL
Posts: 58
Joined: 2/14/2004 From: ardsley Status: offline

What you need to do is the following.
1. Formulas  Analyze the current formula  Test using last N dates (then set the number of dates, 1,3,5, or whatever you are testing) This only tests the formula  it doesn't tune it, but it sets the dates. Now when you tune ANY formula, you are tuning it to that number of dates..
2. Formulas  Formula Tuner  Start Tuner. (Do this until you get over 90% for all formulas)
Thats it.. Now you can repeat step 1 and set a different number for dates..
If anyone is interested I can email the formulas I use to you, but as I said, any formula will do cause you will need to tune it.
PS (When you set the database once you don't need to worry about it. The formula you are using/testing will use the statistics from the database you specify, for example, if one of your formulas is set to Rolling Average of 17 dates this only means that the rolling average of the last 17 dates is used, nothing else, a rolling average of 3 will use data only from the last 3 dates, etc.)
Hope this helps
_____________________________
Excel



RE: ....a new way of thinking  11/27/2004 8:52:32 PM



hellbringer507
Posts: 228
Joined: 8/8/2002 From: las vegas Status: offline

Yep I got it now !
Thank you very much.
Got to love that Denver Nuggets game too !!!
Jeff



RE: ....a new way of thinking  11/27/2004 10:34:08 PM



RPXL
Posts: 58
Joined: 2/14/2004 From: ardsley Status: offline

Only 22 today but lets see how it does.
Need as much input from all as we can get.
Only together will we beat this game
_____________________________
Excel



RE: ....a new way of thinking  11/28/2004 8:47:20 AM



The Ghost Of Crimson
Posts: 210
Joined: 11/25/2003 From: Status: offline

nice idea but here's the problem...first off, you're basically hoping that your formulas are ready to start coming back to normalcy...but what if they aren't...they've just proved their hot (or flukey) and you hope to go against the tide they are supposedly setting...
you didn't mention how big the sample sizes are for the 90% but they're probably only 1020 games or so...1020 games don't give you an expectancy for the next game even if it's 200, so you can't confidently expect anything...
another problem is that if a formula is tuned to say, 101, it will eventually come down in the 50s% but it can do so with a winning record because even a profittable record will bring it way down...again there's no logical expectancy for the future games and no logical reason to expect the formula to even go down to 50%...



RE: ....a new way of thinking  11/28/2004 9:52:47 AM



hellbringer507
Posts: 228
Joined: 8/8/2002 From: las vegas Status: offline

I have 7 formamulas for NBA ATS that are 100% for the last 5 dates. It is hard to find formauals that will be 100% for even 5 dates. They will not be 100% after today. So then lets go the next step and just think of picking one game with these 7. If we can get, lets say 5 of the 7 formaulas to pick just this one team to win ATS then I am betting these 5 formuals will not hit 100%. So today my pick is Sea 9
Coming off 100%, formaulas decay fast and hard. I am sick of being on the wrong end of 100%. I am not saying this is a solid theroy yet but it is worth some work.
Good Luck
Jeff



RE: ....a new way of thinking  11/28/2004 12:05:00 PM



RPXL
Posts: 58
Joined: 2/14/2004 From: ardsley Status: offline

Ghost, first of all thanks for your input, second, you bring up an excellent point but....
I have three sets of 11 formulas  total of 33 formulas that are all at or over 90% (most are actually 100%).
Now if 15 of them point to a "winner" I ask the following question.
Is there a better chance that ALL of these 15 will continue to be right, that is, remain close to 100% or is there a better chance that they will succumb to the Law of Averages and lose ???
I'm still not sure of the answer but logically the latter should prevail. This is why we need to continue testing..
I used to play the other side but kept on losing, then I started playing this way (only few weeks now) and am having much more success.
The downside to this method is having to tune 33 formulas prior to playing.
Thanks, and please keep the inputs coming..
_____________________________
Excel



RE: ....a new way of thinking  11/28/2004 6:27:13 PM



The Ghost Of Crimson
Posts: 210
Joined: 11/25/2003 From: Status: offline

two of you using the same theory, both have one play and they offset in the same game...well, one of you is going to succeed but you see my point...unless these formulas have a proven track record, which they don't, the results are going to be random and therefore there is no law of averages...
i wish you guys well and will be interested in your progress...



RE: ....a new way of thinking  11/28/2004 8:20:47 PM



RPXL
Posts: 58
Joined: 2/14/2004 From: ardsley Status: offline

Ghost, How do you play??
What method have you seen that works for you. I would appreciate your input.
Thanks
_____________________________
Excel



RE: ....a new way of thinking  11/29/2004 4:23:02 PM



The Ghost Of Crimson
Posts: 210
Joined: 11/25/2003 From: Status: offline

i try to rely on the law of averages myself but not by fading...truth of the matter is, no matter how well my formulas perform, and they seem to do well in college football each year, i don't trust my money with them unless they can show me impressive, consistent results and i don't know if that's possible...
as good as this software is, it seems too easy to just blindly play what the computer spits out even though it takes hours and hours to create the formula...



RE: ....a new way of thinking  11/29/2004 6:23:40 PM



RPXL
Posts: 58
Joined: 2/14/2004 From: ardsley Status: offline

You make an EXCELLENT point, and I agree with you to an extent.
I do think that using the formulas can be somehow used to our advantage, we just haven't found the best way yet, key word YET.
The way I look at things, mathematics and statistics have a flaw in them and this is why I am beginning to believe that this new way of thinking might work..
For example, the coin toss...we all know that there is a 50/50 chance of getting heads on 1 toss, and every subseqent toss... but what if heads came up 4 times in a row? What is the chance the 5th toss will be heads??? Math says 50/50 but I ask you this what would be the chance of tossing 5 heads in a row..we all know it woulod be .5 x .5 x .5 x.5 x.5 = 3.125%.
How can in one case the probability be 3.125% that heads will come up (cause we want 5 heads in a row) and 50% in the other case.
What I'm saying is this, although I know there is a 50/50 chance of a formula being right/wrong..I'll lean towards the other...
Thanks Ghost for your comments and lets together try to solve this....
_____________________________
Excel



RE: ....a new way of thinking  11/29/2004 9:18:14 PM



The Ghost Of Crimson
Posts: 210
Joined: 11/25/2003 From: Status: offline

rp, i like the way you think...for arguements sake, and i see it this way anyway, the longer you stretch the coin toss, the longer the odds of course...the problem is that, with a random expectancy, you never know the length of any "streak"...i've done studies on roulette wheels (50%) and blackjack results (52% at best) and what i've found is you'd be better off betting against a series of results instead of one single outcome (a formula) which is 100% random because when streaks end can't be predicted if it's 50/50...
you see, if you isolate your formulas on losing the "next" game instead of maybe "one of the next five or so games", you're clearly at a disadvantage...you'd be better off betting that your formula will lose "soon" as opposed to "now" because that very next decision is always 50% but a succession of the same result is the longshot you're talking about...
you mentioned mathematics and statistics have a flaw in them, and i understand where you're coming from, but they don't...it's just the law of mathematics and it always lays down the "law"...99% of cappers don't understand the laws though...
number freak, who some may know as a prominent member of these forums and a very good friend of mine, taught me all about numerical probabilities and once i understood the big picture i have seen everything in a different light...numbers don't lie, the solid ones hold up year after year...this is a case of numbers 101...



RE: ....a new way of thinking  12/18/2006 3:07:39 PM



kingwasabi
Posts: 3
Joined: 12/3/2006 Status: offline

bump successful strategy or not? Makes a lot of sense.



RE: ....a new way of thinking  1/6/2007 6:33:26 AM



undisputedtruth
Posts: 22
Joined: 11/10/2006 Status: offline

RPXL, your way will not hold up. look at it this way. either the system is legit or it is not. if it is legit (ie will hold up over time, and hence continue at the rate it is going), then you certainly dont want to bet against it. if it is not legit (as you think will be the case), then indeed it will drift towards a 5050 proposition, but will only be 5050 FROM THAT POINT ON. ie. if you flip a coin 40 times, and 30 come up with heads, and you then flip 100 times, you expect 50 heads and 50 tails, and not 40 heads and not 60 tails (so that you have 7070 of the 140 flips) you would expect 80 heads and 60 tails of those 140 flips. this is a very misunderstood concept (by some well educated people). trust me on this......:) PS does anybody have ANY thing long term (long term defined by CONTROL period AND POST period) that holds up?
_____________________________
The Future is Now



New Messages 
No New Messages 
Hot Topic w/ New Messages 
Hot Topic w/o New Messages 
Locked w/ New Messages 
Locked w/o New Messages 

Post New Thread Reply to Message Post New Poll Submit Vote Delete My Own Post Delete My Own Thread Rate Posts 

0.078
