JoelDeze
JoelDeze
Joined: Apr 20, 2016
  • Threads: 43
  • Posts: 343
March 12th, 2017 at 8:40:20 PM permalink
Bracket 8: This is the final bracket but it does not have a 40-game qualifier. (+7 ATS) (71.4% win pct)

GamesWins on System FavoriteLosses on System FavoritePushes on System FavoriteWin Pct
2215610.714285714


12015CharlotteGeorgia St.725023 - 20yes31
32015San Jose St.Oregon St.723035 - 21no140
32015Ole MissAlabama725043 - 37yes61
72015SyracuseVirginia723044 - 38no61
72015Kent St.Massachusetts724015 - 10yes51
72015FloridaLSU724035 - 28no7push
82015Kansas St.Texas724023 - 9no140
162015Air ForceCalifornia722055 - 36no190
162015TCUOregon724047 - 41yes61
162015HoustonFlorida St.723038 - 24yes141
22016IllinoisNorth Carolina723548 - 23no250
32016Michigan St.Notre Dame725036 - 28yes81
42016GeorgiaOle Miss724045 - 14no310
42016BYUWest Virginia725035 - 32no31
42016North TexasRice724042 - 35yes71
42016Wake ForestIndiana725033 - 28yes51
82016TempleSouth Fla.722046 - 30yes161
92016UConnEast Carolina723041 - 3no380
92016Army West PointWake Forest724021 - 13yes81
102016NavyNotre Dame723028 - 27yes11
112016SMUEast Carolina724055 - 31yes241
162016NavyLouisiana Tech722048 - 45no31


The remainder +ATS nets a rough 52% which is not advantageous enough to post.
“Know where to find information and how to use it; that is the secret of success.” – Albert Einstein
JoelDeze
JoelDeze
Joined: Apr 20, 2016
  • Threads: 43
  • Posts: 343
March 12th, 2017 at 8:47:21 PM permalink
Totals:

63.6% was the final from the combined ATS brackets. The inconsistency in the first posting was that my first snapshot was from development data and not from production data and I had not accounted for all games played in that DB. I have not gone through the data to determine any type of additional filtering. So, I do believe this can be improved.

Final is (324 - 185, 14 pushes) (63.6% win pct)

BracketGamesWinsLossesPushesWin Pct
147341210.739130435
244261530.634146341
395573710.606382979
4101633800.623762376
543281320.682926829
6113654350.601851852
758362110.631578947
82215610.714285714
Totals523324185140.63654224
“Know where to find information and how to use it; that is the secret of success.” – Albert Einstein
SOOPOO
SOOPOO
Joined: Aug 8, 2010
  • Threads: 89
  • Posts: 3953
March 14th, 2017 at 10:48:47 AM permalink
Quote: JoelDeze

Totals:

63.6% was the final from the combined ATS brackets. The inconsistency in the first posting was that my first snapshot was from development data and not from production data and I had not accounted for all games played in that DB. I have not gone through the data to determine any type of additional filtering. So, I do believe this can be improved.

Final is (324 - 185, 14 pushes) (63.6% win pct)

BracketGamesWinsLossesPushesWin Pct
147341210.739130435
244261530.634146341
395573710.606382979
4101633800.623762376
543281320.682926829
6113654350.601851852
758362110.631578947
82215610.714285714
Totals523324185140.63654224



Are you saying now that you WERE NOT 417-125, but rather 324-185? My brain hurts trying to go through all the charts....
JoelDeze
JoelDeze
Joined: Apr 20, 2016
  • Threads: 43
  • Posts: 343
March 21st, 2017 at 4:06:34 AM permalink
Quote: SOOPOO


Are you saying now that you WERE NOT 417-125, but rather 324-185? My brain hurts trying to go through all the charts....



I'm saying that after researching and looking at all of the ATS data over the last 2 years that the charts above represent any advantage within each ATS bracket. The totals are the combined results for that research data. Read through the data and determine whether it has any value to you.

That is not my overall record. My overall win pct is barely above 52%. But, I've only been diving into ATS picks over the last two seasons.

I primarily just provide research. My research data over many years was built to provide high probability value for which team would win the outcome of a game. It was never designed to measure ATS. But, like insurance principles, everyone has a risk appetite and so some people like higher value on returns at the cost of more risk. Therefore, I am working more on ATS research data and profiling.

I hope that answers your question.
“Know where to find information and how to use it; that is the secret of success.” – Albert Einstein
WatchMeWin
WatchMeWin
Joined: May 20, 2011
  • Threads: 45
  • Posts: 581
March 21st, 2017 at 8:27:36 AM permalink
JD, Unless you are posting your picks prior to the games starting, you will not convert anyone here. Anyone can post whatever they want post games played where the results are as you wish them to be. If you have done this somewhere in this forum/thread, then my apologies for not reading all the mumbo jumbo.

Look at what Rising Dough are Steeldco are doing... they are posting their picks prior to game starting. These are the only picks I will pay attention to or give merit to.

btw, I went 19-1 the past two weeks in hoops. See what Im sayin?
'Winners hit n run... Losers stick around'
JoelDeze
JoelDeze
Joined: Apr 20, 2016
  • Threads: 43
  • Posts: 343
March 21st, 2017 at 9:00:38 AM permalink
Quote: WatchMeWin

JD, Unless you are posting your picks prior to the games starting, you will not convert anyone here. Anyone can post whatever they want post games played where the results are as you wish them to be. If you have done this somewhere in this forum/thread, then my apologies for not reading all the mumbo jumbo.

Look at what Rising Dough are Steeldco are doing... they are posting their picks prior to game starting. These are the only picks I will pay attention to or give merit to.

btw, I went 19-1 the past two weeks in hoops. See what Im sayin?



I'm not trying to convert anyone and I don't want to be viewed as a tout going forward. The data I posted is for informational purposes only. If you find the data useful then great.

My goal is finding discovery data that may provide an advantage. My sole concentration is on football.
“Know where to find information and how to use it; that is the secret of success.” – Albert Einstein
WatchMeWin
WatchMeWin
Joined: May 20, 2011
  • Threads: 45
  • Posts: 581
March 21st, 2017 at 9:54:21 AM permalink
Quote: JoelDeze

I'm not trying to convert anyone and I don't want to be viewed as a tout going forward. The data I posted is for informational purposes only. If you find the data useful then great.

My goal is finding discovery data that may provide an advantage. My sole concentration is on football.



Are you in the handicapping or sports picking business?
'Winners hit n run... Losers stick around'
bazooooka
bazooooka
Joined: Nov 21, 2016
  • Threads: 3
  • Posts: 182
May 8th, 2017 at 6:44:32 PM permalink
Joel,

Can you post your NCAA data just for 2015 and then select the best trends where you think you system and the line correlate? Having found those trends, how did they do in 2016? If you can do this going back many years that would be indeed be impressive. Isn't it likely though that the correlations and trends are near random and that they don't hold up as well for the following seasons. However if you found a trend back in 2013 and then it worked in 2014-2016 that indeed would be interesting over 1000+picks.
JoelDeze
JoelDeze
Joined: Apr 20, 2016
  • Threads: 43
  • Posts: 343
May 9th, 2017 at 11:58:58 AM permalink
Quote: bazooooka

Joel,

Can you post your NCAA data just for 2015 and then select the best trends where you think you system and the line correlate? Having found those trends, how did they do in 2016? If you can do this going back many years that would be indeed be impressive. Isn't it likely though that the correlations and trends are near random and that they don't hold up as well for the following seasons. However if you found a trend back in 2013 and then it worked in 2014-2016 that indeed would be interesting over 1000+picks.



I'll answer your question by posting the top 3 trends for college.

1. [2016] TSRS (True Statistical Rating of Strength): Accounts for 71.75% Win Probability
1. [2013,2014,2015] TSRS: Accounted for a 70.82% Win Probability

2. [2016] RERDiff (Rushing Efficiency Rating Difference): Accounts for 66.40% Win Probability
2. [2013,2014,2015] RERDiff: Accounted for a 65.88% Win Probability

3. [2016] PERDiff (Passing Efficiency Rating Difference): Accounts for 63.90% Win Probability
3. [2013,2014,2015] PERDiff: Accounted for a 63.99% Win Probability

I have more than 36 Trends that I forecast, with 22 of them accounting for greater than 50% Win Probability. But, since I posted the top 3, I'll explain what they are so I'm not asked the question at a later date.

TSRS

When facing an opponent, each stat category is compared against its opposite. If the opponent is stronger in the comparison, and the team facing them does not deviate from its measured potential, the team will increase slightly in its strength index. TSRS calculates strength of offensive, defensive, and special teams groupings, rates and assigns strength of schedule, and uses covariance between these groups to measure how much these areas change over time. I named it True Statistical Rating of Strength because it handles inflation/deflation values that come with raw statistics. All statistics, regardless of type are converted into a base floating point number. Standard Deviation is used to measure dispersion between teams and lower SD helps determine consistency and quality within each stat grouping. To date, just using TSRS alone produces a 70% or better win probability on any college game.

RERDiff

I created my own Rushing Efficiency Rating formula because it is highly likely you will not find one anywhere on the web, much less one that is accurate. In college, rushing efficiency is slightly more important than passing efficiency. The two are very close but unlike the NFL, rushing is very important in college football. Each offense has a rushing efficiency rating and each defense has a rushing efficiency rating. The greater the RER on offense, the stronger the team is at running the ball. The lower the RER on defense, the stronger the team is defensively at limiting the run. Subtract the two and you have an RERDiff. A team that has a very large RERDiff has a decided advantage over a team that does not.

PERDiff

Pass Efficiency Rating Difference is exact in detail when compared to RERDiff above. It has it's own formula, but the mechanics are pretty much the same as RERDiff so no need to explain this further.



SUMMARY


This year I've built out special AI scenarios for ML and ATS in College and NFL football. I have four unique AI programs that handle scenario driven data.

ML AI for College:

The ML AI for College which I call "Victor Argentum" goes through four base scenarios, and each scenario has up to 15 logic based criteria.

For instance, the four base start with: Power 5 vs. Power 5, Power 5 vs. Group of 5 or IA, Group of 5 or IA vs. Power 5 and finally, Group of 5 or IA vs Group of 5 or IA. Notre Dame is the only exclusion of the group and are considered Power 5 even though they are an Independent. If my system predicts a group of 5 team to beat a power 5 team, it goes through that part of the scenario and is matched to any of the criteria. If it doesn't have greater than a 78% probability after the process is completed, the game is tossed out. Each game fits its own scenario.

ATS is a lot more complicated and I'm pretty excited about all of the scenarios I have in place this year. I won't go into detail here.

--------

I hope that answers all of your questions.
“Know where to find information and how to use it; that is the secret of success.” – Albert Einstein
bazooooka
bazooooka
Joined: Nov 21, 2016
  • Threads: 3
  • Posts: 182
May 9th, 2017 at 12:43:15 PM permalink
JD,

I've read that 6 point favorites win 70% or more of the time in college football. Does you system add value by winning ML bets above 60% on games where the spread is between -5 and +5? And how does your system find edge on ATS bets?

Have you found your system predicted winners, especially if the books have them as a slight dog, or covering at 60%? And have you back-tested your algo on past seasons to see if the relationship has held to past regimes and increased data?

It would be impressive if you found over hundreds of system picks that your algo can cover more often than not if when books are +1 to +3 on a team yet you have it capped as a -3 situation. What other scenarios are you able to find edge?

  • Jump to: