CFB faves...Do they cover ATS when ahead (but not already covering) at HT?

I came across something interesting while researching CFB games this year.  This system will help you determine whether or not a favorite will cover the game spread based on the lead at halftime relative to the spread and a middle.

Criteria

1.  Eliminate ALL games where the favorite has already covered the game spread by halftime.

2.  Eliminate ALL games where the score is tied or the underdog is ahead at halftime.

3.  Consider only games where there is a middle of 6 or more (see my blog on calculating middles)

If team A is favored by X points and is leading at the half four scenarios are possible:
  a.  Team is covering the X point spread at the half.  Do not consider this scenario.
  b.  Team is covering by about half the spread (2H line will most likely be close to the difference) presenting little or no middle.
  c.  Team is covering by less than half of the original line (2H line will most likely be half of the original spread, presenting a good middle)
  d.  Team is covering by more than half of the original line (2H line will most likely present a good middle in the opposite direction)

Examples:

Week 7: NMEX playing BYU.  The line is BYU -23.  BYU leads at the half by 4 points (7 to 3).  Clearly 4 is much less than half of 23.  The 2H line is BYU -10 presenting a middle of 9.  Qualifying play.  We would then expect BYU to fail to cover the game spread.  BYU wins the 2H 14-0, and the game 21-3, but fail to cover the 23 point spread.

Week 1: NTSU playing KSST.  The line is KSST -29.  KSST leads at the half by 28 points (28 to 0).  Clearly 28 is much more than half of 29.  The 2H line is KSST -13.5 presenting a middle of 12.5.  Qualifying play.  We would then expect KSST to cover the game spread.  KSST wins the 2H 17-6, and the game 45-6, thus covering the 29 point spread.

Observations:

 

In most of the college football games, either the underdog is tied or leading at the half or the favorite is covering at the half.  

 

For games with a middle of 6 or more, when a favorite is covering below half of the original spread, the chances that the favorite will ultimately cover the original spread is low.  Conversely, when that favorite is covering above half of the original spread, the chances that the favorite will ultimately cover is greater.

 

 

Results

 

In all games up through week 8 of the 2008 college football season, the following results were obtained using the 4 criteria listed above.  33 games fit the criteria above.

 

Favorites covering less than half of the original spread at halftime ultimately fail to cover the game spread 95% of the time.  Those favorites are 1-18 ATS.

Favorites covering more than half of the original spread at halftime ultimately cover the game spread 92% of the time. Those favorites are 12-1-1 ATS.

 

However, the difficulty in knowing this is that we cannot go back in time and place or change our bets on the game spread after the game has already started.  However, this serves up an excellent opportunity for those involved in “live betting” to get in on the action, provided the site contains game lines that do not change.  

 

The key is identifying these “live” games and waiting until the right moment to pull the trigger should an opportunity arise.

 

Does anyone have any ideas on how they would apply this to 2H betting without live betting to gain an edge when one of the games falls into this scenario?  Would it involve game spread and 1^st half betting scenarios?  Thoughts? 

 

Madden... thanks for the comment.

For what it's worth, I am not looking at middles to choose a second half bet.  With the way the oddsmakers make the 2H lines, this would be too easy to figure out.  It would take digging into those 33 games to figure out at what point along the spectrum (between 1 point  and the game spread) that the halftime (effective spread) falls. 

Porky's system fails for small middles, because that gap is so small, it would be almost equally likely for the middle to miss to one side or the other.


Visualize with a prototype number line we all learned in math....

<-(DOG LEADS)-><-----------(FAVE LEADS)---------------->

               0         ATS/2         ATS
<--------------x---------@@@@@----------x-------------->
        A      B     C     D      E     F       G



At halftime, the game will fall into one of seven scenarios. 
  (A) - dog leads at the half
  (B) - score is tied (favorite leads by 0)
  (C) - favorite covering by less than half of game spread
  (D) - favorite covering by about half of game spread
  (E) - favorite covering by more than half of game spread
  (F) - favorite covering game spread exactly (denoted by "ATS")
  (G) - favorite covering by more than the original game spread

In some cases where the original game spread is on the 1/2 point, there is no (F).

It doesn't seem to matter where in the upper half (E) or lower half (C) of that spectrum, it is clearly divided into these area.  So there is no subrange within (E) or (C) that gives us a better probability than the other.  The "gray" area (D) may be as low as three-points or as high as six.

In your scenario, it sounds like you are wanting to know how large the middle (between F and someplace in G) would have to be in order to secure an ATS win.  This is why I simplified things by choosing a 6-point middle.  Anything less, and the system starts to fail.  With that said, there needs to tired winning percentage because virtually every game falling in (C), the favorite has failed to cover ATS for game and every game falling in (E), the favorite has covered ATS for game. 

Only one time in (C) has a favorite covered, and one time in (E) has a favorite failed to cover.  

Now we are getting into "Spreads Analysis 401"!! 

After I plug in the results for week 9, then I will post an update.

I just love this.... no spellchecker. 

The last paragraph should read...

With that said, there doesn't need to be a "tiered" winning percentage because virtually every game falling in (C), the favorite has failed to cover ATS for game and every game falling in (E), the favorite has covered ATS for game.

What this boils down to is is there a way to bet both 1H and full game spreads and hedge with a 2H bet some kind of way to minimize loses and maximize gains.... along with the hope that the halftime line will produce a middle of 6 of more.  Otherwise everything goes to pot!

Any more ideas? 

Madden,

In your previous post, you say since C rarely is greater than F and E rarely is bellow F...

You missed it,  C and E are ranges of pointspreads and are always less than F.

Here's an example,

Let's use the NMEX/BYU example.  Look at our number line.


<--(NM LEADS)-><------------(BYU LEADS)---------------->

               0          -11'         -23
<-----U--------x---V-----@Wx@@--X-----Y-x------Z------->
        A      B     C     D      E     F       G


I have denoted points on my line (V, W, X, Y, Z) where a game could fall at the half.

U may be where NMEX leads by 7 (not relevant to this post)
V may be where BYU leads by 4 (relevant - 2H line may be 10 - as in my example, we have a middle)
W may be where BYU leads by 10 (relevant - 2H line may be 12 - middle too small)
X may be where BYU leads by 17 (relavant - 2H line may be 12 - we have a middle)
Y may be where BYU leads by 21 (relavant - 2H line may be 7 - we have a middle)
Z may be where BYU leads by 30 (not relevant)


Also, I tend to disagree that the middle only hits 25% of the time, because you have to take into account the size of that middle.  If your middle is 0.5, you will never land in between.

I believe I understand your question now.  Let's see if I can get you those results.

I just compiled the WK 9 scores/lines and found 10 games landed somewhere between B and F.  Turns out all of the effective ATS values (points) land somewhere in that shaded D range, causing the middles to be under 6, so the results above will not change for another week. 

Here are my stats:

All results are excluding the situation where the favorite is covering those games at the half and when dogs are tied or leading..

For middles of 0.0 – 11 games – favorites cover 77% (7-2-1); region C (2-2), region E (4-1), midpoint (0-1)
For middles of 0.5 - 11 games - favorites cover 30% (3-7-1); region C (2-2), region E (1-5-1)
For middles of 1.0 - 14 games - favorites cover 36% (5-9); region C (0-2), region E (5-7)
For middles of 1.5 - 8 games - favorites cover 37% (3-5); region C (1-2), region E (2-3)
For middles of 2.0 – 8 games – favorites cover 62% (5-3); region C (1-1), region E (5-1)

For middles of 2.5 – 12 games – favorites cover 54% (6-2-1); region C (4-1), region E (2-4-1)

For middles of 3.0 – 7 games – favorites cover 17% (1-5-1); region C (1-2), region E (0-3-1)

For middles of 3.5 – 6 games – favorites cover 50% (3-3); region C (0-3), region E (3-0)

For middles of 4.0 – 11 games – favorites cover 63% (7-4); region C (1-2), region E (5-3)

For middles of 4.5 – 7 games – favorites cover 43% (3-4); region C (1-2), region E (2-2)

For middles of 5.0 – 3 games – favorites cover 66% (2-1); region C (0-1), region E (2-0)

For middles of 5.5 – 8 games – favorites cover 37% (3-5); region C (2-2), region E (1-3)

 

I do not see a trend.  I have segregated these into region C or region E plays (still no trend).  Because the middle is so small, one TD or FG one way or the other can tip the middle in the opposite direction.  I see no value to this data. 

 

It becomes increasingly difficult to determine which region (C or E) a game lands on when the middle is zero, and this directly relates to that gray area D.  The smaller the middle, the thinner that gray area is, and more likely the plays become toss-ups.  If you can use them to help “simplify” things, be my guest.  The plays that are relevant as set forth by criterion #3 clearly define the C and E regions as area D widens.

Hope you can make sense out of this. 

Week 10 saw two games fit this criteria - one in the low (C) range, the other in the high (E) range.

C-range game:Arizona St./Oregon St.
E-range game: Washington/USC

C-game explanation:
Oregon St (-14) at home led 7-6 at the half.  Oregon St was only -1 at the half.  The 2H line was Oregon St -7.  The middle was 6.  Qualifying play in the lower range (1/14 = 7th percentile).  Our play on team is the underdog Arizona St.  We would expect Oregon St to not cover the game spread of -14.  Oregon St wins the 2H by only 1 (20-19) and thus fails to cover ATS.  Also, we notice Oregon St fails to cover the 2H ATS. 

E-game explanation:
USC (-46) at home led 42-0 at the half.  USC was -42 at the half.  The 2H line was USC -17.  The middle was 13.  Qualifying play in the upper range (42/46 = 91st percentile).  Our play on team is the favorite USC.  We would expect USC to win the 2H by at least 4 to cover the game spread of -46.  USC wins the 2H by only 14 (14-0) and thus covers ATS.  However, USC fails to cover the 2H ATS. 


With these two games added in:

Favorites covering by less than half of the original spread (lower 50th percentile) at halftime ultimately fail to cover the game spread 95% of the time.  Those favorites are now 1-19 ATS.

Favorites covering by more than half of the original spread (upper 50th percentile) at halftime ultimately cover the game spread 93% of the time. Those favorites are 13-1-1 ATS.

I have crunched more numbers within this subset to come up with a range of games that can apply to 2H betting.  Stay tuned. 

Yo Kangy, thats some gangster shittt.... keep it up and good luck

More analysis of the games fitting this system

In all 20 games that fall in the lower 50th percentile of the spread range (the C-range games), all of the favorites are at home and are 1-19 ATS.  Using Porky's system, all of the "play-on" teams would be road dogs as they would provide the best opportunity to miss the middle.  The road dogs cover the 2H spread 60% (12-8) of the time, which means the non "play-on" home favorites also fail to cover the 2H spread 60% of the time.  Keep in mind, as noted above, that the non-"play-on" favorites (all home favorites thus far) fail to cover the game spread 95% of the time.  When the home favorites do cover the 2H spread, it usually is not enough to overcome the full game spread.


I broke the 20 C-range games thus far into groups based on original game ATS lines:

If spread is < 20, "play-on" dogs are 6-0 2H ATS
If spread in 20's, "play-on" dogs are 4-7 2H ATS
If spread is > 30, "play-on" dogs are 2-1 2H ATS


What does this mean? 
Games where the favorite is leading by less than half of the game spread, the "play-on" team, dogs of less than 20 or greater than 30, are a combined 8-1 (89%) YTD.  


Of the 15 games that fall in the upper 50th percentile of the spread range (E-range games), these "play-on" teams, always favorites,
are 13-1-1 ATS, but only 9-4-2 (69%) ATS 2H line.  Of these games, "play-on" home favorites are 9-0-1 ATS, but only 5-3-2 2H ATS.  The "play-on" road favorites are both 4-1 (80%) ATS and 4-1 (80%) ATS 2H.  


I broke the 15 E-range games thus far into groups based on original game ATS lines:

 If spread is < 20, "play-on" favorites are 0-0-1 2H ATS
 If spread in 20's, "play-on" favorites are 7-2-1 2H ATS
 If spread is > 30, "play-on" favorites are 2-2 2H ATS
   

What does this mean? 
Games where the favorite is leading by more than half of the game spread, the "play-on" team, favorites with lines in the 20s is 7-2-1 (77%) YTD. 



Summarizing the C and E range system plays,
If a game lands in region C (low end) and the middle is 6+, bet the "play-on" dog on 2H line if original spread is either less than 20 or more than 30. 8-1 YTD

If a game lands in region E (hi end) and the middle is 6+, bet the "play-on" favorite on the 2H line if the original spread is between 20 and 29.5. 7-2 YTD

Combined these two system plays are 15-3 (83%) YTD


USC was not an E-range bet because the original line was -46, outside the 20-29.5 range.  Arizona St. WAS a C-range bet because the original line was -14, less than 20. 
Arizona St covered the 2H line of +7.


I will try to combine both systems into one post when posting the plays.  Also you will note that one system may say to play one team, the other may say play the opposite or no play.  And both systems may go against NJPorky's original 2H system.  A lot of times, the systems will agree.  You will have to use your best judgment.

Thanks .   Bookies beware!  It's time to get our money back!!

GANGSTER KANGY  --->    <--- BOOKIE  

Update after Week 10:

Oklahoma was an "E-range" bet because the original line was OU -27, inside the 20-29.5 range and they were leading by 24 at the half.  The 2H line was OU -10.5, presenting a middle of 7.5.  We would expect OU to cover the remaining 3 points for game.  Since the original game line is in the 20's, this presents a play.  OU was the 2H 28-14 and not only covered the game spread, but covered the 2H as well.   
Oklahoma covered the 2H line of -10.5

Florida Atlantic was an "E-range" bet because the original line was FAU -22.5, inside the 20-29.5 range and they were leading by 19 at the half.  The 2H line was FAU -10, presenting a middle of 6.5.  We would expect OU to cover the remaining 2.5 points for game.  Since the original game line is in the 20's, this presents a play.  FAU won the 2H 21-7 and not only covered the game spread, but also covered the 2H.   
Florida Atlantic covered the 2H line of -10


This brings the E-range record to 9-2 YTD and 17-3 (85%) combined.

Hey Lippsman, Florida did not make the system because they were already covering the game spread by halftime.  Florida game line was -23, but were up by 35-0 score, thus violating the first rule:

1.  Eliminate ALL games where the favorite has already covered the game spread by halftime.

The system is designed for favorites leading at the half by between 1 and the game spread minus 1/2.

No problem man! 

we have a play tonite kangy? 
We SHOULD have hit last night!
that sucked!