Correlated Plays, does anyone know how to work out the edge?

Hi,

I would like to know if anyone could tell me how to work out an estimated edge/advantage from same game correlated 'parlay/multiplier' bet in American Football?

Please note: example 1 is a hypothetical 'would never happen(unless the bookmaker was stupid)' example and example 2 is a 'real bet' example .

Also note: all games are $1.90 lines (so about -111 American)


Example 1(would never happen in real life) : a game of football has the dog @ +34.5*** and the total @ 34.5***,
then the edge on the Dog and Under would be would be +80.5% / the Favorite and Over would be the same.

*** Please note: The edge would always be 80.5% if the total and line were the same as each other.


Example 2 (Oregon vs Oregon State^^^):

Oregon State @ +26.5 and Under 62.5 and

Oregon -26.5 and Over 62.5

both bets for $100 @ $3.61 (+ 261), so total bet outlay of $200 for a potential profit of $161.


^^^ : Because I am betting and live in Australia, the home team is first on the ticket, according to
www.espn.com/chalk/liveOdds
it would be this game(from an American Perspective):

211 Oregon State
212 Oregon (Date: 11/25)

Update (About 620 PM, Queensland Australia Time, Date 11/21)

Maximum Edge working out (see example 1 above post)

The way I get the 'maximum edge', is to:

1. Divide the line by the total, 34.5 / 34.5, which is 100%

2. Divide the figure in '1' by 4 , which is 25%

3. Then I add 25%*** to the amount in figure '2', which is 50%

4. Multiply figure '3' by the odds, so 50%. x 3.61 = 180.5%, RTP

5. Minus '4' by 100% to get the player edge, which is about 80.5% player edge

6. No 'potential edge error' adjustment necessarry for the above working out


Estimated Edge (working out)

The way I get a 'ballpark estimate', is to first use a ':

1. Divide the line by the total, 26.5 / 62.5, which is about 42.4%

2. Divide the figure in '1' by 4 , which is 10.6%

3. Then I add 25%*** by the figure in '2', which is 35.6%

4. Multiply figure in '3' by the odds, so 35.6%. x 3.61 = 128.5...%, RTP (theoretical)

5. Minus '4' by 100% to get the player edge, which is about 28.5...% player edge (theoretical)

6. Then i make a 'potential edge error' adjustment: by dividing figure in '5' by 4,
which is about 7.12...%


***: 25% is used as the baseline chance, as it is assumed that each individual leg would have an independent (non-correlated) chance of 50%,
therefore 2 legs would be 50% x 50% = 25%.


Also, I can not stress enough that this could be a potentially very flawed 'ballpark estimate',

Lastly, if anyone with more advanced knowledge in maths or sports betting than me could help 'fix' or improve the way of estimating the chances of correlated plays, then I would greatly appreciate your input

-------------------
Update after RS's post below:

The above 'theoretical advantage', would only be applicable for working out the Dog and Under, (my favorite and over reasoning is flawed)

With parlay #1, if the favorite covers then it MUST go over. If the dog covers it MIGHT go over or it might go under (ie: dog wins 7-0, it's under. Dog wins 28-21, it goes over). I'm not seeing how you'd come up with an 80.5% advantage, though, unless you were assuming a dog and under had 100% correlation...?

Btw, I have two correlated futures bets for this season. Both had 100% correlation (okay, maybe 99.999%), but were also very big long shots. One already lost but the other is still a possibility.

Quote: RS

With parlay #1, if the favorite covers then it MUST go over. If the dog covers it MIGHT go over or it might go under (ie: dog wins 7-0, it's under. Dog wins 28-21, it goes over). I'm not seeing how you'd come up with an 80.5% advantage, though, unless you were assuming a dog and under had 100% correlation...?

Btw, I have two correlated futures bets for this season. Both had 100% correlation (okay, maybe 99.999%), but were also very big long shots. One already lost but the other is still a possibility.

Thanks for your post,

you are correct for the Favorite and Over (silly mistake on my part),

I still think I am right for the Dog and Under (you would get the 80.5% edge for the Dog and Under in my original post's first example)
the Favorite and Over's edge in the same example would not be measurable using the same 'working out' (but would be less than the maximum of 80.5%, though I still doubt it would be a negative edge).

Interesting side note: To guarantee a profit in the first example, you would take the Dog and the Under (both 34.5) @$ 3.61 and bet the total Over 34.5 @ $1.90 to lock in a profit of about 24.48...% on investment (would never happen in real life though)

Quote: RS

With parlay #1, if the favorite covers then it MUST go over.

Which means you're getting +260 on a 50% chance the favorite covers.

Likewise, there is a 50% chance the game goes under, in which case the dog/under parlay hits.

So if each individual bet -- dog, favorite, over, under -- is a 50% chance, that means the dog/over is the only way to lose as favorite/under is impossible. So that makes is it a 35% edge.

So that means a side / total ratio of 100% is a 35% edge. Roughly a side / total ratio of 33% has a 0% edge when bet blindly at +260. We could graph historical data to find what the edge is when the side / total ratio is 42%. If it is linear, it would be a 4% edge. I would be very surprised if it didn't end up looking like a logarithmic graph. Which would make the edge higher.

Quote: RS

Btw, I have two correlated futures bets for this season. Both had 100% correlation (okay, maybe 99.999%), but were also very big long shots. One already lost but the other is still a possibility.

What book lets you parlay futures?

Quote: RS

With parlay #1, if the favorite covers then it MUST go over. If the dog covers it MIGHT go over or it might go under (ie: dog wins 7-0, it's under. Dog wins 28-21, it goes over). I'm not seeing how you'd come up with an 80.5% advantage, though, unless you were assuming a dog and under had 100% correlation...?

Btw, I have two correlated futures bets for this season. Both had 100% correlation (okay, maybe 99.999%), but were also very big long shots. One already lost but the other is still a possibility.

Go browns!