Calculating expected win %

[Igetp2s]

Suppose Team A and Team B both played 100 games against exactly the same competition. Team A won 70% of the games they played, and Team B won 40% of the games. Now Team A plays Team B on a neutral site. How would you calculate Team A's expected win %?

I was thinking it would be 70 / (70+40) = Approx. 64%, but that seems wrong. If in those 1st 100 games the competition had exactly a 50% winning % against other teams, and Team A was able to win 70%, they should do even better against Team B, not worse.

So what would be the correct way to calculate this?

Thanks for the help.

[Dunder]

----------

[Meestermike]

I use a formula I call Win % edge.

Ex.
Team A Win % = .672
Team B Win % = .489

Add both = 1.161

Divide Team A win % .672 by 1.161 = .579 ~ Win % Edge

Team B is .579 better than Team B

The following chart, may prove to be very helpful. It involves win percentages and the points from a spread. First you determine which team has the edge, and by what percentage for each game. Then locate the suitable percentage on the chart and see how many points it is worth.

Percentage Edge		Point Value
.524			½
.545			1
.565			1½
.583			2
.600			2½
.615			3
.630			3½
.643			4
.655			4½
.667			5
.677			5½
.688			6
.697			6½
.706			7
.714			7½
.722			8
.730			8½
.737			9
.744			9½
.750			10
.756			10½

Use the table, which shows percentages converted into Money Lines to begin searching for live wagers.

WIN % EDGE		REQUIRED MONEY
HOME OR ROAD 		LINE
.524				-110
.545				-120
.565				-130
.583				-140
.600				-150
.615				-160
.630				-170
.643				-180
.655				-190
.667				-200
.677				-210
.688				-220
.697				-230
.706				-240
.714				-250
.722				-260
.730				-270
.737				-280
.744				-290
.750				-300
.756				-310
.762				-320
.773				-330
.783				-360
.792				-380
.800				-400

Look for discrepancies between what the chart tells you the Money Line should be and what the line is that the bookmakers are quoting. Look for a difference of at least 20 cents when a game is handicapped nearly even and 30 to 40 cents as it approaches .667 or 200. As a game that is handicapped approaches .750 or 300, the difference should be a difference of at least 40 cents but closer to 50 cents.

Charts and reasoning taken from a paperback book entitled "GOING FOR IT! THE WINNER'S GUIDE TO SUCCESSFUL SPORTS BETTING" by author Larry Humber

[Formulawiz]

I use a formula I call Win % edge.

Ex.
Team A Win % = .672
Team B Win % = .489

Add both = 1.161

Divide Team A win % .672 by 1.161 = .579 ~ Win % Edge

Team B is .579 better than Team B

The following chart, may prove to be very helpful. It involves win percentages and the points from a spread. First you determine which team has the edge, and by what percentage for each game. Then locate the suitable percentage on the chart and see how many points it is worth.

Percentage Edge        Point Value
.524            ½
.545            1
.565            1½
.583            2
.600            2½
.615            3
.630            3½
.643            4
.655            4½
.667            5
.677            5½
.688            6
.697            6½
.706            7
.714            7½
.722            8
.730            8½
.737            9
.744            9½
.750            10
.756            10½

Use the table, which shows percentages converted into Money Lines to begin searching for live wagers.

WIN % EDGE        REQUIRED MONEY
HOME OR ROAD         LINE
.524                -110
.545                -120
.565                -130
.583                -140
.600                -150
.615                -160
.630                -170
.643                -180
.655                -190
.667                -200
.677                -210
.688                -220
.697                -230
.706                -240
.714                -250
.722                -260
.730                -270
.737                -280
.744                -290
.750                -300
.756                -310
.762                -320
.773                -330
.783                -360
.792                -380
.800                -400

Look for discrepancies between what the chart tells you the Money Line should be and what the line is that the bookmakers are quoting. Look for a difference of at least 20 cents when a game is handicapped nearly even and 30 to 40 cents as it approaches .667 or 200. As a game that is handicapped approaches .750 or 300, the difference should be a difference of at least 40 cents but closer to 50 cents.

Charts and reasoning taken from a paperback book entitled "GOING FOR IT! THE WINNER'S GUIDE TO SUCCESSFUL SPORTS BETTING" by author Larry Humber

I believe the author has the win% to line conversion incorrect. As an ex: 54.5% should calculate into a -3.5 pt fav and not -1 pt fav. Big difference.

[Sinister Cat]

So what would be the correct way to calculate this?

Thanks for the help.

This may be what you are looking for:

http://forum.sbrforum.com/handicapper-think-tank/122312-home-field-advantage.html#post1253669

Note that this method also allows you to easily factor in a home field advantage.

[20Four7]

I believe the author has the win% to line conversion incorrect. As an ex: 54.5% should calculate into a -3.5 pt fav and not -1 pt fav. Big difference.

The value of a 1/2 point and the value of different points will depend on the sport being played. All 1/2 points are not created equal.

[Igetp2s]

Thanks for the replies guys, but I don't think I really got what I was looking for. The Ganchrow thread that was linked to was beyond my understanding of statistics. I was hoping for a simple formula with as little symbols as possible.

I'd like to not focus on the conversion to spread aspect, just want to get a straight win % if possible. Also, don't want to focus on home/road adjustments, so in my example I used a neutral court.

Perhaps the formula from my original example would be (70/(70+40)) * (.5/.4), where .5 is the opposition's win % in the 1st 100 games, and .4 is Team B's win %?

Does that make sense?

Formulawiz, I think your formula is wrong. Why would team A's win % go down if Team B is worse than average?

[Sinister Cat]

Perhaps the formula from my original example would be (70/(70+40)) * (.5/.4), where .5 is the opposition's win % in the 1st 100 games, and .4 is Team B's win %?

Does that make sense?

That seems to give an answer that's close.

Here's the formula from that thread I posted:
(H-H*A)*C/((H-H*A)*C+(A-H*A)*(1-C)

So it's (.7-.7*.4)*.5/((.7-.7*.4)*.5+(.4-.7*.4)*.5 = .21 / (.21 + .06) = .777

The .5 is what you can use for home field advantage; so .5 would be neutral.

[sharpcat]

good threaf

[Formulawiz]

Thanks for the replies guys, but I don't think I really got what I was looking for. The Ganchrow thread that was linked to was beyond my understanding of statistics. I was hoping for a simple formula with as little symbols as possible.

I'd like to not focus on the conversion to spread aspect, just want to get a straight win % if possible. Also, don't want to focus on home/road adjustments, so in my example I used a neutral court.

Perhaps the formula from my original example would be (70/(70+40)) * (.5/.4), where .5 is the opposition's win % in the 1st 100 games, and .4 is Team B's win %?

Does that make sense?

Formulawiz, I think your formula is wrong. Why would team A's win % go down if Team B is worse than average?

I am not saying that. All I am saying is the conversion from expected win % to a line is incorrect. In the example I previuously gave the line is 2.5 points off.

[u21c3f6]

I don't think you can realistically compare the two this simply. However, If I did have to compare the two this simply, I think your original calculations are a close approximation without many more comparisons. The 50% does not apply after you have data. Team A apparently played teams that were only 30% successful against it. Now they are facing Team B that had a 40% success rate. Therefore in this simple comparison, Team B has actually performed better than the teams that Team A has already faced and therefore Team A's chances should be less than 70% (as you calculated) and not more than 70%. Again, I don’t think this is a good way to compare.

Joe.

[Meestermike]

I believe the author has the win% to line conversion incorrect. As an ex: 54.5% should calculate into a -3.5 pt fav and not -1 pt fav. Big difference.

Formula I believe the author was attempting to provide a 2 to 2.5 point edge advantage so you would make a play if the line was actually -1 or lower like pick-em or +1 you would have that much more of an edge

[Formulawiz]

Formula I believe the author was attempting to provide a 2 to 2.5 point edge advantage so you would make a play if the line was actually -1 or lower like pick-em or +1 you would have that much more of an edge

Makes sense then

[Igetp2s]

I don't think you can realistically compare the two this simply. However, If I did have to compare the two this simply, I think your original calculations are a close approximation without many more comparisons. The 50% does not apply after you have data. Team A apparently played teams that were only 30% successful against it. Now they are facing Team B that had a 40% success rate. Therefore in this simple comparison, Team B has actually performed better than the teams that Team A has already faced and therefore Team A's chances should be less than 70% (as you calculated) and not more than 70%. Again, I don’t think this is a good way to compare.

Joe.

I don't think that makes sense. The teams A has faced have a 30% win percentage against team A, but they have a 50% win percentage in all other games amongst themselves (I am assuming this), and a 60% win percentage against team B. So Team B is worse than Team A's and Team B's historical competition, not better.

[Igetp2s]

That seems to give an answer that's close.

Here's the formula from that thread I posted:
(H-H*A)*C/((H-H*A)*C+(A-H*A)*(1-C)

So it's (.7-.7*.4)*.5/((.7-.7*.4)*.5+(.4-.7*.4)*.5 = .21 / (.21 + .06) = .777

The .5 is what you can use for home field advantage; so .5 would be neutral.

Thanks Sinister. I am still a little confused about your formula.

How does (.7-.7*.4)*.5 for example translate into .21? Wouldn't that be -.105? I think there might be a sign wrong somewhere.

[Sinister Cat]

Thanks Sinister. I am still a little confused about your formula.

How does (.7-.7*.4)*.5 for example translate into .21? Wouldn't that be -.105? I think there might be a sign wrong somewhere.

Sorry, close bracket at the end of the formula was missing, should be
(H-H*A)*C/((H-H*A)*C+(A-H*A)*(1-C))

But yeah, (.7 - .7 * .4) * .5 is .21, not sure how you got -.105

[Igetp2s]

Sorry, close bracket at the end of the formula was missing, should be
(H-H*A)*C/((H-H*A)*C+(A-H*A)*(1-C))

But yeah, (.7 - .7 * .4) * .5 is .21, not sure how you got -.105

The way you wrote it,

Step 1 is .7 (the second one) * .4 = .28
Step 2 is .7 (1st one) -.28 = -.21
Step 3 is -.21 * .5 = -.105

How are you getting to .21?

[bleedblue]

For step2:

.7 - .28 = .42

what you did was .07 - .28 = -.21

[Igetp2s]

For step2:

.7 - .28 = .42

what you did was .07 - .28 = -.21

Duh. You'll have to believe I'm not really as stupid as that.