# Closing Line \ EV

Pages: [1]
 Author Topic: Closing Line \ EV  (Read 794 times)
Shawn2k
Newbie

Karma: 0
Posts: 11

« on: August 23, 2017, 09:06:58 AM »

Hi,

I know there is some, probably many posts close to this topic, but it's the one most important factor in valuebetting, so I'm taking it up once again.

First of all some facts - Correct me if I'm wrong: Closing line, more specifically Pinnacle's closing line is what most valuebetters look at, to make sure they have positive EV. Over a big volume, if you beat Pinnacle's closing line (including the juice), we should have a positive EV. This should at least work in the bigger leagues and markets.

Let's take an example:

We make a bunch of bets, lets say 1000, where the odds is 4,20. Pinnacle's closing line on all the bets are 4,00, with a RTP (return to player) of 97,5%.

This makes the true closing line close to: 4,00/0,975 = 4,102.

So.. if we've got an average of 4,20 on these 1000 bets, we should have 4,20 / 4,102 = 2,39% positive EV on each bet.

Is all of this correct?

A couple of other questions \ things to discuss:

1) If all of these thousand bets were delivered one hour before the game, and we didn't know the closing line, but Pinnacle's odds at the moment were 4,0 (And we make bets on 4,20) - Should we expect lower, the same or better odds compared to the closing line? From a logical point of view this should flatten out over time, but I have no stats on this.

2) Is there differences in value on Pinnacle closing lines, if the odds is high or low? What I mean is: Burnley to beat Chelsea gives us 12,50 at Pinnacle. Liverpool to beat Chelsea gives us 4,0. Both has 97,5% RTP. Will still 12,50 / 0,975 = 12,82 give us the correct odds, or do we need more if the odds is high? Variance will obviously be the devil if we use odds like 12,50, but is still i.e 13,25 value in this spot?

 « Last Edit: August 23, 2017, 09:28:23 AM by Shawn2k » Logged
barbero
Getting better

Karma: 16
Posts: 239

« Reply #1 on: August 23, 2017, 11:49:53 AM »

2) Usually more commission might be applied on the leg with higher odds.
 Logged
Malteser
Newbie

Karma: 2
Posts: 24

« Reply #2 on: August 23, 2017, 05:27:43 PM »

1) On average I would expect the odds one hour away to be same as the closing odds.
2) It is different: extracting true odds from Pinnacle odds by dividing by overround is not a good way. It is better to use something called Shin probabilities - search on Google Scholar.

Good questions by the way. Shows you have some understanding.
 Logged
Shawn2k
Newbie

Karma: 0
Posts: 11

« Reply #3 on: August 24, 2017, 01:13:35 PM »

1) On average I would expect the odds one hour away to be same as the closing odds.
2) It is different: extracting true odds from Pinnacle odds by dividing by overround is not a good way. It is better to use something called Shin probabilities - search on Google Scholar.

Good questions by the way. Shows you have some understanding.

Have read a little bit about Shin probablities, but I haven't found a way to calculate it. I could always assume what the margins are on a odd around 5 at a match with 2% total margin, but would have been nice to get a deeper understanding of what we can expect the margin to be on each outcome, depending at the size of the odds - Because it's obviously different for low and high odds - And most of all: It's vital if we want to be sure of making a valuebet.

If someone just could say something like "When Pinnacle have margins around 2%, we can assume the margin for odd between 1,20-1,50 are around 1,35%, odd between 3,0-4,0 are around 3% etcetc"..... that would have been nice

 Logged
Malteser
Newbie

Karma: 2
Posts: 24

« Reply #4 on: August 25, 2017, 05:32:53 PM »

Erik Strumbelj's papers in 2014 and 2016 will be helpful, and focus on the relevant part of Shin's work.

Bookies often distribute the overround in an ad-hoc fashion, where they manually decide what odds map to which probability ranges. But it is closer to the Shin approach rather than proportional distribution.
 Logged
bm1125
Newbie

Karma: -2
Posts: 7

« Reply #5 on: August 28, 2017, 11:53:23 AM »

I am not really aware of any way to determine the bookie margin of each outcome. As Malteser said, the margins are higher for higher odds. Have a look at this table here and you can see that the lower the probability (higher odds) the more costly the bookies mistakes will be . It is more difficult to estimate the probability of a rare event and thus bookies will hedge their risk by taking much higher commissions of those rare events.
 Logged
maletaja
Probably a Pro

Karma: -25
Posts: 470

« Reply #6 on: August 28, 2017, 04:31:15 PM »

That explains why Federer opponents have shitty odds like 1.01-12
 Logged
 Pages: [1]