# Grasping the Concept of value betting

Pages: [1]
 Author Topic: Grasping the Concept of value betting  (Read 933 times)
mikey
Newbie

Karma: 0
Posts: 17

« on: October 11, 2017, 11:54:22 PM »

Hello Everyone,

I've been doing quite a bit of research over the past little while about value betting. It's taken some time to get a decent grasp of the concept and the probabilities, managing bankroll, etc.

I've just begun paper trading to see if I'm doing the calculations correctly. I was hoping to get some feedback on whether or not I've got the basics.

Here is an NHL ice hockey value bet from the other day:

Edmonton Oilers vs. Winnipeg Jets

Pinnacle odds on Edmonton Oilers (3 hours before game time): 1.636, or 100/1.636 = 61.1% 'true' probability
As a hockey fan, I would have actually given Edmonton a higher probability to win the game

Marathonbet: 1.94, or 100/1.94 = 51.5% implied probability

value = (0.611 x 1.94) - 1
v = 1.185 - 1
v = 0.18534, or 18.5% value (+EV)

As for managing bankroll and/or staking amounts - using fractional Kelly method for, lets say 10%

K = [(P x odds) - 1] / (odds - 1)
K = [1.18534 - 1] / (1.94 - 1)
K = 0.18534 / 0.94
K = 0.174, or 17.4% (bankroll at 100% Kelly)
at 10% Kelly, 0.174 x 0.1 = 0.0174, or 1.74% of bankroll
i.e. if bankroll was \$10,000, I'd be betting  (10,000 x 0.0174) \$174 on Edmonton Oilers to win

While Edmonton did end up losing this game 5 - 2, it was still a good bet to take because there was value.
Making 1000+ or 10,000+ similar bets should yield a positive result.

I hope I got it.

Thanks for the input,

Mike

 Logged
Transat2
Newbie

Karma: -1
Posts: 10

« Reply #1 on: October 12, 2017, 05:19:19 AM »

Hi,

You need to take in consideration Pinnacle margin. Let's say it's 3%.

The true probability of Edmonton winning is then (61.1 * 100)/103 = 59,34%  (without taking Favorite Longshot bias into consideration).

You can do it faster using https://www.pinnacle.com/en/betting-resources/betting-tools/margin-calculator. Just write odds of each game in the boxes.

For calculating value it is good you have just to take the new true probability of 59,34% in your formula.

Same thing with Kelly Criterion you're calculating it the good way, just put the new probability of Edmonton. Also, there is a minor error cause 0.18534/0.94 = 0.1971 or 19.71%. It is indeed the good way of calculating the Kelly Criterion.

Hope it helps.

 Logged
maletaja
Probably a Pro

Karma: -25
Posts: 470

« Reply #2 on: October 12, 2017, 08:44:48 AM »

I think 100 bets with 1.7-2.3  odds are enough for valuebetting . For my perspective i havnt got any major downswings, if u bet some 5% edge
I have heard that suggestion for new bookie were 200k free cash balance!!!
 Logged
Chipmunk
Newbie

Karma: 2
Posts: 38

« Reply #3 on: October 12, 2017, 10:00:49 AM »

I think 100 bets with 1.7-2.3  odds are enough for valuebetting . For my perspective i havnt got any major downswings, if u bet some 5% edge
I have heard that suggestion for new bookie were 200k free cash balance!!!

With 100 you can have drastic swings. Wait till you have 1000+.
 Logged
gamblehappier
Has some experience

Karma: -2
Posts: 157

Gambling is the key in order to survive.

« Reply #4 on: October 13, 2017, 02:32:27 AM »

Chimpuck is right.I believe with 5/100 value you need at least 300 bets,maybe more.If you have 10/100 value i think with150 bets you are ok.
 Logged

Believe to yourself is the secret to success.
Chipmunk
Newbie

Karma: 2
Posts: 38

« Reply #5 on: October 13, 2017, 02:58:05 PM »

I've got a spreadsheet sim for this. With half-Kelly and mean 5% edge, you have 37% chance of losing money over 100 bets. If you are in the neg there it likely won't be too bad though, the chance of losing more than 33% of your roll by that stage is 1.7%.

At 250 bets, those figures go to 30% and 4.7% (yes the latter increases initially!). At 500, 22% and 6%, at 1000 14% and 5%, at 2000 6% and 3%.

Note at full-Kelly, the volatility is _much_ worse. If you absolutely must be more aggressive, use 0.6 or max 0.66 Kelly. There's almost no point going over that, the increased returns are minimal (going from 0.66 to full Kelly gives you +11% bankroll growth, but triple the chance of a -67% drawdown).
 « Last Edit: October 13, 2017, 03:06:54 PM by Chipmunk » Logged
maletaja
Probably a Pro

Karma: -25
Posts: 470

« Reply #6 on: October 13, 2017, 06:08:59 PM »

I dont know what u talk about. Im making 100 bets per day and i have max 20bets drawback

But u can bet 10k i roulette!
If u talk about kelly bookmaker needs 1billion free cash to take 100k roulette bets!!! Its nonsense
 « Last Edit: October 13, 2017, 06:10:30 PM by maletaja » Logged
mikey
Newbie

Karma: 0
Posts: 17

« Reply #7 on: October 20, 2017, 03:13:36 AM »

Thanks for the replies everyone.

If starting with a bankroll of 5k, I understand that a better needs to turnover his bankroll quite frequently

I've read online that some suggest 200 bets per week! This seems like a lot to me. Is 200 recommended? Can someone do 100/ week?
 Logged
Alfa1234
Pro

Karma: 19
Posts: 646

« Reply #8 on: October 20, 2017, 08:01:06 AM »

I struggle to see where Maletaja finds 200 valuebets every week in bigger leagues and big events...so yes you can start with fewer events but the fewer events, the more swings you'll have.
 Logged
 Pages: [1]