Prediction PAFC Punters' Thread

Remove this Banner Ad

Sleezy

Premiership Player
Mar 17, 2015
3,365
6,646
Melbourne
AFL Club
Port Adelaide
Good luck.

earning money in punting takes a lot of dedication and most commonly a pretty strong understanding of maths. Selectivity and self-control are majors stumblings blocks for most people. Long term winners are in the fraction of a percentage of the population. It's best not to put forecast values on it rather try and improve each month and I can garantee it will be rocky early. I'll admit my knowledge of sports betting compared to racing is minimal and I rarely bet on AFL but try to be selective. From afar I would suggest it would be extremely hard to have a positive ROI betting 9 head to head games a week.
Thanks for your comments.

I most definitely recognise the likelihood of success is low. I've stated my goals because i need something to work towards. Whether they are realistic or not is something i will learn as i go. I've done my research, now it's time to test my theory woth cold hard cash.

I've also committed to posting my results here because it will drive me to see it through. Putting my thought on paper for others will help me with losong weeks i'm sure. It will also help with critical analysis from others outside the bubble.

As for the particulars of my strategy, i know no-one bets like this, and i think i can see why. It is high risk. It relies on accuracy instead of value. A week of 6 will more likely than not lose me a little money. The question is, will the 7s, 8s & 9s make up for the 4s, 5s and 6s?

Do i risk not flat betting each game, only betting games i see value in? I don't have the system or the bank to do so. Over the next couple of years though i do plan vastly improving the sophistication of my tipping techniques. Can only do it with data though, and you gotta start somewhere.
 

Log in to remove this ad.

Power Trip

Debutant
Oct 7, 2008
107
79
Adelaide
AFL Club
Port Adelaide
You will go broke before the mid year byes

You would need to tip 7+ winners every single week to even make your money back

Going on the current odds if you backed all the favs and the 7 lowest odds ones got up you will lose $0.96 for the week

Sorry but this system will not work


Sent from my iPhone using Tapatalk
 

Sleezy

Premiership Player
Mar 17, 2015
3,365
6,646
Melbourne
AFL Club
Port Adelaide
I've done some more detailed analysis of the odss I can expect to be betting into, and my own tipping habits and biases.

Some interesting outputs:
  • There were three teams I tipped at less than 50% accuracy last year: Richmond (i didn't think they were any good), Melbourne (Also thought were terrible) & Port (blind faith)
  • There was a pretty even spread of high, mid & low table teams across the spectrum of tipping accuracy, it clearly showed my biases for or against certain teams however
  • I underrated the importance of home ground advantage for similarly matched teams in the first half of the year. I corrected this after the bye and improved my tipping by about 1 tip per round
  • Tipping pre-bye was volatile, wild swings from 5s to 8s and back to 5s, with higher odds. Tipping post bye was stable, weeks of 7s in a row at lower odds.
  • There is a negative correlation between high average odds rounds and tipping accuracy, which works in your favour to limit losses but works against you by limiting winnings as well
  • Working on averages, flat betting makes you money with a tipping accuracy of 6.5ish. Average odds per round do not take into account outliers (ie, bulldogs defeating sydney or essendon toppling hawthorn). Remove the outliers and the average odds reduce significantly - making 7+ tips per round a requirement for profit.
  • Tipping 4 averaged a 30% loss, Tipping 5 averaged a 10% profit (only 1 round of 5 for me last year, a bye round), rounds of 6 a 7% loss, rounds of 7 a 3% gain and rounds of 8 a 25% gain.
  • Gimme wins (odds of 1.10 or under) are harmful to the bottom line as they do not contribute enough to overcome the money lost on poor tips
Some lessons for me from this:
  • My tipping is not accurate enough to make a profit flat betting head to head matches
  • My biases will cost me big time
  • The weakness of both Carlton & Essendon will mean that 80% of rounds will have at least 2 games with odds less than 1.10, making success even harder
  • One poor week will wipe out a month of gains
  • Betting during the bye rounds is risky business - one adverse result can cause a loss
  • accuracy early in the season is vital, the odds area better as the pecking order is yet to be established
Changes to the strategy:
  • Development of a system to increase tipping accuracy and reduce impact of biases (more on this to come - this has been a bit of fun)
  • Track two systems across the season. One selective based on predicted probability versus available odds, the other flat betting as previously highlighted
  • Use of this season as a data collection and model validation exercise. Therefore, I will not be chasing down sign-up bonuses as I don't want to waste that until I have a system that is at least somewhat reliable
  • Goals for the year revised to a profit of any kind and a partially validated predictive model
 
So what does the Essendon situation do to odds this year.

A mate of mine who is a senior economist about 7 years ago did a study into footy betting and he found that Essendon and Collingwood have shorter odds to win than they should because of the sheer volume of money because of their supporter base, compared to similar games with another team having the same sort of chance of winning as either Essendon or Collingwood had.

Will Essendon supporters in big numbers, bet against their own side to lose and try and collect something? Will they be stupid enough to back their own team? Will they stop betting? Will they move their funds to another game? Or to another sport?
 
Jan 30, 2013
16,166
16,559
AFL Club
Port Adelaide
I've done some more detailed analysis of the odss I can expect to be betting into, and my own tipping habits and biases.

Some interesting outputs:
  • There were three teams I tipped at less than 50% accuracy last year: Richmond (i didn't think they were any good), Melbourne (Also thought were terrible) & Port (blind faith)
  • There was a pretty even spread of high, mid & low table teams across the spectrum of tipping accuracy, it clearly showed my biases for or against certain teams however
  • I underrated the importance of home ground advantage for similarly matched teams in the first half of the year. I corrected this after the bye and improved my tipping by about 1 tip per round
  • Tipping pre-bye was volatile, wild swings from 5s to 8s and back to 5s, with higher odds. Tipping post bye was stable, weeks of 7s in a row at lower odds.
  • There is a negative correlation between high average odds rounds and tipping accuracy, which works in your favour to limit losses but works against you by limiting winnings as well
  • Working on averages, flat betting makes you money with a tipping accuracy of 6.5ish. Average odds per round do not take into account outliers (ie, bulldogs defeating sydney or essendon toppling hawthorn). Remove the outliers and the average odds reduce significantly - making 7+ tips per round a requirement for profit.
  • Tipping 4 averaged a 30% loss, Tipping 5 averaged a 10% profit (only 1 round of 5 for me last year, a bye round), rounds of 6 a 7% loss, rounds of 7 a 3% gain and rounds of 8 a 25% gain.
  • Gimme wins (odds of 1.10 or under) are harmful to the bottom line as they do not contribute enough to overcome the money lost on poor tips
Some lessons for me from this:
  • My tipping is not accurate enough to make a profit flat betting head to head matches
  • My biases will cost me big time
  • The weakness of both Carlton & Essendon will mean that 80% of rounds will have at least 2 games with odds less than 1.10, making success even harder
  • One poor week will wipe out a month of gains
  • Betting during the bye rounds is risky business - one adverse result can cause a loss
  • accuracy early in the season is vital, the odds area better as the pecking order is yet to be established
Changes to the strategy:
  • Development of a system to increase tipping accuracy and reduce impact of biases (more on this to come - this has been a bit of fun)
  • Track two systems across the season. One selective based on predicted probability versus available odds, the other flat betting as previously highlighted
  • Use of this season as a data collection and model validation exercise. Therefore, I will not be chasing down sign-up bonuses as I don't want to waste that until I have a system that is at least somewhat reliable
  • Goals for the year revised to a profit of any kind and a partially validated predictive model

it's quite hard sticking to a single bookmaker for a given week and make a profit in the long term. more accounts = more ability to find best odds which can reduce expected loss from about 6% to 3-4%. the more selective you are the lower the expected loss and in some cases an expected positive return... or even a no risk scenario.
 

Sleezy

Premiership Player
Mar 17, 2015
3,365
6,646
Melbourne
AFL Club
Port Adelaide
I've done some more detailed analysis of the odss I can expect to be betting into, and my own tipping habits and biases.

Some interesting outputs:
  • There were three teams I tipped at less than 50% accuracy last year: Richmond (i didn't think they were any good), Melbourne (Also thought were terrible) & Port (blind faith)
  • There was a pretty even spread of high, mid & low table teams across the spectrum of tipping accuracy, it clearly showed my biases for or against certain teams however
  • I underrated the importance of home ground advantage for similarly matched teams in the first half of the year. I corrected this after the bye and improved my tipping by about 1 tip per round
  • Tipping pre-bye was volatile, wild swings from 5s to 8s and back to 5s, with higher odds. Tipping post bye was stable, weeks of 7s in a row at lower odds.
  • There is a negative correlation between high average odds rounds and tipping accuracy, which works in your favour to limit losses but works against you by limiting winnings as well
  • Working on averages, flat betting makes you money with a tipping accuracy of 6.5ish. Average odds per round do not take into account outliers (ie, bulldogs defeating sydney or essendon toppling hawthorn). Remove the outliers and the average odds reduce significantly - making 7+ tips per round a requirement for profit.
  • Tipping 4 averaged a 30% loss, Tipping 5 averaged a 10% profit (only 1 round of 5 for me last year, a bye round), rounds of 6 a 7% loss, rounds of 7 a 3% gain and rounds of 8 a 25% gain.
  • Gimme wins (odds of 1.10 or under) are harmful to the bottom line as they do not contribute enough to overcome the money lost on poor tips
Some lessons for me from this:
  • My tipping is not accurate enough to make a profit flat betting head to head matches
  • My biases will cost me big time
  • The weakness of both Carlton & Essendon will mean that 80% of rounds will have at least 2 games with odds less than 1.10, making success even harder
  • One poor week will wipe out a month of gains
  • Betting during the bye rounds is risky business - one adverse result can cause a loss
  • accuracy early in the season is vital, the odds area better as the pecking order is yet to be established
Changes to the strategy:
  • Development of a system to increase tipping accuracy and reduce impact of biases (more on this to come - this has been a bit of fun)
  • Track two systems across the season. One selective based on predicted probability versus available odds, the other flat betting as previously highlighted
  • Use of this season as a data collection and model validation exercise. Therefore, I will not be chasing down sign-up bonuses as I don't want to waste that until I have a system that is at least somewhat reliable
  • Goals for the year revised to a profit of any kind and a partially validated predictive model
So I mentioned in my previous post that I was putting together a predictive model to help with my little betting project. Well V1 of the model has been completed and been put through it's paces in round 1 of the NAB challenge.

How the model works (not too much detail):
  • Splits each team into fwds, mids backs
  • Each player gets an offensive and defensive rating out of ten, based on a comparison table I put together of representative players
  • The rating represents their average performance. The best offensive player (probably Danger or J Kennedy) does not get a 10 and the worst offensive player (probably Zac Dawson or Steven Morris)
  • There is also some inherent variability in performance integrated into the model
  • Based on the individual players in each line, a team gets 6 scores for fwds, backs & mids offence & defence
  • This is then compared to the scores of the opposition and a percentage likelihood of either team calculated
Model Performance so far:
  • 7 from 7 so far (not that it has been particularly difficult to pick)
  • Weaknesses where I don't know the depth of a teams list that well, particularly NSW & QlD teams and s**t teams (cartlon, melbourne), players just get a standard rating based on a guess
  • This weakness should be less prevalent during the season proper when best 22s are played, and I will be getting better and better knowledge of the teams
  • Bias for teams with quantity over quality. The Eagles/Crows game was an interesting case study - they had a line-up with about 10 defenders in the 26 where as adelaide really only had 6 forwards. Consequently, the 10 average defenders scored higher than the 6 really crows forwards, which included two high class options in tex and betts.
  • This weakness will also be less prevalent in the season proper due to the reduced bench size. I will have to consider how to include the 'no match-up' factor in however (ie no kpd v a kennedy or hawkins)
  • At this stage there does not seem to be much correlation between % likelihood of winning and score - so no line bets at this stage. Early days though.
All in all, pretty happy with the performance so far. We'll see if it can keep it up in the season proper.
 

Sleezy

Premiership Player
Mar 17, 2015
3,365
6,646
Melbourne
AFL Club
Port Adelaide
So I mentioned in my previous post that I was putting together a predictive model to help with my little betting project. Well V1 of the model has been completed and been put through it's paces in round 1 of the NAB challenge.

How the model works (not too much detail):
  • Splits each team into fwds, mids backs
  • Each player gets an offensive and defensive rating out of ten, based on a comparison table I put together of representative players
  • The rating represents their average performance. The best offensive player (probably Danger or J Kennedy) does not get a 10 and the worst offensive player (probably Zac Dawson or Steven Morris)
  • There is also some inherent variability in performance integrated into the model
  • Based on the individual players in each line, a team gets 6 scores for fwds, backs & mids offence & defence
  • This is then compared to the scores of the opposition and a percentage likelihood of either team calculated
Model Performance so far:
  • 7 from 7 so far (not that it has been particularly difficult to pick)
  • Weaknesses where I don't know the depth of a teams list that well, particularly NSW & QlD teams and s**t teams (cartlon, melbourne), players just get a standard rating based on a guess
  • This weakness should be less prevalent during the season proper when best 22s are played, and I will be getting better and better knowledge of the teams
  • Bias for teams with quantity over quality. The Eagles/Crows game was an interesting case study - they had a line-up with about 10 defenders in the 26 where as adelaide really only had 6 forwards. Consequently, the 10 average defenders scored higher than the 6 really crows forwards, which included two high class options in tex and betts.
  • This weakness will also be less prevalent in the season proper due to the reduced bench size. I will have to consider how to include the 'no match-up' factor in however (ie no kpd v a kennedy or hawkins)
  • At this stage there does not seem to be much correlation between % likelihood of winning and score - so no line bets at this stage. Early days though.
All in all, pretty happy with the performance so far. We'll see if it can keep it up in the season proper.

The representative player rating table - should inspire some interesting debate

upload_2016-2-25_18-50-11.png
 

(Log in to remove this ad.)

Sportsbet currently running a special.
  • Place 25 dollar bet on a team to win the premiership
  • Recieve a 5 dollar bonus bet for every win that team has up to round 12.
Port are $11 to win the premiership, so I put 25 down on them. Plus we have a good early draw, so should score some good free bets. :)
 
Mar 17, 2014
3,615
4,429
AFL Club
Port Adelaide
Sportsbet currently running a special.
  • Place 25 dollar bet on a team to win the premiership
  • Recieve a 5 dollar bonus bet for every win that team has up to round 12.
Port are $11 to win the premiership, so I put 25 down on them. Plus we have a good early draw, so should score some good free bets. :)
Thanks for the tip. That reminds me, I had an email from UBet that had something about a special offer of $2 for Port to make the 8. Better investigate that one!

Edit: The offer was for entering the locker room competition.
 
Last edited:
Sportsbet currently running a special.
  • Place 25 dollar bet on a team to win the premiership
  • Recieve a 5 dollar bonus bet for every win that team has up to round 12.
Port are $11 to win the premiership, so I put 25 down on them. Plus we have a good early draw, so should score some good free bets. :)
William Hill got a similar one. They've had some cracker specials these last few days.
 
Feb 13, 2013
4,151
7,185
Alberton
AFL Club
Port Adelaide
Sportsbet currently running a special.
  • Place 25 dollar bet on a team to win the premiership
  • Recieve a 5 dollar bonus bet for every win that team has up to round 12.
Port are $11 to win the premiership, so I put 25 down on them. Plus we have a good early draw, so should score some good free bets. :)

Can't find that in the app :(


Sent from my iPhone using Tapatalk
 

BIG BILLY F

All Australian
Jan 15, 2016
630
528
AFL Club
Port Adelaide
I have been on big losing streak lately....

I am finding I am getting up on some stage on my weekends but giving it all back and then some....

But on my holiday at Singapore I was 3500 down on the casino there my misses was pissed...

Long story short on my 2nd last day I had all the shrapnel of coins that I had no idea what to do with.....so I take it down the casino...it turns out it was 14 dollars worth of coins...

I proceed to to take that 14 dollars and turn it into 5300in 12 hours of intense gambling......my misses was still pissed off though
 
I take all credit
Met the turnover requirement pretty quickly and made a modest profit on top of what I'd made from the Robbie Gray bet.

Have had a lot of luck with live sports betting. Correctly guessed the final score of a soccer match (Azerbaijan Premier league, no less o_O) paying $9.50.

Edit: Moved the quote in here, instead of continuing on in last week's game-day thread.
 

Peck

Todd Marshall fan
10k Posts
Jul 30, 2009
10,286
6,792
Melbourne, VIC
AFL Club
Port Adelaide
Other Teams
Left is Best
I have been on big losing streak lately....

I am finding I am getting up on some stage on my weekends but giving it all back and then some....

But on my holiday at Singapore I was 3500 down on the casino there my misses was pissed...

Long story short on my 2nd last day I had all the shrapnel of coins that I had no idea what to do with.....so I take it down the casino...it turns out it was 14 dollars worth of coins...

I proceed to to take that 14 dollars and turn it into 5300in 12 hours of intense gambling......my misses was still pissed off though
No offense mate but I would stay the **** out of casinos if I were you.
 

Remove this Banner Ad

Back