balmainforever
Dibs
- Sep 4, 2003
- 25,736
- 30,596
- AFL Club
- Richmond
- Other Teams
- Balmain, GreenBay, Edmonton, Celtic
- Moderator
- #3,502
Last edited:
-
AFLW 2024 - Round 9 - Indigenous Round - Chat, game threads, injury lists, team lineups and more.
AFL 2021 Brownlow Medal (any results after no posting is automatic card)
- Thread starter TheJanuaryMan
- Start date
- Tagged users None
- Status
offthebusgus
Debutant
- Oct 9, 2019
- 85
- 198
- AFL Club
- Carlton
what is a discord?
- Oct 6, 2011
- 35,540
- 10,942
- AFL Club
- GWS
- Other Teams
- Patriots, Golden State, Wildcats
All I know is I got booted out of a few in my times.
what is a discord?
Analysis Paralysis
Senior List
- Jul 25, 2018
- 187
- 315
- AFL Club
- Brisbane Lions
He deleted you, not the discord.Hey all!
Just a heads up for next year, LucasParker, salaterra and whichever other alternate account he is using…
This guy made a discord a month or so ago, invited a few of us who shared our counts throughout the year on it. We shared our counts, team votes, halfway leaders etc. and he didn’t share a single piece of information and dodged the question whenever asked. Constantly refused to show any sense of accountability and ignored anything in relation to him sharing his own information. He has since deleted the discord.
Lesson learnt for myself next year, if someone is creating a discord, make sure they share their own info first and don’t leech/mooch off of others.
Just a heads up for next year
Been a hugely fun ride throughout the year, but this was a disappointing way to end it.
So I deleted any counts/analysis and left (not that they are worth anything anymore).
Low.
LionsOrDieTrion
All Australian
- Jun 23, 2020
- 802
- 1,127
- AFL Club
- Brisbane Lions
We will create one for 2022 soon, must do own count, full transparency. First sign of BS, gaawwwwnn.
Tigers04
Senior List
- Jul 10, 2012
- 264
- 208
- AFL Club
- Richmond
I jumped on this thread from about round 13-14 onwards for the top 5/10/20 posting value bets each week. Some were good and some turned out average. Had a lot of fun this year, was a shame with macrae and Kelly underpolling but overall still walked away with a nice profit, should probably buy the mrs something nice as I didn’t pay her much attention throughout the year when the footy was on. Think I had around 150 bets on for the night, some got up and some didn’t but keen for next year and will post my own count starting round 1.
Bring on round 1 already!
Bring on round 1 already!
- Oct 6, 2011
- 35,540
- 10,942
- AFL Club
- GWS
- Other Teams
- Patriots, Golden State, Wildcats
Just went through a bit more analysis of the '3 Vote Markets' in 2021.
Someone was good enough to private message me about the Round 7 odds for North Melbourne v Melbourne, so that meant I had the full dataset.
Sportsbet in the 196 matches with outright favourite for the '3 votes', 133 of them won. (67.857% on the night). I have excluded the 2 tied outright odds on Merrett/Stringer @ $2.25 in Rd 18, and Taranto/Josh Kelly @ $1.90 in Rd 22. Sportsbet odds for '3 Vote' favourite outright ranged from $1.005 to $2.75.
Compared it against my model, and I admit I am not good at tipping outright favourites compared to them. In 198 matches, my model with the outright favourite for the '3 Votes' , 127 of them won the 3 Votes. (64.141% on the night). In my models I had no outright tied odds. My odds for '3 Vote' favourite outright ranged from $1.02 to $4.01.
Sportsbet Bookmaker margins for '3 Vote odds':
Mean: 25.9054%
Maximum: 63.93%
Minimum: 10.40%
Median: 25.9044%
Sportsbet '3 Vote odds' on favourite (196 matches excluding the 2 matches with equal favourites):
Mean: $1.42
Median: $1.315
All information to keep on mind for next year.
As a side not, I had a bit more time to go through the 2nd half of the season, and a surprise that popped up for me was neither Max King or Dom Sheed polling the 3 in Rd 19 , it was the $21 Tim Kelly taking full votes.
Someone was good enough to private message me about the Round 7 odds for North Melbourne v Melbourne, so that meant I had the full dataset.
Sportsbet in the 196 matches with outright favourite for the '3 votes', 133 of them won. (67.857% on the night). I have excluded the 2 tied outright odds on Merrett/Stringer @ $2.25 in Rd 18, and Taranto/Josh Kelly @ $1.90 in Rd 22. Sportsbet odds for '3 Vote' favourite outright ranged from $1.005 to $2.75.
Compared it against my model, and I admit I am not good at tipping outright favourites compared to them. In 198 matches, my model with the outright favourite for the '3 Votes' , 127 of them won the 3 Votes. (64.141% on the night). In my models I had no outright tied odds. My odds for '3 Vote' favourite outright ranged from $1.02 to $4.01.
Sportsbet Bookmaker margins for '3 Vote odds':
Mean: 25.9054%
Maximum: 63.93%
Minimum: 10.40%
Median: 25.9044%
Sportsbet '3 Vote odds' on favourite (196 matches excluding the 2 matches with equal favourites):
Mean: $1.42
Median: $1.315
All information to keep on mind for next year.
As a side not, I had a bit more time to go through the 2nd half of the season, and a surprise that popped up for me was neither Max King or Dom Sheed polling the 3 in Rd 19 , it was the $21 Tim Kelly taking full votes.
- Moderator
- #3,807
Pathetic.
- May 6, 2009
- 10,619
- 25,573
- AFL Club
- St Kilda
Wouldn't you know it- bang that was another O'gorman umpired match.
That's at least 3 noticeably surprising results from this year alone where he umpired.
I said it before the count and I'll say it again, anyone with the time and energy to build a model which accounts for actual umpires (even at a rudimentary level, ie- places a red flag 'do not bet' against known weird vote givers) will do very well for themselves
Last edited:
Prochard123
Norm Smith Medallist
- Jul 28, 2014
- 7,147
- 5,773
- AFL Club
- Collingwood
Definitely should keep track a list of umpires that issue dodgy votes.
I'm sure there are a few troll umpires that doesn't give a flying f about Brownlow
I'm sure there are a few troll umpires that doesn't give a flying f about Brownlow
invisible_
Debutant
- Oct 16, 2020
- 100
- 290
- AFL Club
- Adelaide
Had some time to do some final analysis on the count, in particular to see how my model performed relative to other online counts (ESPN, BetFair, Before you bet, Sporting News, Daily Sports Buddy and our very own TheJanuaryMan).
The metric i'm using to measure error is mean square error, where the error is calculated for every single player in every single game, for example:
Count predicts 2 and player gets 3, then error is 1 and squared error is 1.
Count predicts 0.5 and player gets 2, error is 1.5 and squared error is 2.25.
Count predicts 0 and player gets 0 (there should be A LOT of these), then error is 0.
To measure the accuracy of the model/count in a single number, the average squared error from every player across every game is taken.
The results (lower number the better, indicating lower average error)
BF: 0.127
Before you bet: 0.1456
Daily sports buddy: 0.139
Sporting news: 0.168
TJM: 0.1742
ESPN: 0.1294
My model: 0.0973
Must note, it's so much harder to achieve a low error margin when giving exact 3,2,1s as done by TJM, SN and DSB.
Pretty excited by this, can see room for a few minor tweaks in the offseason but otherwise it's just plug and play next years data and we should be on again for another profitable year.
Have attached the csv file showing votes as provided by each of the above 7 counts if anyone is interested in doing their own analysis.
Edit: attached second csv file which only contains players who were expected to poll (by anyone, including my model >0.05 votes) or did actually poll which should be a bit more readable if you just want a brief look
The metric i'm using to measure error is mean square error, where the error is calculated for every single player in every single game, for example:
Count predicts 2 and player gets 3, then error is 1 and squared error is 1.
Count predicts 0.5 and player gets 2, error is 1.5 and squared error is 2.25.
Count predicts 0 and player gets 0 (there should be A LOT of these), then error is 0.
To measure the accuracy of the model/count in a single number, the average squared error from every player across every game is taken.
The results (lower number the better, indicating lower average error)
BF: 0.127
Before you bet: 0.1456
Daily sports buddy: 0.139
Sporting news: 0.168
TJM: 0.1742
ESPN: 0.1294
My model: 0.0973
Must note, it's so much harder to achieve a low error margin when giving exact 3,2,1s as done by TJM, SN and DSB.
Pretty excited by this, can see room for a few minor tweaks in the offseason but otherwise it's just plug and play next years data and we should be on again for another profitable year.
Have attached the csv file showing votes as provided by each of the above 7 counts if anyone is interested in doing their own analysis.
Edit: attached second csv file which only contains players who were expected to poll (by anyone, including my model >0.05 votes) or did actually poll which should be a bit more readable if you just want a brief look
Last edited:
- Oct 6, 2011
- 35,540
- 10,942
- AFL Club
- GWS
- Other Teams
- Patriots, Golden State, Wildcats
Hate to say this as well, one games votes doesn't affect the outcome of a whole count. But these 'micro' bets on these games can be fixed and obviously have a huge affect on unders/overs and all.
In one case with one betting company, You could technically do a multi , for instance on Tim Kelly to Poll 3 Votes in Round 19 (roughly $21) and Wines to Win Brownlow (roughly $3.50)
21 x 3.50= 73.50 multi
I'm not sure umpires are 'clean' in all of this..........there are a lot of ways all of this could be done under cover and no one in the footy world would ever notice.
In one case with one betting company, You could technically do a multi , for instance on Tim Kelly to Poll 3 Votes in Round 19 (roughly $21) and Wines to Win Brownlow (roughly $3.50)
21 x 3.50= 73.50 multi
I'm not sure umpires are 'clean' in all of this..........there are a lot of ways all of this could be done under cover and no one in the footy world would ever notice.
- Oct 6, 2011
- 35,540
- 10,942
- AFL Club
- GWS
- Other Teams
- Patriots, Golden State, Wildcats
I'm also happy to share my spreadsheet for all the matches I compiled this year for 2021. Alot of the odds are based on the 25% bookmaker margin (as that is what you usually get with Sportsbet) and I usually gather 5-14 players who are potentially in the running for the votes and then to see who can poll the 3. The odds are based on an opinion, statistics and to give what I think is 'fair' odds or 'low odds'. (Sometimes I would refer to some more reputable outlets as a bit of an ice breaker) I probably wouldn't release something like this pre-count for next year though. Not sure if anyone else has a system like this in place.......but would be interested to find ways at improving. (Think I saw INVISIBLE spreadsheet, and it looks like a similar thing to what I have done)
So long as the margins or percentages add up to 125% or so, is what you will get with Sportsbet and the guys offering '3 vote markets'.
Sometimes I would go through a 2nd review and do a weighted average- which I think is a good thing. (but in some cases bad as my gut had Zorko as BOG in Rd 23, but later changed to Rich )
- Thread starter
- #3,814
Gift
Norm Smith Medallist
Nah I dont believe it coz I literally lost every 3 vote bet that I made and they barely let me on for anything still in the later rounds. People will always weigh up risk v reward when doing dodgy stuff. No sane person would risk their job/being exposed in the news by placing bets which are capped at like $100 on these type bets.
Its more likely the umps are tired after 3hrs of umpiring and just wanna go home so they will sometimes pick someone silly because not that much though as been put in.
The volume sportsbet would allow on these markets wouldnt be enough for an umpire to risk their job and integrity. Wouldnt worry about any deliberately dodgy shit. They are incompetent though, we see that every week with their decisions...
Gift
Norm Smith Medallist
A place where blokes trade some personal favours for Brownlow info
Dont listen to those that complain about giving out info, they got stuff in return. Cant ask for a refund if the guy does a bad job
Check out the stts for employees stealing in the workplace. People are literally risking their jobs for more self gain. If I was an umpire I would find a way to bet on this. They aren’t paid very much at the end of the day.
Billionaire Panda
Senior List
- Jul 13, 2021
- 175
- 219
- AFL Club
- St Kilda
Damn I can't believe I missed out on betting this game, had it as a raffle between Kelly King and Sheed. TK was definitely best on at 3qtr time and had quite a few wow moments with Nic Nat giving it to him on a platter as he broke away from a stoppage.
One thing for me to consider and I'm not sure if you guys have experienced this, but games where my Saints lost were my most accurate games for predicting votes. I tend to watch my team so hard and only notice an opposition player when they are actually having a huge impact without bias. Eg Dusty not polling v Saints with Bolton getting 2.
Analysis Paralysis
Senior List
- Jul 25, 2018
- 187
- 315
- AFL Club
- Brisbane Lions
Didn't notice that from you, but your Gawn analysis was on the money in the end! Nice job.
Billionaire Panda
Senior List
- Jul 13, 2021
- 175
- 219
- AFL Club
- St Kilda
Yeah Gawn single handedly made it an okay night after Steele disappointed. Was anchored into a lot of multi's. Loves pulling random 2 vote games does Max.
Anyone backing Josh Kelly was asking for trouble, the eye test I feel really showed he had an average season , doesn't really stand out in my opinion... Macrae is so stiff though, I feel I got stooged on Oliver rd 5 leader 7.00 tab also took Dusty at 8.00 and he missed out on votes vs Stkilda, thought Oliver could definately win the rd 5 leader outright and overpolled by a vote or two in the early games to have it on track, then missed out completely vs Geelong when all I needed was a vote to win it outright, think he was slated for 1.5 votes on betfair or possibly 2 behind petracca, good ole MaXY Gawn snagged the 1 here... also had my 10 legger ruined by the duncan thing but was only a $20 bet at odds of 2.50 lol~!!
Sorry, I did not create the discord to share spreadsheets, I apologise if you felt used, I said this in the discord. I created it for a place we can discuss that is offline, to share tips, and share notifications for when markets drop.
Rob Ogurman, also umped Macraes 40 touches 1 goal, 0 vote game, flag!!
Hi all,
Hope everyone went well on Sunday night. Personally I made 5x my outlay which was nice, helped mainly by a 63/1 multi of top 5, 10 and 20 place getters boosted by Boak top 10 $3.20, Guthrie top 20 $1.64 and Mills top 20 $2.63.
Would’ve loved Oliver to miss out on the 3 in the final round as I had Wines/Bont/Walsh tri in order and boxed but that’s the way the cookie crumbles at times.
Ive done my own spreadsheet and count for a few years now but in nowhere near the detail I find on some counts here.
Is anyone willing to share a blank spreadsheet with some more detail? I’m not great with computers but have figured out how to make an automatic leaderboard as I enter in round votes but that’s as far as I can get.
TIA
Hope everyone went well on Sunday night. Personally I made 5x my outlay which was nice, helped mainly by a 63/1 multi of top 5, 10 and 20 place getters boosted by Boak top 10 $3.20, Guthrie top 20 $1.64 and Mills top 20 $2.63.
Would’ve loved Oliver to miss out on the 3 in the final round as I had Wines/Bont/Walsh tri in order and boxed but that’s the way the cookie crumbles at times.
Ive done my own spreadsheet and count for a few years now but in nowhere near the detail I find on some counts here.
Is anyone willing to share a blank spreadsheet with some more detail? I’m not great with computers but have figured out how to make an automatic leaderboard as I enter in round votes but that’s as far as I can get.
TIA
- Status
Similar threads
- Sticky
