Football & Sports Betting Thread 2019/2020
- Thread starter Levo
- Start date
- Status
| Salernitana v Crotone, Crotone -0.25 @ 2.10 |
| Elche v Las Palmas, Elche -0.25 @ 2.31 |
| Großaspach v Jena, Jena 0.25 @ 1.99 |
| Man United v Everton, Everton 0.75 @ 2.03 |
Bet the first three last night and all have shortened quite a bit already. Small stake on Everton this morning after they drifted out overnight. Schneiderlin is out and Digne questionable (but probably in) but I still like them at this price. They are much better than their league position and I feel United's counter-attacking style is much more effective against the bigger sides.
I also wanted to bet Kingstonian (7th division) +1 at home to Woking (5th) in the FA Trophy. Hosts kicked off +0.75 against Fylde in the FA Cup a fortnight ago who are also in 5th tier but much better than Woking. The visitors confirmed they will go as strong as possible but so did Fylde and Kingstonian more than held their own and lost 0-2 (fair result at least 1-2). Their league form is also very good at the moment (3-2-0, 10-2 goals in their last five). Limits are still very low and Pinnacle are not up yet and odds have dropped a little over night. Will monitor it until kick-off.
Last edited:
Early start acca for today:
Motherwell/Rangers >1.5, FFC Turbine Potsdam W, Partizan Belgrade, Desna/Dynamo Kiev >1.5, Botev Plovdiv, PSG W, Cotonsport, Atletico Madrid B, UD Logrones, Covadonga, Piast Gliwice or draw @21.94
Thinking about it, the Desna v Kiev game could easily end up 0-0.
Motherwell/Rangers >1.5, FFC Turbine Potsdam W, Partizan Belgrade, Desna/Dynamo Kiev >1.5, Botev Plovdiv, PSG W, Cotonsport, Atletico Madrid B, UD Logrones, Covadonga, Piast Gliwice or draw @21.94
Thinking about it, the Desna v Kiev game could easily end up 0-0.
Small loss.
Out = 10 units.
5 bets = no return - Feralpisalo, Bordeaux, Vise, Bamaseda, Monopoli
1 bet money back - Betis (1.00)
4 wins - Osnabruck (2.025), Montegiorgio (2.10), Sivasspor (2.05), Carpi (1.85)
Total: 9.025
-0.975
Gonna go for a draw tonight, Palace v Brighton. Backed us for the draw yesterday but annoyingly in an acca with Wolves draw!
Kingstonian won 3-1. Staked the straight win at close to 5/1 in the end as the handicaps didn't appeal.
I'm on Plymouth tomorrow in their FA cup replay at home to Bristol Rovers. Rovers coach Graham Coughlan announced that he's considering his future after their league game at the weekend citing personal reasons. He's on the verge of joining Mansfield Town who are 18th in League Two (Bristol are 4th in League One!). To say this was unexpected would be an understatement. Local press has described it as an "tornado that has been sent through the club." None of this is official yet and Coughlan might even be in charge of the game tomorrow but he has already told his players he's leaving and this is obviously not the kind of preparation you want. There'll be another press conference this afternoon which should clear things up.
First leg (1-1) was a pretty even affair which Plymouth slightly edged. They also missed a 96th minute penalty. Both teams are fighting for promotion in League One and Two (Plymouth are in 8th) respectively but that doesn't always tell the full story and in reality Rovers are a mid-table League One side at best and Plymouth one of the best teams in League Two. There isn't much between the two sides and with the managerial mayhem at Bristol I have to side with the hosts. They are 2.30+ general quote.
Post from Rovers forum that gives a few more details and captures the mood at the club today:
Last edited:
I didn't followed it because I don't bet at weekends (family time and stuff like that), but this was pretty much an amazing hit rate tipping. Off course there will be days where the hit rate will fail a bit because you can't predict the unpredictable, but keep up the good work.
I'm fancy on NO BTS for the Palace - Brighton match. It has a nice odd and I think it might happen. thoughts:
Only new at this but I'm thinking I'll take a break. Well kind of. I want to develop some spreadsheet/database skills to help me rather than make picks for a little while.
Been using a spreadsheet from the start that I've copy/pasted form tables and overall tables into then trying to price up matches based on some calculations I had it set up do. I don't know if what I've been doing is sound or remotely accurate, just went with it. I've been getting odds close to the bookies (minus an overround) but nothing was tested at all. If the spreadsheet was throwing up potential value I'd then try and make judgement calls myself.
I don't have any skills as such, surprised the spreadsheet even functioned but I want to do some testing and learn a few things.
Got inspired by reading this last night - https://www.football-data.co.uk/ratings.pdf
Found someone who is actually running that, plus a couple of other rating systems to make predictions - https://sinceawin.com/data/archive/?list/2019-2020/20/0/11/3
I find all that stuff fascinating. Would like to be able to do that kind of thing myself with multiple rating systems that I either find online or make up myself off the top of my head.
Had a look at this today as well, same sort of thing:
So tonight I've been just trying to learn a little. Made something not even designed to be predictive just to learn may around Libre Office Calc a little bit. It's the percentage of home wins in the PL in 05/06 vs shots on target supremacy (home team shots on target - away team shots on target) in the actual games. Kind of pleased about being able to get this far in one evening so just for the hell of it:
Show the obvious, there was a trend between having more shots on target than your opponents and winning! Small dataset but I just wanted to learn how to produce these kinds of charts. Not predictive at all as it doesn't in any way attempt to address how to go about forecasting that a team will have more shots on target than their opponent before a game takes place. That's the sort of thing I might actually be doing with this ultimately.
Might have to learn how to scrape websites, because if I don't see a lot of manual inputting of figures in my near future where the specific bit of data aren't already nicely packaged and available for download.
Been using a spreadsheet from the start that I've copy/pasted form tables and overall tables into then trying to price up matches based on some calculations I had it set up do. I don't know if what I've been doing is sound or remotely accurate, just went with it. I've been getting odds close to the bookies (minus an overround) but nothing was tested at all. If the spreadsheet was throwing up potential value I'd then try and make judgement calls myself.
I don't have any skills as such, surprised the spreadsheet even functioned but I want to do some testing and learn a few things.
Got inspired by reading this last night - https://www.football-data.co.uk/ratings.pdf
Found someone who is actually running that, plus a couple of other rating systems to make predictions - https://sinceawin.com/data/archive/?list/2019-2020/20/0/11/3
I find all that stuff fascinating. Would like to be able to do that kind of thing myself with multiple rating systems that I either find online or make up myself off the top of my head.
Had a look at this today as well, same sort of thing:
So tonight I've been just trying to learn a little. Made something not even designed to be predictive just to learn may around Libre Office Calc a little bit. It's the percentage of home wins in the PL in 05/06 vs shots on target supremacy (home team shots on target - away team shots on target) in the actual games. Kind of pleased about being able to get this far in one evening so just for the hell of it:
Show the obvious, there was a trend between having more shots on target than your opponents and winning! Small dataset but I just wanted to learn how to produce these kinds of charts. Not predictive at all as it doesn't in any way attempt to address how to go about forecasting that a team will have more shots on target than their opponent before a game takes place. That's the sort of thing I might actually be doing with this ultimately.
Might have to learn how to scrape websites, because if I don't see a lot of manual inputting of figures in my near future where the specific bit of data aren't already nicely packaged and available for download.
Only new at this but I'm thinking I'll take a break. Well kind of. I want to develop some spreadsheet/database skills to help me rather than make picks for a little while.
Been using a spreadsheet from the start that I've copy/pasted form tables and overall tables into then trying to price up matches based on some calculations I had it set up do. I don't know if what I've been doing is sound or remotely accurate, just went with it. I've been getting odds close to the bookies (minus an overround) but nothing was tested at all. If the spreadsheet was throwing up potential value I'd then try and make judgement calls myself.
I don't have any skills as such, surprised the spreadsheet even functioned but I want to do some testing and learn a few things.
Got inspired by reading this last night - https://www.football-data.co.uk/ratings.pdf
Found someone who is actually running that, plus a couple of other rating systems to make predictions - https://sinceawin.com/data/archive/?list/2019-2020/20/0/11/3
I find all that stuff fascinating. Would like to be able to do that kind of thing myself with multiple rating systems that I either find online or make up myself off the top of my head.
Had a look at this today as well, same sort of thing:
So tonight I've been just trying to learn a little. Made something not even designed to be predictive just to learn may around Libre Office Calc a little bit. It's the percentage of home wins in the PL in 05/06 vs shots on target supremacy (home team shots on target - away team shots on target) in the actual games. Kind of pleased about being able to get this far in one evening so just for the hell of it:
Show the obvious, there was a trend between having more shots on target than your opponents and winning! Small dataset but I just wanted to learn how to produce these kinds of charts. Not predictive at all as it doesn't in any way attempt to address how to go about forecasting that a team will have more shots on target than their opponent before a game takes place. That's the sort of thing I might actually be doing with this ultimately.
Might have to learn how to scrape websites, because if I don't see a lot of manual inputting of figures in my near future where the specific bit of data aren't already nicely packaged and available for download.
I fiddled a long time with similar stuff, and it can be a nightmare to check whether the quoted odds are correct. I remember having done a lot of calculations for potential value-bets, only to figure out that a huge .csr sheet from football-data.co.uk did not have the same closing odds quoted as oddsportal did. There's loads of trial and error, but it's interesting. Always wanted to create a betfair model, but it takes a lot of time and effort to learn to code, and also to find out how to scrape the data, and in the end my day-to-day job, is just better investment of time vs money. Was pretty interesting never-the-less and had a lot of ideas, that I didn't follow through on though. The bookies are so hard to beat in the long run, as they close your accounts if you're consistently profitable, so that's also a pretty downer when having invested a lot of time. But it's an interesting journey nevertheless. Hit me up on a PM if you wanna share ideas or knowledge, or just someone to be the devil's advocate.
@Chipper: Your starting point should always be the closing odds of previous matches which are used to determine market implied ratings. If team A kicks off a -0.5 goals favourite against team B, and B were a 0.25 goals underdog against C, than A should be a -0.25 favourite against C, roughly speaking. Do these with all matches and you get ratings for every team in the league (Excel's solver function can do this). Of course these ratings have to be tweaked and adjusted for team news and form. Also these ratings can change and sometimes the market is wrong. But over a large enough sample you will see that there is a 1:1 correlation between Pinnacle's closing lines and actual results. They are countless studies and research papers that all come to the same conclusion. Previous closing odds predict future results better than anything else. Not goals, not points, not the current position in the table, not even expected goals. The market ratings take things like xG (which is just a measurement of actual performances on the pitch) into account, though.
Here's a recent example from Bundesliga II that I observed. Aue are currently in 6th and have been around the top 5 all season practically. Yet they continue to be priced like relegation candidates. How can that be? Well looking at the xG table, they are in 17th (from 18). They are also 16th in shots taken, 15th in shots conceded, 18th in corners. Doesn't matter which metric you use (xG just happens to be best one available in my opinion), they actually are a relegation candidate. Which is also what the expectation was before the season. They hosted Regensburg last weekend and looking just at the table (7th vs 6th, both on 23 points) you would conclude that these are two equal teams which, factoring in home advantage, would result in the home team being roughly a -0.25 favourite. Yet it was Regensburg who kicked off below evens on the 0 handicap. The market knows that they are a solid mid-table side and Aue one of the worst in the league. That the home team fluked another win after being completely dominated and outshot is beside the point. Your future evaluation of them should be based on market prices.
Here's a recent example from Bundesliga II that I observed. Aue are currently in 6th and have been around the top 5 all season practically. Yet they continue to be priced like relegation candidates. How can that be? Well looking at the xG table, they are in 17th (from 18). They are also 16th in shots taken, 15th in shots conceded, 18th in corners. Doesn't matter which metric you use (xG just happens to be best one available in my opinion), they actually are a relegation candidate. Which is also what the expectation was before the season. They hosted Regensburg last weekend and looking just at the table (7th vs 6th, both on 23 points) you would conclude that these are two equal teams which, factoring in home advantage, would result in the home team being roughly a -0.25 favourite. Yet it was Regensburg who kicked off below evens on the 0 handicap. The market knows that they are a solid mid-table side and Aue one of the worst in the league. That the home team fluked another win after being completely dominated and outshot is beside the point. Your future evaluation of them should be based on market prices.
Last edited:
You guys are quickly becoming living encyclopedias.
But if a team is consistently beating xG data maybe there are valid reasons. A keeper in amazing form, or strikers scoring chances they "shouldn't"?
Two bets tonight:
Increased my stake on Plymouth as they have drifted out. Don't really know why (see above), Bristol will also be without their suspended captain. Also something small on South Shields in the FA trophy replay. They lead their division (7th level) with three games in hand, whereas Southport are probably a slightly above average 6th division side. South Shields were much the better side in the first leg and should have won but their goalkeeper gifted Southport a late equaliser.
| Plymouth v Bristol, Plymouth -0.5 @ 2.47 |
| Southport v South Shields, South Shields 0.25 @ 2.04 |
Increased my stake on Plymouth as they have drifted out. Don't really know why (see above), Bristol will also be without their suspended captain. Also something small on South Shields in the FA trophy replay. They lead their division (7th level) with three games in hand, whereas Southport are probably a slightly above average 6th division side. South Shields were much the better side in the first leg and should have won but their goalkeeper gifted Southport a late equaliser.
This basically describes italian teams from the late 80's and 90's: deep defenses & fast and skilled forwards. The other teams played football while they scored the few chances they had.
To use my own example from Bundesliga II, only 17 games have been played and it is not even that unusual to see such a discrepancy between actual points and expected points. It happens every season, in every league. At this stage just getting lucky with refereeing decisions can make a big difference in terms of position in the table. It needs a much bigger sample to conclude their actual level of performance is indeed higher. The metric has also proven to be too precise and predictive to question it too carelessly. Even without delving into advanced analysis, just looking at their actual goal difference of +1 gives you an indication that their results are not sustainable. Goal difference doesn't nearly have as much predictive power as xG for example, but it is better than points or table position. See here.
As I said, the market can be wrong in individual cases. It is highly unlikely though, that teams get consistently over or undervalued by such a margin. Particularly at this level.
Pretty big on Villa -1.5 and -1.75. Was going to do a proper write up but short on time. Obviously pool is u23s (and missing their best u23s) and Villa is a rotated side but still strong. Don't think the market fully grasps the gulf between the sides. Astana United was a decent comparison. United side is way way better than the Pool team and Villa would beat Astana by a few goals. Watched Pool in u19s and u23s this year and their defence is a complete mess. Salz should have scored 5/6 last week against them. Fleetwood were 1.4 against the u21 in the EFL trophy. Can't see how Villa don't win by 2 with the gulf. It's a huge gulf in physicality, should bully them and have better quality also. I've got Villa -2.5 as fair line, think it'll be something like 4/5-0. Hopefully they don't expect to just turn up and win though.
Live handicap is still -0.75. Unless Villa completely take their foot off the gas, this seems really generous as well.
2 at Newbury tomorrow
12.50 Shishkin
2.25 The Big Breakaway
Both should win, double pays 11/8
12.50 Shishkin
2.25 The Big Breakaway
Both should win, double pays 11/8
Won easily. They couldn’t have played any worse either, show how big the gulf was. Won some on correct score too. Nice evening.
I went on Villa -2 @2.50. The bookies were completely blindsided on this game.
Fair enough. Do the bookies tend to be well aware of this nowadays and price more off xG than the table?
Great analysis...wish I'd seen it last night!
Treading water at the moment. Plymouth should have won but when the opposition keeper pulls off one of the best saves you'll ever see and you then concede a freak goal off a deflected clearance you know it's not going to be your day.
Thankfully I had all sorts of handicap and goalline bets on Villa in-running.
Yesterday someone claimed in another thread that the bookies are running balanced books which couldn't be further from the truth these days. Prices have to be as accurate as possible and that's why they follow the big (and therefore mostly sharp) money in Asia. The people/syndicates putting that money down use analysis tools that go way beyond what the average punter has at their disposal to squeeze every last bit of margin out of the market because they can be confident their own price reflects the true probability. The likes of Tony Bloom and Matthew Benham who run two of most well-known betting syndicates in Europe used advanced metrics (such as xG) long before it became mainstream and made fortunes off it.
I'm not sure how much pricing the bookies do themselves these days to be honest (maybe @Spiersey or someone working the industry knows more about it). In lower tier/obscure leagues that aren't traded (yet) in Asia sure. In these cases they might just use the current league table and the most recent results to come up with a price. That's why these leagues provide many more betting opportunities for someone who's willing to do a little bit of extra research compared to the Premier League where every little bit of data is easily available.
If the league table or the general public's opinion (or just this forum's) determined the market, then Liverpool would be the best team in the country. In reality, that is still Man City and both are priced accordingly in the market.
Last edited:
Going to risk on CR7 & Greenwood to score, combined odd of 7.20
Haha, read some of the stuff in this thread, put a double on Crystal Palace v Brighton to draw and I thought Plymouth to win. Checked my account this morning and actually had Bristol Rovers, so made a nice profit.
Also went on Liverpool -2. The mexicans totally sucked against a very weak team.
Bookies aren’t really bookies anymore (in the traditional sense of pricing things up themselves) they just follow Asia religiously basically. Every single bookie bar 365/Marathon and 1xbet just copy the prices from the other bookies for obscure.
365 will price obscure stuff themselves first, then some stuff like Efl cup etc and maybe some more mainstream early on as well, but would be small limits.
Most traders I speak to nowadays are completely clueless as all the skill has been removed from their job and all they do is monitor prices and make sure they are following Asia/365 prices
Bloom and Benham are incredible. Syndicates are unbelievably good at betting on prem etc. Shame their entire organisations are so secret, would love to get a proper insight/tour of their offices one day.
So true. About once a year I sit down to do some googling and find out more about them only to be greeted by the same old articles from years ago.
Do you know about Stratabet/Strategem? They operated a trading and analytics platform that very much looked like you would imagine something the syndicates operate only to completely shut down from one day to the next and with it almost all traces of its existence. Rumour has it that it was a maybe a little bit too good to be publically available and the syndicates intervened.
I'm a good while off bringing anything to the table but I absolutely will, thanks. It's all about acquiring those spreadsheet skills for now, not even going to attempt any models for a while
I've decided to make a start of testing fivethirtyeight's data against that on football-data.co.uk. What you say about oddsportal differing from footballdata is a bit of a worry for the future though. Maybe I could use both and go with whoever has the worst as it would be the more robust (and hopefully not impossible) test.
I know fivethirtyeight is a loser already, Pinnacle have tested it before, but that's not really the point for me at this stage, I just want to learn spreadsheets. I mean I literally learned 'find and replace' last night as footballdata and fivefirtyeight call different teams by slightly different names. Also how to use more than one sorting key at a time so I could get everything in the same order and even conditional formatting to highlight cells in different colours based on the what's in them. I've a long way to go before I can start producing reports of results by league, season, favourites, ranges of perceived value etc. etc. but that's what I want to able to do at some point. 538's data is sizeable, and it's on topic so it seems like the ideal material to learn with to me.
For now, and it's really nothing, I can say that since 17/18 backing all Austrian home teams that would have shown up as value on 538 would have lost me -38.18 points on 273 bets. This is going to take a while....
Cheers, makes perfect sense. I don't really know what else to say to that other than thanks. Nothing to ask, add or argue.
How do you get the data from 538 into one big sheet? is it available somewhere in one click as .csv or .xls? They're actually pretty good at predicting outcomes
This is what I was using - https://github.com/fivethirtyeight/data/tree/master/soccer-spi
The first one. It has their calulated probabilities, results, projected scores, XG where available.
Just thinking, if nothing else that's xG data all in one easy to grab place for possible future testing. They seem to be using a different model to Understat for that but it could be handy. Not encountered the non-shot xG before that they talk about in their description of they how they do what they do. Literally no idea what to make of that but it's an interesting concept.
Was on Liverpool v Monterrey BTTS & Over 2.5 @ 1.90. Good start.
Have a bit on Barca -1AH @ 2.05 against Madrid later too.
Have a bit on Barca -1AH @ 2.05 against Madrid later too.
I foolishly took Liverpool -1 with skybet. Useless.
I foolishly though Barca would beat Real tbf
Luckily I won a few quid on cards in that one, Suarez was a cert, as was casmeiro.I foolishly though Barca would beat Real tbf
- Status