Football Season Stats Review

[harry_rag]

I'll start off by saying this may well be of no interest to anyone other than me and I'm not sure if this is the best place to post it (it's not really "chat or banter" and isn't specific to a single league so it doesn't seem to fit in the footy section) but here's my thoughts on some stats I've been looking at across the "big 5" leagues.

• Average goals per league fairly steady at 2.81 compared to 2.8 last season. Some ups and downs in the various leagues but nothing likely to be significant.
   
• Penalties scored down from 553 to 471 (10.8% of all goals to 9.2%). I assume that's down to less being awarded rather than poorer conversion but I'll look for the pens awarded numbers. Maybe things are settling down VAR wise.
   
• Own goals fairly steady, up from 155 to 161 or 3.03% to 3.14%.
   
• Goals scored by subs up from 666 to 707 or 13.02% to 13.77%. Across the 2 seasons it's only been 9.16% in the Premier League (3 subs max) as opposed to 14.49% in the other 4 leagues (5 subs max). So we can expect an increase in England next season when the 5 sub rule is adopted. Obviously relevant to anytime goalscorer betting in terms of less goals being scored by the starting 11 when the number of subs is increased.
   
• The % of goals that had an assist was up from 66.8% to 68.2%, largely down to less penalties being scored. If you exclude own goals and penalties the % of "assisted" goals is fairly steady at 77.8% from 77.6%. So an average of 2.81 goals per games but only 1.92 assists. I have a strong feeling that most punters don't appreciate that gulf which makes the anytime assists market even more profitable for the bookies that the anytime goalscorer. I'm still tilting at backing in the goalscorer market (we'll see how that goes next season) but am also happily a seller at bigger prices. I've learnt my lesson with regard to assists though. Never a buyer or a backer be!
   

Will probably look at the bookings related stats next.

[harry_rag]

Some stats rating to bookings markets. I've added Portugal to the mix here given it's higher average.

League	Games	Yellow	Red	Av points	Av. cards	Points/card	Fouls pg	Fouls/Tackles	Pen pg	Refs used
Portugal	306	1740	120	66.67	6.08	10.97	29.93	0.99	0.35	25
Spain	380	1897	92	55.97	5.23	10.69	26.48	0.85	0.35	20
Italy	380	1696	93	50.75	4.71	10.78	26.68	0.88	0.37	45
France	380	1429	103	44.38	4.03	11.01	23.43	0.69	0.32	26
EPL	380	1299	43	37.01	3.53	10.48	20.23	0.62	0.27	22
Germany	306	1055	24	36.44	3.53	10.33	23.78	0.74	0.28	24
Total	1826	8061	451	50.32	4.66	10.71	25.09	0.80	0.32	27.00

Nothing that really surprises me in the points averages or the rankings. France has a higher ratio of reds to yellows hence the highest points per card number. I was quite surprised by how many different refs got a game in Italy! 

Not surprising that the top 3 leagues have the most fouls per game and highest fouls/tackles ratio. (Those numbers will be slightly off as the data had 1 game missing in both Germany and Portugal.)

In terms of real world betting observations, as I only ever sell bookings where the price is over 50 I haven't had many bets in France, Germany or England. For the top 3 my historical numbers look like this (not just last season).

Spain: 256 bets (140 wins and 116) losses, 69 points loss (biggest ever loss 98 points) - average sell price = 59.05

Thoughts - biggest sample size, slight overall loss (less than the single biggest loss), virtually break even so far, might as well carry on and see if it improves (still a small sample in the grand scheme of things)

Italy: 162 bets (99 wins and 63 losses), 542 points profit (biggest ever loss 81 points) - average sell price = 56.52

Thoughts - carry on, obviously!

Portugal: 24 bets (11 wins and 13 losses), 153 points loss (biggest ever loss 77 points) - average sell price = 67.79

Thoughts - carry on for now as it's such a small sample size. Not that many games get priced up in this league anyway.

I suppose I could work out a "magic number" for France/Germany/England e.g. where the price is at least x% higher than the league average but I'm not sure it's worth the effort (bookings overall is only a marginally profitable market). Maybe something to paper trade for a while come the new season.

[harry_rag]

On 6/22/2022 at 4:09 PM, harry_rag said:

Penalties scored down from 553 to 471 (10.8% of all goals to 9.2%). I assume that's down to less being awarded rather than poorer conversion but I'll look for the pens awarded numbers. Maybe things are settling down VAR wise.

If my sources are accurate I'm getting penalties awarded down from 687 to 585 so just shy of a 15% reduction in penalties awarded and scored. The conversion rate is steady at around 80.5%. That's quite a big drop!

I'm inclined to think there's some money to be had by laying the penalty taken market but it's hard to work out the "right" price to offer. Big swings from one season to the next don't help!

[harry_rag]

Here's some more "DOLITA" (Data of limited interest to anyone) and thoughts on a possible system for the new season.

From	To	Total	<40	<45	<50	<55	<60	<65
45	49.5	209	34.93%	55.50%	56.94%	70.33%	72.73%	79.90%
50	54.5	353	28.33%	46.74%	48.44%	63.74%	65.16%	74.79%
55	59.5	298	26.17%	39.26%	40.27%	53.69%	54.70%	64.09%
60	64.5	130	16.92%	36.92%	36.92%	52.31%	53.85%	66.92%
65	69.5	63	12.70%	26.98%	26.98%	39.68%	41.27%	53.97%
70	Plus	34	8.82%	14.71%	14.71%	26.47%	26.47%	38.24%

The data is a sample I have of 1087 games where I recorded the midpoint of the bookings spread price and the actual make up, so for games where the midpoint was between 45 and 49.5, 55.5% of the games had a bookings points total of under 45. The average for the whole sample was midpoint 55.34, sell price 53.9 and make up 52.52. The numbers align reasonably well with a table in a book I have which shows the expected unders numbers based on a team's average total bookings make ups.

Now fixed odds betting on bookings is a bit of a pain (trawling all the firms to find out the lines and prices, hoping that the attractive price isn't with a firm who only allow me a tiny stake) and there's a strong case to be made for betting on unders rather than overs. So I'm wondering if a viable approach would be to bet on BF aiming to get a price that equates to the "true" % with a bit of an edge added. Much will depend on what lines they offer on which leagues. I'll see what's offered where as the new season gets underway.

Selling bookings on the spreads seems reasonably profitable so it's not a giant leap to assume that there may be some mileage in looking to bet on unders (most people prefer to buy and most people prefer to back overs).

[harry_rag]

Nowhere else to post this bet so will stick it in here. Done doubles and up on a header scored in the following games with Hills.

Norwich v Millwall at 2.35

Sevilla v Valladolid at 2.35

BMG v H Berlin at 2.15

Lyon v Troyes at 2.1

Early days in terms of the data gathering for this (sparked by the obvious value in backing headers in the Women's Euros). There won't be any mileage on backing blindly at the prices offered but there are a few potential angles emerging. I've had a couple of bets so far and both have won (in the midweek Rangers and West Ham games). Don't want to go gung ho yet so, as there were 4 bets that made some appeal, I've gone for a modest stakes perm. 

[harry_rag]

The French game was the only one to see a headed goal so no return. A reminder to carry on gathering the data and hold back on the bets perhaps, at least doing the perm minimised the losses compared to singles.

[harry_rag]

I've played it in the Hibs v Rangers game at 8/5.

A standout bet by several measures and Rangers very much one of the teams always worth considering on the available data.

("Holding back on the bets" meaning betting by exception rather than not at all.)

[harry_rag]

1 hour ago, harry_rag said:

I've played it in the Hibs v Rangers game at 8/5.

A standout bet by several measures and Rangers very much one of the teams always worth considering on the available data.

Appears to be in at the 3rd time of asking! 

[harry_rag]

Most interesting angles thus far on the headed goals bets:

Back where the spreads total goals price (midpoint) is >3: Profit of 2.27 points from 24 bets with an ROI of 9.45%

Back where the best price is > than my "true" price (based on the spreads for headed goal minutes): Profit of 4.55 points from 11 bets with an ROI of 41.36%

Bonkers theory - it might be profitable to back <2.5 goals and a header scored in every game. 

I'll be looking to disprove that one (without chucking any money at it) unless the data makes an extremely compelling argument.

I'll do a summary of where things stand after the weekend.

As I say, I'm only betting by exception so far (122 games collated now and just the 3 singles and 1 perm bet to date).

[harry_rag]

Going for it in the Inter game at 10/11. Odds on not ideal but Inter are one of the prime teams for headed goals in any league and they’re significant favourites here in a game with an unusually high goals expectation.

[harry_rag]

No joy last night, playing E Frankfurt v Koln at 5/4.

One of those "better than true odds" bets (profitable but a tiny sample) and Koln had a very high headed goal % last season.

[harry_rag]

Another loser yesterday, probably wise to conclude that there will rarely be value bet other than in exceptional circumstances (e.g. what was suggested by the evidence of the Euros in terms of Women's International games). There are only two firms offering odds that I'm aware of and one of them has never offered a clear best price that I've seen, so it's essentially Hills or nothing. The data is up to 137 games now which is still very small but it's starting to look more predictable now. Here's a few observations.

• Total goals running at 3.01 per game versus an expectation (based on spreads midpoint) of 2.84. Blindly buying total goals would have been profitable but I'm pretty confident there's no mileage in that as a strategy. With a bigger sample size I'd expect buying or selling both to be loss making. I think there's been an unusually high number of goals in what is a small sample size.
   
• Total goal minutes; 412 goals with an average goal minutes of 48.3 (the rough rule of thumb is 50 minutes per goal). You'd have made a loss blindly buying or selling total goal minutes which is pretty much what you would expect.
   
• Headed goal minutes; 69 headed goals with an average goal minutes of 46.2. You'd have made 283 points profit selling headed goal minutes in every game compared to a loss of 428 points if buying. It wouldn't surprise me if selling was profitable in this market long term as I imagine most punters would want to buy but it's a low price point to be selling at. You'd need a lot of winners to make up for the occasional heavy loss.

So much for the spreads, fixed odds thoughts to follow.

[harry_rag]

So, that's 137 games with 412 goals, 69 of them headers and at least one header scored in 57 games. 16.75% of the goals have been headers, entirely consistent with last season's data for all the major leagues. 57 out of 412 games gives simple "true" odds of a shade over 2.4.

Header scored odds have been available for 116 of the games (at an average of 2.26) and if you'd bet blindly in every game you'd have lost 16.66 points with an ROI of -14.36%.

I've calculated the following "true" odds:

• True odds based on the % of goals being headers implied by the spread prices for each game (this has ranged from around 14% to 20%, presumably reflecting the firms' views of the aerial threat posed by the teams involved). You'd have lost 4.35 points (-3.75%) had you been able to get those odds in all of the 116 games where a price was offered. Backing all the selections where the odds available were bigger than this "true" price is showing a profit of 4.3 points (35.83%) after 7 winners from 12 bets. So an interesting angle but too small a sample to leap to any conclusions on.
   
• True odds based on assuming 16.5% of all goals will be headers (based on last season's data) Had you been able to get this price in all of the 116 games you'd be 1.12 points up with an ROI of 0.97%. There hasn't been a single game so far where you could have got on at this price (probably making it a reasonable yardstick for a true price).

I'll carry on collecting the data for a while and having a small bet on the so far profitable selections where the best price is > than the true odds as mentioned above, I'm not convinced it'll be that easy long term but pure paper trading can be a little too dry.

[harry_rag]

22 hours ago, harry_rag said:

I'll carry on collecting the data for a while and having a small bet on the so far profitable selections where the best price is > than the true odds as mentioned above,

On at 6/5 in the game between Red Star Belgrade and Maccabi Haifa ("true" odds 2.16). I don't have headed goal stats for those teams' domestic leagues but their CL games so far are reasonably promising. MH scored a headed goal in the first leg and scored 2 headed goals in one leg of each of their earlier ties. RSB scored one in one leg of their only other tie. So 6 headed goals in 4 out of 7 games all scored by the teams involved tonight.

[harry_rag]

No joy last night, going again in the Rangers tie (won in 1st leg) at 11/10.

SPIN have unusually high prices in that game (creating a 3 point arb) and the Trabzonspor game (a whopping 6 point arb) so I'm slightly wary of reading too much into them. If they dropped their price in the Rangers game to close the arb then the bet would still be value so I've taken that one but, as it wouldn't be the case, not the other game.