Effect of Early Goal

[jonnyg]

this is a key point < why were Rosenborg priced so short to score goal 2 < incredibly I have the answer < hang on >

“In this masterwork Pullein explains how you can work out what is likely to happen …. Kevin Pullein is the leading football commentator for the Racing Post. … for instance, a team which is losing by 0-1 is 6% more likely to score the next goal “

the above comment is in effect how the in play betting algorithm is set up < you will have betting affiliates advising Rosenberg dnb / to score 2nd goal etc <

[jonnyg]

Haugesund 1-0 HT

8/11 5/2 9/2

over 2.5 is offered at 5/6 < should be around 6/4

Huag 2nd goal = 15/8 no 2nd goal = 7/2 , Rosenberg 2nd goal = 5/6

In the Table below I have added the last 44 games in Norway Tippeligaen where the home team, opened the scoring 0-10 and 1-0 Half time , and 41% ended over 2.5 goals

The home teams won 70%, 23% ended a draw and the away teams won 7%

23% of the games in the table below followed the pathway to 1-1 with the 2nd goal in the 46-70 time band.

data correct to May 1st 2017

still 1-0 on 73 minutes

in effect how likely were Rosenberg to fight back given the events of the first half = survival analysis < the betting industry had much higher hopes of a fightback then they should have because of the first half events >

[jonnyg]

Will provide a summary of the in play analysis this afternoon

1. we had that game in Korea < away fightback > 1-1 in 2nd half < 1-0 HT > with lowered expectation of an away win < FT 2-1

2. HJK game < market had much higher expectation of over 2.5 goals at HT and HJK winning the 2nd half given the events of first half <

3, the market at Ht has Rosenborg at a much higher expectation of scoring goal 2 = survival analysis < given the first half events >

very solid afternoon < dont worry about my image < I am retired

[jonnyg]

very pleasing afternoon > hope you enjoyed it < as a point of interest because I am reading a book on artificial intelligence < is it possible that the betting industry in play algorithm is not as sophisticated as the majority of people believe ??

[jonnyg]

I can see what I consider a pricing error in the Huagesund game <


Rosenberg are offered at 4/5 to score goal 2 when the home team given the events of the first half are just as likely / i would say more likely to score goal 2 <

I have looked at 1-0 HT goal 0-10 minutes in most leagues around the globe <

in many leagues > 1-1 in 46-70 is not common at all < so in that time period as time decays > 2-0 will be far more likely then > 1-1

Rosenberg did not go > 1-1 in 46-70 time band

and on 71 are now 15/8 to score goal 2
[/quote]

on 83 rosenborg are offered at 5/1 to score goal 2
[/quote]

FT 1-0 > spot on

[jonnyg]

Posted on: September 12th, 2015
In the table below I have added the last 26 games in the Korea K League Classic where the home teams were winning 1-0 at half time and there was a fightback ( 1-1 ) with a goal in the 46-70 minutes time band, 69% ended over 2.5 goals and 19% ended over 3.5 goals

The home teams won 38.5%, 38.5% ended a DRAW and the away teams won 23%


Jeju Utd away fightback 10/3 Evens 2/1

on 60 minutes over 2.5 is offered at 1/2

in this profile in K league < you will see lowered expectation of an away win so we can see in theory that Jeju should not be fav at 1-1
over 3.5 goals in this profile in K league will be a rare event >

0n 69 > 7/2 5/6 12/5 < it is Jeju that should be around 7/2

bet advice < at 1-1 Gangwon draw no bet at 6/5

2-1

FT 2-1

bet advice at 1-1 > Gangwon draw no bet @6/5 Won

SPOT on

[jonnyg]

KuPs 0-1 HT

28/1 11/2 1/7

over 2.5 is offered at 6/5 and should be around 15/8

HJK Helsinki are offered at 10/11 to win the 2nd half ( 2nd half betting ) and this price looks way out of line with expectation given the events of the first half < the price should be much bigger > ditto HJK are offered at 8/11 to score goal 2 and that price should be bigger


why ?

0n 63 minutes over 2.5 is now offered at 12/5 with HJK now offered at 7/5 to win 2nd half

on 72 over 2.5 is offered at 4/1 with HJK to win 2nd half at 7/4

on 78 HJK are now offered at 9/4 to win 2nd half

SPOT ON

[jonnyg]

I find football data and analysis better then sex so naturally you may not share my passion for football data and analysis > probably the most trolled football data and analyst in the history of social media <

on a serious note < football data and analysis in general is at a really weak level < I fear for the young people that have an interest in football data and analytics > things can only get better ???

[jonnyg]

Newells 1-0 HT

In the table below I have added the last 55 games in Argentina Primera ( regular season ) where the home teams opened the scoring with a goal in the 0-10 minutes time band and 38% ended over 2.5 goals

44-7-4 < the home teams won 80%

9/55 followed the pathway > 1-1 with the 2nd goal in the 46-70 minutes time band and 2/9 ended 1-1.

“In this masterwork Pullein explains how you can work out what is likely to happen …. Kevin Pullein is the leading football commentator for the Racing Post. … for instance, a team which is losing by 0-1 is 6% more likely to score the next goal …” < well not in this profile of game

Trading Strategy

In Argentina Primera in games where the home teams open the scoring 0-10 minutes and the half time score is 1-0 ; back against ( lay ) the 1-1 correct score and if still 1-0 on 69 minutes then scratch the trade by backing 1-1.

1-0 half time to the home team in Argentina Primera added to the early goal metric and 1-1 with a goal in the 46-70 minutes time band will not be a common event and if this does happen then 1-1 full time will be an unlikely outcome.

data correct to March 27th 2017

[jonnyg]

would your jaw drop if I told you > since 2011-2012 in the Premier League > 1-0 HT to home team 0-10 minutes > 1-1 with 2nd goal in 46--70 minutes and only 2 games ended 1-1 FT ?

[jonnyg]

In the table below I have added the last 20 games in the Premier League where Chelsea opened the scoring away in the 0-20 minutes time band and 0-1 half time , 65% ended over 2.5 goals , 1/0.65 = 4/7

The last 20 away games played by Chelsea when they opened the scoring in the 21 minutes to half time band and were winning 1-0 at half time , 35% ended over 2.5 goals , 1/0.35 = 15/8


data correct to Jan 14th 2017

we can see in general 0-1 HT goal 0-20 and 0-1 HT goal 21-HT < are not in the same ball park > as a result data in terms of sample size can be a monster red herring <

[jonnyg]

they say everyone likes an underdog < just not in football data and analytics

[jonnyg]

if you have read the book the numbers game > an epic flaw in one of the themes >

You will often hear commentators claim that teams are more vulnerable to conceding a goal just after they have scored themselves. True or false?

Peter Ayton and Anna Braennberg tested the validity of this belief. They did so by considering matches across two seasons in the English Premier League that ended in 1-1 draws (this, they reasoned, shows that ‘each team was capable of scoring against the other’). There were 127 such instances during the 1994-95 and 1995-96 campaigns. I came across their study while reading Myths and Facts about Football: The Economics and Psychology of the World’s Greatest Sport (2008).

So, is the team that trails to the game’s opening goal more likely to score just after they have conceded themselves than at any other time during the match?

For each game, the time left after the first goal was divided into four quarters. Say, for instance, the opening goal was scored after 50 minutes. The remaining 40 minutes is divided into four 10-minute quarters. If teams are more likely to concede immediately after scoring, then, naturally, you would expect more equalising goals in the first quarter than the fourth. Is this the case?

Mr Ayton et al did not recognise that it is an early goal in general that is more likely to result in a quick response and a very low % of early goal games will end 1-1

in effect the sample of 127 would have included very few early goal games >

[jonnyg]

the monster flaw in the book was that football is 50% random based on Birmingham beating Chelsea 1-0 when Chelsea dominated the game < a game I watched on video many times >

here is my response > in 2013

The Numbers Game written by Chris Anderson and David Sally is a very interesting book which I am currently reading and on page 39 the book discusses ‘riding your luck” .

Anderson and Sally use the example of Birmingham v Chelsea which was played in the 2010/2011 season to advise that teams can dominate games and lose .

I have just watched the first half of the game and keyed the shot strength the same way as it is done in football trading houses.

A delivery is basically a cross into the box, a near miss is a attempt at goal which you consider that there should be a goal and Chelsea had two in the first half and in the first one, Drogba only had Foster to beat from close range but shot straight at Foster and in the second one, Drogba headed the ball from close range onto the inside of the post with Foster beaten and the rebound was fired wide.

A chance is when there is less expectation then a near a miss of the shot ending in a goal and a half chance is one step down from a chance in terms of expectation of a goal and you can use shot strength to have expectation of the score in the game and I award 0.01 for a delivery, 0.05 for a half chance, 0.1 for a chance and 0.3 for a near miss so we can see that in the first half although the game was 1-0, in terms of the shot strength data it was Birmingham 0.13 v Chelsea 1.59 so we can see that Chelsea did dominate the first half.

In terms of expectation of the outcome of the second half we should not assume that because Chelsea dominated the first half that there will be expectation of Chelsea winning the second half because it is half time game state that will give the expectation of the accuracy of both the teams in the second half.

The question is was Birmingham beating Chelsea a random event and my argument is that although Chelsea did dominate the first half we need to look at the variables in the game to define if the game can be considered random.

Chelsea did not have Terry at the heart of the defence and the goal was a result of the centre backs both going for the same ball which allowed Bowyer to score unmarked from close range.

Essien and Lampard were also missing and Anelka was playing fairly deep and Drogba was basically the big threat in terms of Birmingham conceding and Foster countered the threat of Drogba by having an outstanding game in goal.

Chelsea had lost their previous league game at home to Sundeland and went onto win just one of their next seven league games after losing to Birmingham although it was their first back to back defeat since the 2005-2006 season.

If we look at 1-0 half time game state around the world then there is expectation that both teams will have similar accuracy in the second half but that the home teams shot on target production will be on average higher then the away teams in the second half.

In the 2010-2011 season there were 83 games that were at 1-0 half time game state and Stoke were the only away team to win and there were 16/83 19% games that were drawn.

At half time there was little expectation of Chelsea winning given the historical 1-0 half time game state and they may have had more possession in the second half but their shot on target production in the second half was only two.

Birmingham in the 2010-2011 season lost only 5 league home games and only conceded 22 goals at home and won all 4 of the games when they were winning 1-0 at half time at home and in the second half of the games their shot on target production was 6 in the second half of the 4 games and they scored with 50% of their shots on target and conceded 12 shots on target in the second half of the 4 games and conceded only one goal.

In the 2010-2011 season Chelsea did not win any games away when they failed to score in the first half and in the 4 games where they were losing 1-0 at half time away, their total shot on target production in the second half was 9 so they scored with 22% of their shots on target in the second half of the 4 games and the home teams had a total of 3 shots on target in the second half courtesy of Arsenal as the other three teams did not have a shot on target in the second half.

Conclusion

A goal in football is certainly a rare event and teams do win games and have less shots on target and I think the key to the Birmingham game is that Chelsea were exposed at the heart of the defence for the goal and it was only Drogba that tested the Birmingham keeper in the first half and in the second half given all the data that there is for 1-0 half time game state, resulted in little expectation of Chelsea fighting back and winning the game.

Chelsea had 32 shots of which 9 were on target, 15 off target and 8 blocked and 21(66%) of the shots were inside the penalty area.

This season only Everton and Manchester United have won away in the league when losing 1-0 half time away and it is not a common event.

[jonnyg]

this is the massive faw re football data and analysis > you will hear loads of people advise they have 20/30 /40 years of experience of betting >

but very few people want to push football data and analytics to the next level < the metric expected goal totals is considered by the mainstream as a great new metric when it has been used for over 10 years

and has severe limitations