Tuesday, 31 March 2015

Starting XI Green All Over

A quiet weekend in the FTL with the top leagues idle, but it was a memorable round as today's post has a very appropriate 'green all over' look about it. Every one of the eleven active entries made a profit, with the total profit of 49.24 in large part owed to Crewe Alexandra's 90th minute winner at Sheffield United in League One on Saturday, a game Crewe were priced at 10.26 to win.

Not surprisingly, that one result alone made a big difference to the FTL table, and sees the March monthly prize currently heading to Mortimer (currently up 21.80 points this month) who may well become the first contestant to win two monthly prizes having previously been January's winner.

BettingTools.co.uk reclaimed first place with a small profit of 1.47 points, and flying out of the red up ten places into fifth was Mortimer who made 13.99 points.

Slightly less spectacular was TFA_Raz's climb up six places after making 9.89 points on the round and he also moves out of the red.

Fairfranco stayed in 10th place making 2.79 points and Bounty Boy Football Elite also emerged from the red zone with 7.39 points moving him up five places into 13th place. The 14 entries in positive territory are:

Still in the red, but moving up three places, was fellow Bounty Boy Skeeve who made 2.53 points.

Drawmaster continued a good recent run picking up 2.81 points, and @ValueBankFooty made 2.22 points.

Bounty Boy TFA Draws made 4.35 and moved up three places, while Football Investor was idle. Mountain Mouse picked up 1.61 points while Sjosta made 0.19 points.

Saturday, 28 March 2015

Tempestuous Seasons

Approximately 50% (which, for any non-statisticians reading this, is about half) of the hits on this blog come from outside the UK, so I do try to write as clearly as possible. Despite my efforts, one individual (and only one) seems to have great difficulty understanding what I am saying, or rather what I am trying to say, about the Bundeslayga system.

Perhaps some of the more subtle nuances of the English language are lost despite my best efforts, or perhaps he is the ‘tortured genius’ or tortured artist of betting, brilliant in his own field, yet lost outside of it, or perhaps he just likes attention. There’s an old saying that says:

Arguing with a statistician is a lot like wrestling with a pig. After a few hours you begin to realise the pig likes it
I'll try again. My review of the Bundslayga system yesterday was just that - a review. It was a verification of how the system has performed, in the past, and the results given represented the very minimum return possible. In the real world, a little effort would boost those profits.

There is no hidden agenda, and the review made no claims about the efficacy of the Bundeslayga in the future. In fact, readers of this blog will know that I have often written that it should not work, that I am surprised it does still work, and that at some point it should cease to work, and I have also clearly stated that this is not a ‘get rich quick’ system. However, it does have value to someone in regard to mitigating Premium Charges, either now or in the future.

Why the individual doesn’t focus his attention and undoubted statistical skills on finding a better way to avoid the Premium Charge I can only guess, but the fact is that for many years this method has had an edge. Pinnacle Sport’s over-round on the Bundeslayga markets averages 2.02%, so any ROI above -2% should be celebrated, not derided.

The edge will likely not continue, and I am certainly not saying it will continue. 258 bets isn’t a huge sample, but the real-life experience of profitability when laying on the exchanges goes back prior to the 2012-13 season. I started paying the Premium Charge in September 2008 and the Super Premium Charge in October 2012, and I identified this possible mitigation strategy in 2010.

It has worked for me. If it works for you, try it. If it doesn’t, ignore it. The system is free – adopt, adapt, improve or ignore. If you don’t use Exchanges for your betting, then its value is diminished, and if you don’t use Betfair (and need to generate more turnover), it is even more diminished, but why there is such negativity over an idea that has at least some proven merit baffles me. Another English expression - "if you can't say anything nice, don't say anything."

Steve hits the nail on the head - maths is important, but not the most important aspect of betting. Sjosta (Shosta in English) would be running away with the FTL if that were true, but last time I checked, that wasn't the case.
For anyone interested, the link is active here.

While an understanding of basic statistical principles (obviously including probability) is important to have when betting, any more than that is at best unnecessary, and quite possibly an impediment. There’s an expression in England – you can be too smart for your own good. It means “believing that you know everything to the point to where it can harm you or work against you”. The genius often proclaims “I am a statistician and I can tell you…” which is wonderful, but for practical purposes, so what?

Dmitri's thoughts were very similar to those Tage shared...Imagine!
For most of us, success in betting is about how much money we make rather than in proving our academic brilliance. I certainly don’t claim to be a statistician, (a real statistician tells me quite often I am not, in case I should have any delusions of grandeur) and I don’t believe Steve M (Daily 25) claims to be one either, yet we have both had success over the years. Yes, success is relative, but when an estimated 95% of bettors lose money (citation needed), making any profit is arguably a success. Worth mentioning that Steve’s style is a lot more ballsy than mine, which just shows that there is more than one way to skin a cat (another English expression – this post is turning into quite the English class).

A couple of others out there who have blogged away over the years and have also had much (relative) success are Mark Iverson and Ian Erskine who, and if I am insulting either of them about their lack of statistical qualifications, I don’t mean to.

What the four of us do have in common however, is a positive attitude, which may or may not be a success factor, and I’m guessing that the door to sports betting for all of us was probably sports rather than statistics. Maybe a life spent supporting teams like Tottenham Hotspur and Crystal Palace prepares one well for the ups and downs of betting?

I suspect that for most bettors, statistical skills up to a certain point are helpful, but beyond that, the law of diminishing returns rapidly kicks in. It may even turn negative if it leads to an ultra-cautious approach. Back to the Bundeslayga since 2012; an ROI of (at least) +1.68% versus an expected return of -2% may well not be statistical proof of an edge, but for how long do you wait in the real world?

I think at some point you have to say the edge is good enough to warrant getting involved. We’re not talking about getting married or a life and death situation. If things change as they ultimately will, then stop. Hopefully your profits while the edge lasted exceed the losses after it vanished. John Maynard Keynes said:

“In the long run we are all dead. Economists [Statisticians?] set themselves too easy, too useless a task if in tempestuous seasons they can only tell us that when the storm is long past the ocean is flat again."
People were very eloquent in ye olden days. More recently, Matthew Iglesias wrote in Slate: 
Compared to most contemporary economists, Keynes was an extremely deep thinker about the question of probability. A point he makes elsewhere about the long term is that, if you try to plan decades in advance, you'll find that unquantifiable uncertainty dominates quantifiable risk.
To me, this means that at some point, you need to take the plunge. If you wait too long, any edge that may exist will be gone.


The stereotype of a statistician is that they're male, they wear pocket protectors, they're very formal, they’re logical thinkers and they are socially awkward, (some might say rude), but admittedly that of the IT worker (e.g. me) is almost as negative (although one or two people who have met me think I’m kind of OK. Well, one if you don’t count my Mum). As for my Dad, well, wherever he is right now, I know he's looking down on me. 

He's not dead - he's just very condescending.
Mark, Ian, Steve and myself have all been around for a long while, and I’m pretty sure Ian has made seven figures from betting, Steve has made (give or take) AU$417,902.36 and both Mark and myself have hit the Super Premium Charge. It’s possible of course that were we all qualified statisticians, Ian and Steve would have made far more money, and neither Mark nor myself would be paying the pesky Super Premium Charge at all having found a way around it, but it does show that it is quite possible to make a (relatively) decent profit without a knowledge of advanced statistics.
I agree with Ian who has said that the key ingredient to being successful in betting is not so much in IQ but more in the approach and attitude one takes. 

He’s also observed that you can teach a group of people the same strategy, and then observe that some profit from it, while others don’t. Success at betting is about more than just facts and data, and of course it has to be. Any edge from data is (usually) quickly self-correcting, so there has to be an ‘intuitive’ factor required which perhaps is more readily available to some personality types than others. Successful betting is more art than science.

Friday, 27 March 2015

DC v UR v Sq

A couple of comments on my proposal to modify the formulae for Lay bets. Marty was straight to the point:

I object to revising the settlement of lay bets
Well, fair enough, although I’m not sure what FTL entry (if any) Marty is voting on behalf of.

Jamie A, who actually IS an entry, suggested an alternative approach (I’ll call it the Under-Round):
So back bets have a 2% margin and lay bets have 4%!
I would understand making the lay bets have a higher margin for the rest of the season since they had a lower margin for the season to date - but the fairest way is still:
Pinnacle back price * (pinnacle margin)^2 = lay price
Then lay and back markets have the same margin.
While Jamie’s formula is solid enough, it seems to me that this would be ideal if the back prices they are derived from are Exchange prices, but they are not, since we don’t have Exchange prices available.

Using the Double Chance [DC]  formula means using actual, and verifiable, prices so I still prefer my Double Chance approach for a couple of reasons.

The first is that a Lay bet is nothing fancy or mysterious, it is simply a convenient way of backing all other possible winners in the market. In a binary event, a Lay of A is a Back of B. Lay Federer / Back Djokovic – it’s the same bet.

Football has its pesky Draw, but a Lay of H is still simply a Back of both D and A. Using the DC is merely splitting the one Lay bet into two Back bets. Joseph Buchdahl suggested using the +/- 0.5 prices which would indeed be the perfect (except for Draw lays) answer here, but those prices are not available, at least not where the Match Odds Pinnacle Sports Prices are available, and where I need them.

The second reason is that this method is directly using prices that are extracted, and are available on the Football-Data.co.uk web site.

Although the FTL is intended to be more for comparison purposes (hence the use of one book’s prices rather than the best), the Double Chance method does more closely reflect the real world experience.

The differences are small anyway. A lay at around the 2.0 mark, and the DC formula records a profit of ~0.93 as does Jamie’s UR method. At shorter prices, the DC method is slightly higher (a lay at 1.65 for example returns ~1.41 rather than ~1.39) while at higher prices, the UR returns are slightly higher (a lay at 2.4 for example returns ~0.66 for the DC, ~0.67 for the UR).

So, FTL entrants, we have three options:
  1. Leave the method as it is today - Status Quo (Sq)
  2. Use the Double Chance formula
  3. Use the Under-Round formula
Since this is a quiet weekend, feel free to submit your votes with your entries before the April 4th round. April already! 

To satisfy my own curiosity, I ran the DC formula numbers for the Bundeslayga since 2012-13 (the first season that Pinnacle Sports prices are reported), and they show a profit of 4.33 points from 258 selections. 

As I have been saying since at least 2013 about the Bundeslayga:
It's proven profitable for several seasons, and used more as a Premium Charge mitigation system than one that will make me rich.
An ROI of a verifiable minimum 1.68% (Pinnacle Sports match odds can usually be beaten) may not be exciting or glamorous, but if you are either paying, or may be paying in the future, the Premium Charge, methods like this are very useful. (I suspect any critics are currently exempt from this charge). The rules used for this verification were Home team between 1.3 and 2.0 inclusive on Pinnacle Sport. 1.3 to 1.39 inclusive would have been a higher ROI%, but then I am stupid.
Although Tage may well be the most brilliant statistician to come out of Aarhus, (why do I always want to sing "in the middle of our street" when I hear that name?), I'm not sure he would score so highly on the arguably more important Emotional Intelligence scale.

As the input from the rather more polite Jamie shows, there is no 'correct or wrong' answer to the question. In the absence of a +/- 0.5 price, the Lay profit needs to be derived. The original Lay rules were based on real-life experience for a Friendly competition, were made clear from the start, and apply to everyone.

Anyway, the Double Chance and Under-Round methods are both fine with me, and I'll go with whichever receives the most votes.
Very well put FT member Daily25. Please vote! Steve's blog is one of the best out there, in large part because of his positive attitude even when things aren't going well.
It seems that a lot of the services Steve uses are having bad runs right now. It happens.

Too many people are only too quick to assume the worst and jump to (usually wrong) conclusions. Or they put 2 and 2 together and come up with 5 (not Tage - he's a retired statistician). And an argument over in-play trading versus pre-game betting is ultimately pointless since it's a personal preference. For some sports, trading is (in my opinion) the only way to be profitable long-term. For others, where in-play has little liquidity, the value is (in my opinion) likely to be pre-game. If you're making a profit, excellent. It doesn't mean that others are not, and more than one style can be profitable at the same time. They are not mutually exclusive. Your results are not improved by wasting time finding fault with the approach of others; in fact if that time were spent on fine-tuning your own methods, your results might well be improved. Betting is all about opinions - if we all had the same opinion about everything, there'd be no market - but the debate over opinions should at least be civil and criticisms constructive.

A puzzle, courtesy of @DavidJBodycombe :

Wednesday, 25 March 2015

FTL Update 24.March And FTL Update

A very quiet midweek, with entries from Talkies Tips and Mortimer only. Talkies Tips was the winner, making 2.92 points and moving up two places, while Mortimer was the loser dropping 3.00 points, three places, and into the red. Overall a small loss of 0.08 points on the round.

Here are those entries along with the innocent, idle bystanders moving either up or down:

The penultimate monthly prize will be headed to Australia unless someone can overtake Gecko from the abbreviated schedule coming up.
Not to jinx him, but Gecko would be the seventh unique monthly winner. 

Also, unless I hear any objections, I'll revise the settlement of Lay bets moving forward effective immediately, to use the odds on the other two options (Double Chance) to calculate the returns.

So for a match where the odds are for example (over-round 200%)
a lay of the three options will now return (rounded):
This compares with the old formulas returns of 0.29, 0.38 and 0.92. Small changes that will only affect Layers, and by very little, but in the absence of the AH +/- 0.5 prices, this method does have the advantage of still using Pinnacle's prices. And as stupid as I am, it didn't take me too many days to come up with the formula.
As with the other bets, the Pinnacle returns represent the minimum return you could expect. In real-life they will be higher - how much higher depends on where you can bet. 

Tuesday, 24 March 2015

The Good, The Bad And The Downright Rude

It's strange how people can be so different in their attitudes. One might be out running errands on a Saturday morning, dropping off alterations for example, and have car problems. Some people are all too happy to help, running around finding jump leads - others quite the opposite.

It's the same on the Internet, where the relative anonymity allows people to behave in ways that they wouldn't in the 'real world'. Human nature being what it is, some people are happier to continue convincing themselves that their ideas have merit, rather than accept fact based evidence or logic.

Resorting to harassment as a means of deflecting attention is, of course a rather silly idea, with the police far more willing to issue warnings - Police Information Notices - to trolls these days.

We all know I have been threatened myself with this despite the fact that I don't troll or harass, and have been waiting for several months, but some people do troll and harass.

In case the perpetrator is blissfully unaware of what harassment is, writing five blog posts on one subject (about which they have no direct knowledge incidentally) in one day qualifies as harassment. Using someone else's name without their permission qualifies as harassment. There are also laws covering libel, which can be a problem when you insist on writing about a subject you know nothing about.

For a Police Information Notice to be issued, it is also necessary to show a "course of conduct", which apparently means showing that it is not just one event. A thirteen month campaign qualifies with ease.

I trust this friendly advice will see an end to this. Ignoring and blocking worked fine, but when it is brought to my attention that legal boundaries are being crossed, action is needed. 

In the event that it continues, I'll be interested to hear form anyone else who has crossed paths with this individual. Thanks to those who have already provided valuable advice and information. I do know that there are good people out there too.     

Monday, 23 March 2015

FTL Update 22.March

No Monday matches for a change, so an earlier update than usual. Highlights from the weekend were that no less than four entries dropped into the red this week, leaving just 12 remaining in profit, and we have another change at the top. The Erskine Cup semi-finals went to form, and overall the entries combined lost 29.15 points, making March the worst month of the season.

First the league, and with just the dozen entries in profit, I'll lump them all together.

Randolph was the big points winner of the week going 12 for 12 and making 6.74 points. He won his Erskine Cup semi-final, as did Gecko, who takes over from BettingTools.co.uk in first place after making 0.83 points. On another tough weekend for draws (10 in the 49 Big Five matches) the biggest loser was XX Draws which lost 9.77 points and Trend to end who lost 7.00 points. Club Havana made 5.42 points and climbed three places as did the idle Betcast thanks to the woes of others. Gecko, Mortimer and BettingTools.co.uk are the March leaders.

The group down by less than 10 points is:
The only winner here was TFA_Raz, who made 2.65 points and climbed four places, while Football Elite fell seven places after losing 6.34 points but remain top of the Bounty Boys. Jamie A lost points, but moved up two places.

The rest of the table looks like this:
Rubicon was the big winner in this group winning on 7 of his 8 EPL selections, and made 3.99 points. Drawmaster made 0.30 points profit and moved up two places, while TFA Euro Draws and @ValueBankFooty also made small profits. Mountain Mouse was the biggest loser followed by Talkies Tips (down 5.37) who along with XX Unders, was eliminated from the Erskine Cup.

March isn't over yet, although with the International Break there will be less activity next weekend (I plan my holidays around the FTL you'll be pleased to hear), but here are the results by month - a steady decline is rather noticeable:
I would have expected the opposite with models absorbing promoted teams and improving as the season goes on, but that appears not to be the case. A reminder if anyone needs it, that profits can turn into losses, and vice versa, and even a season is a small sample size.  

The Erskine Cup played out as below, with the Final between the current 1st and 3rd placed 'teams' in the league. For a competition with a large element of randomness, the cream certainly rose to the top for the final:
The 'portfolio' numbers continue to decline:
Cassini drops into the red overall, and there were losses for the Bounty Boys. Football Elite has a Bounty liability of £425 closely followed by Skeeve (£475). Skeeve's official bets were profitable, but because there are limitations on what prices are available in the Football Data statistics, (i.e. no handicap prices), FTL Skeeve has to settle for straight wins at shorter prices rather than the actual -1.5 or whatever in his official bets. Football Investor (£500) was taken ill, while TFA's Draws (£500) dropped three more points.

Sunday, 22 March 2015

Billy Walters

I've written about Billy Walters before. He has one home (he has several) close to that of Mrs Cassini's parents, and plays golf at a club where an acquaintance is the club professional, so I hear second-hand stories about him, but his story is an interesting one anyway for many of us given that his success appears to be down to data analysis rather than access to inside information. 
Here's a timely article from ESPN:  
On a mid-September night, Ezekiel Rubalcada swaggers through the doors of Las Vegas' M Resort Spa Casino. He moves to the half-moon-shaped bar overlooking the sportsbook, catching up with old friends. He exchanges high-fives and playful F-bombs with a couple of bartenders, then makes a raunchy pass at a brunette waitress who escapes in a near sprint.

For 38-year-old Rubalcada, being at the M is a pleasing trip down memory lane, a visit to his primary workplace throughout 2010 and 2011. Back then, he had nearly $1 million in his account at the M. Dressed in slacks and a sport coat, he would saunter in and bet six figures a week on NFL and college games. He was, M Resort staffers say, one of the sportsbook's "bigger guys" -- a high roller who could afford to bet very, very big.
But he wasn't that at all.
In fact, Rubalcada was a faceless grunt in the most successful gambling enterprise of all time.
The divorced father of two says he was paid $1,200 a week by an outfit called ACME GroupTrading to sip vodka tonics or Bud Lights until a man he'd never met called or texted him on his small Nokia phone. The voice would tell Rubalcada, known as Lubbock, how many thousands of dollars to place on which games -- immediately. But the ultimate orders came from the greatest and most controversial sports gambler ever: William T. "Billy" Walters.
For almost four decades, Walters, now 68, is thought to have bet more money more successfully than anyone in history, earning hundreds of millions of dollars. Federal and state investigators sniff around his operation regularly. Scores of bettors and bookies have tried to crack his methods so they can emulate him. Even Walters' employees, like Rubalcada, have tried to figure him out so they can win alongside him. Walters has outrun them all.
What's clear, according to dozens of interviews and thousands of pages of legal documents, is that Walters beats the odds everywhere -- in the stock market, real estate, criminal proceedings and his true wheelhouse, sports gambling. His talent brought him from a life of poverty in rural Kentucky to one of wild success. He owns a fleet of car dealerships, several high-end golf courses, a private jet and fabulous homes in places like Palm Desert and Cabo San Lucas.
Walters, according to interviews, is both kind and a bully, charming and scheming. He maintains an opulent lifestyle but also gives lavishly to charity as well as to presidents, governors and city council members. Among everyone who knows him -- from employees to bookmakers to politicians and investigators -- he elicits admiration, fear, jealousy and consternation in nearly equal measures.
Walters doesn't want to talk much about any of it. Over a period of several months, he begs off numerous interview requests from ESPN, saying his lawyers want him to keep quiet because of the latest investigation against him. Last spring his name appeared in headlines next to those of pro golfer Phil Mickelson -- a friend and occasional golf partner -- and billionaire investor Carl Icahn as targets of a federal insider-trading investigation. Walters has denied wrongdoing, but the case remains open.
In one phone call, he rails about his distrust of the media. Another call ends with a referral to his attorney. In late January, he answers a few questions and agrees to a fuller interview if ESPN agrees not to ask him about the insider-trading case or write about allegations that he once provided information to an FBI agent. ESPN declines the offer.
Walters repeatedly voices frustration over what he says are draconian laws that drive a massive market of betting dollars to poorly regulated offshore locales. "Why don't we license this thing?" Walters says. "Why don't we regulate it? Why don't we keep the bad guys out? Why don't we generate some jobs?
"Everybody in this country bets on sports," he continues. "If you were to take away the office pools, fantasy pools, people betting on sports, I guarantee the [TV] viewership would be cut in less than half."
Even when there isn't an FBI investigation looming, Walters isn't willing to talk much about his work. One of the few times he got close -- in a fawning 60 Minutes profile in 2011 -- he spoke only in general terms. He's been more open about his background: He was born to a poor teenage mother in southern Kentucky and raised primarily by his grandmother until she died when Billy was 12. He started placing bets as a young boy and lost all his savings -- $75 -- betting on the New York Yankees against the Brooklyn Dodgers in the '55 World Series. In the early 1980s, he left two failed marriages and a car salesman gig behind in Kentucky and, after a misdemeanor gambling conviction, headed west with a tiny bank account and a heavy drinking habit.
After arriving in Vegas, Walters connected with the people who would help him turn his gambling habit into a career. In 1980, Dr. Ivan Mindlin and Michael Kent had formed the now-legendary Computer Group, which pioneered the use of computer algorithms for sports betting. Mindlin was the self-indulgent frontman, a surgeon-turned-gambler. The technology was run by Kent, a mathematician who developed nuclear submarine technology. By the early '80s, the Computer Group had burgeoned into the first national network of sports bettors, betting hundreds of thousands a day. The collective's gamblers, handicappers and investors began earning millions.
Walters was inexperienced, but Mindlin was impressed with his moxie and recommended his hiring to Kent in 1983. Walters was tasked with exploiting the weakest betting lines with bookies, then eventually with moving millions every week in exchange for a cut of the profits. "I gave birth to him," Mindlin says.
After a few years, Walters quit drinking for good and morphed into a member of Las Vegas' influential elite, developing golf courses, subdivisions and industrial parks. According to Jack Sheehan, a longtime Vegas pal, Walters fancies himself an across-the-board genius whose business acumen stretches far beyond betting. He prefers being seen as a successful entrepreneur, the friend says, not just a "Las Vegas gambler." Walters says sports betting now occupies less than 7 percent of his time.
But gambling was how he made his name. "I can tell you nobody has ever approached sports betting with as much analysis, as much technical capability, computer analysis," says Sheehan, who is writing a biography on Walters. "And he has a work ethic that is just ridiculous. If you and I had $300 million, we might play golf five days a week. Or be on a beach with three gals in bikinis. Billy works as hard today as he did when he was a used-car salesman in Kentucky."

Saturday, 21 March 2015

Down And Out

It's been a tough start to my Spring Break. The weather may be sunny and warm (in the 20Cs) but my mood is anything but.


My inbox was bursting at the seams on arrival, and buried among them was an FTL entry from Randolph, the Erskine Cup semi-finalist matched up against the XX Unders. On closer inspection, all Randolph's selections were Friday night games, twelve of them. 'Let me have a look', I thought - big mistake, holiday ruined -

All twelve won!

My Erskine Cup hopes are all but over, and my holiday may have to be cut short. This will probably be Randolph's biggest win of the season too! Seriously, hard to beat a 100% strike rate. For the record, the selections included six lays, including a lay of the draw.

On this topic, (calculating Lay wins), the esteemed Joseph Buchdahl commented that:
I think there is a correct number. It's the +0.5 Pinnacle handicap. That is backable in the real world.
I just don't see the point in showcasing impossibilities. You don't need to use a special formula, just the +0.5 handicap odds.
That would indeed be the perfect and correct number for laying either of the two teams, assuming the price was extracted at the same time as the others. Unfortunately, the +/- 0.5 Pinnacle Handicap prices are not included in the Football Data spreadsheets, or at least not that I have seen, so where is this correct number to be found on a Monday when the results are updated? 

The other problem is that there would still be the issue of laying the draw, as Randolph (lucky devil) wanted to do in last night's Troyes v Creteil match.

Joseph continues:
I just don't see the point of applying artificial formulae when real world odds are available to do the job. 
The problem is that real world odds are not (to the best of my knowledge) available anywhere. The point of an artificial formula is to allow entrants who like the Lay option to be able to join in the FTL. For some, having a formula to allow Lay bets was a pre-requisite for entering. 

Referring to the Bundeslayga system, Joseph writes:
Your Bundeslayga also appears to use some sort of formula (the same one?). This is a subscription service. Results for 13/14 are quoted as 4.34% from 96 selections. But applying real world prices (I've assumed a Pinnacle's typical Bundesliga Asian handicap market would be a shade over 1.02 margin) I get a yield of 2.1%.
The formula is the same - I run the FTL and track the Bundeslayga. The Bundeslayga is not a subscription service. It is a (now) well known 'system' first identified and written about in this blog in 2010. It's been mentioned many times since, and it's a great Premium Charge mitigation system which shouldn't continue to be profitable year after year, but somehow is. There is no model generating the selections, although I have started looking at adding a filter with the goal of improving results, and while I do include the selections and update the results as part of the XX Draws / latterly Cassini service, no one (at least no one in their right mind) pays for Bundeslayga selections since all they need to do is look for Home, odds-on favourites in the Bundesliga. No one buys a system that has already been given away. The results from the individual 'Cassini' systems are also clearly broken down, week by week and by season to date. Any subscriber can attest to that.

I do document the Bundeslayga selections and results because I bet on these matches myself, I noticed the opportunity, and I'm interested in seeing how long it will keep giving.

It's a freebie and I have no idea if anyone (subscribers or others) bets on these or not. My members are incredibly quiet, which is usually a good sign. People join the service for the XX Draw selections which have been around almost as long as I have, starting in 2010 with the 'strong draw' idea.

As I showed in my screenshots from last weekend's actual bets, (and what could be more real world than these?) for the Bundeslayga an average of 0.025 has been close to real world experience, which is why that formula was chosen. (I used to use +0.04 and found that was shorting my results). [I admit here that I have no if this average is reasonably realistic for Lays in other leagues]. Here they are again:

The four differences were 0, +0.1, +0.4 and +0.5, which (purely by chance) averages out at... +0.25. The number was based on empirical evidence rather than plucked out of the sky. I suppose I should include the 50% Super Premium Charge too. Joseph continues:
You just set yourself up for criticism that you are deliberately trying to inflate your returns (even if that is not what you are doing), which is why I've ended up commenting here.
Since the FTL formula applies to everyone the same, I have to presume this comment refers to the Bundeslayga, in which case I'd ask why on earth would I bother, when the system is publicly known, generates a lowly maybe 40 to 50 winners a season, and selections going back several seasons can be verified by anyone using whatever formula they want to apply? One person has criticised this calculation in nearly two years. There are 750 Cassini selections to date this season, of which just 32 are Bundeslayga winners. If we deflate those 32 by making the average adjustment a less real-life (see screenshots) +0.04 instead, the return drops by less than 0.04 points per winner. Seriously? I think the way my bets are recorded is scrupulously fair. Many services choose a far more arbitrary way of deciding their official prices. Every one of mine is verifiable, and has been since 2013-14.   

If the Bundeslayga interests anyone, they can run their own numbers using whatever formula suits them and see if it is worthwhile. 

Choosing  a simple system generating just 46 winners a season for 'inflation' would be a very strange thing for me to do. The Bundeslayaga is not claimed to be a big money-spinner, (if I thought it was, I'd have kept quiet about it!) but it does have extra 'churn' value if you are paying either the Premium Charge, Super Premium Charge, or may be doing so at some point. 

Having said all that, I know an average spanning 1.3 to 2.0 is not ideal. I'm open to better ideas, bearing in mind the limitations of not having +/- 0.5 Pinnacle Handicap prices available.

For what it's worth, the Bundeslayawayga is more profitable this season, (+11.09 from 13 winners due mostly to Borussia Dortmund's troubles) but I am, of course, inflating the numbers (kidding) - here they are using the same +0.025 average adjustment:
I'm expecting the profits on both systems to descend towards zero as the season winds down anyway.

One subscriber wrote:
You should probably just take lays out of the FTL altogether, it will save you a load of grief
A 'load' of grief? One person has complained, although out of 28, that's a high percentage I guess! I'm not even sure why the complainant complained since they are not impartial to using the Lay option themselves.
For official recording of the Bundeslayga, if you can't get the price from Betfair that the game goes off at KO, then maybe ask your members what price they took and take an average and assume 5% comm.
Or maybe proof to the SBC and get them to do it!
Even if I could get the price from Betfair at KO, can you imagine the outrage that would ensue for matches where the price movement between Friday afternoon and kick-off (sometimes more than 72 hours later) meant the Lay price was lower than the Back price? The scandals of Profumo, Watergate and er, Hillary Clinton's emails would all be mere footnotes in history in comparison.

Also, the formula applies to all FTL eligible leagues, not just Gerrmany, so asking for input on Bundeliga Lays would be of limited value for the other leagues. I imagine lays in less popular leagues would need a larger average adjustment. 

The SBC (Secret Betting Club) I have little knowledge of (the 'Secret' always puts me in mind of Enid Blyton for some reason, and seems a bit silly when they promote themselves so actively on Twitter) but asking them how they proof Lay bets in general isn't a bad idea. And as mentioned earlier, the selections are not the product of a model so they can proof at will - the only variable would be how they decide which selections around the evens point are qualifiers. I personally have lays that are unofficial because the Pinnacle price is above 2.0.

The negativity floating around was alleviated somewhat by a positive comment from an Erskine Cup semi-finalist (Gecko - not his real name - is it?) still in with a chance of winning, unlike me - he wrote:
ps read your comments and hope you do continue this, I appreciate the amount of work (I was a mod on OLBG for many years and scored a number of comps on there, took huge amount of time and often wondered why I bothered!) but some people really do appreciate it and I'm in that category so thanks very much for all the hard work

hope I can carry on my decent spell and if I manage to win anything I will hopefully be able to donate some back to next season's pot

anyway, good luck to you and others this w/end and I'm very wary of the pommie underdog lol
I'll decide on next season later. I suspect this season's bonus contributions from Ian Erskine and Betcast, and the bounty offers from Football Elite, Football Investor, Skeeve and The Football Analyst may be a one-off, but I appreciate the comment.

Graeme, proprietor of TFA, was bemoaning the general misfortunes for services right now:
Hard to keep updating the ratings at the moment if I'm honest but I've been here a few times before. No doubt we're due an upturn again. If the other main ratings services were coining it in at the moment, I'd be seriously worried about my ratings. From what I'm seeing from blogs and via emails, the other ratings services are having an identical run to TFA. Not much good for anyone following all of us in a portfolio unfortunately but it shows it's a wider issue here than just my ratings. It's always hard to explain losing runs like this and I'm not sure anyone can.
I know how he feels, since the Cassini Value Selections are having a poor season after a strong performance last season. The strange thing about last season was that the low number of draws meant a loss on the XX Draws, while it was Away wins that appeared to make up the numbers. This season it's the reverse as the draws are back, but Homes are vanishing! Yet another reminder not to read too much into small samples. 
As for the question of where are the draws going, and in support of my “diminishing home advantage at the top level” theory, the answer is that they are almost all becoming Away wins.
Of the 63 Draws ‘lost’ this (2013-14) season (based on 9 year averages because Serie A changed increased its membership in 2004-05), only one of them became a Home win, while a massive, and almost unbelievable, 62 (98.4%) became Away wins.
As Danny suggested a few weeks ago, Jamie also, not unreasonably, suggests that the Cassini Value Selections may have been beneficiaries of the draw drought. I accepted this possibility at the time, (obviously fewer draws means more homes and aways) but having looked at the numbers, since these value selections are almost always home selections in the 2.0 to 2.6 price range, it is obvious that if they have benefited at all from the dearth of draws, it is only by the minutest of margins.
The fact is that this season, anyone selecting anything other than away teams, would have struggled to meet expectations. The decline in draws is only one piece of the puzzle. The other side of this is why the Away win numbers are so high this season?
For those who like numbers, here are the 9-year averages across Europe, followed by 2013-14's results:
Away wins were at record (10 year) highs in England, France, Italy and Spain. Only Germany failed to set a new high, although they were still 8 wins above their 10-year average of 89.
With still a little over 25% of the season to go, (818 played) the Homes, Draws and Aways are on course for the following totals:
The recurrence of 818 is a coincidence. Homes are down again (spectacularly in Serie A), Draws are returning to their long-term average, and Aways are down on last season's high, but continue to push the average higher each season.
It's Happy Birthday blog today, in case any of you missed it. Seven years, 1960 posts, 3463 comments, and over a million page views. I must be doing something right, even if the last year has been rather dominated by FTL updates. Apart from the time it takes, one other downside of running the FTL is that time that might have been spent on general blog posts is now taken up with posts on the FTL. While this may be fascinating to the 28 entrants, I’m told by friends that it doesn’t make for a particularly riveting read for the vast majority of viewers.

Although I've not used them for many years, at one time I did have funds invested with WBX )World Bet Exchange) and they featured in the banner for this blog. This week saw their demise however,
A note from the CEO, Malcolm Gray:
To our loyal members and affiliates,
I have always taken a long view with WBX and believed that we'd become a genuine alternative to Betfair on the strengths of our great platform and commitment to putting every exchange player back in control. Unfortunately the business strategy is no longer viable due to the rising costs of regulatory compliance and levies in the United Kingdom, along with the increasingly prohibitive stance of various betting jurisdictions.
Though I am of course disappointed that WBX is closing, I am proud of my team and the fairness in which the exchange has operated for almost 10 years.
Please rest assured that WBX will continue to service its Members and affiliates in the same spirit during the wind-down period.
Thank you for your support.
And finally get well soon wishes to Football Investor Stewboss.
I am currently taking a break from the service as I have been certified unfit to work by a doctor. I hope to return on the 24th March and the next update will be 26th March.

Sorry for any inconvenience caused.
Stew's ailment doesn't sound too serious fortunately, but it explains his inactivity this weekend.  

Thursday, 19 March 2015

FTL Update 18.March

Are midweek rounds tougher to predict than weekend? It certainly feels that way, and something to look at after the season perhaps, but this midweek round cost the FTL as a whole another 23.74 points, and with only eleven entries active, that was relatively high. There were only two winners - the big winner was Mortimer, who recovered from a slide into the red at the weekend with a strong rebound back into the green after gaining 3.87 points and five places, and the other was @ValueBankFooty with a more modest gain of 0.66 points.

Less successful were TFA_Raz who replaced Mortimer in the red after losing 5.92 points and dropping seven places, Graeme's TFA Bounty Entry Draws which lost 5.00 points and Sjosta who fell to last place after losing 6.46 points.

BettingTools.co.uk lost 2.41 points, but hangs on to the lead, while Fairfranco and Football Elite both lost 3.00 points. Talkies Tips lost 2.00 points, while Mountain Mouse and Drawmaster lost peanuts.

The 16 entries remaining in profit are:

The bottom half of the table looks like this:
With an international break coming up, the March monthly competition is closer to wrapping up than you might think. The leaders are:
No one has yet won two monthly prizes, but Mortimer and BettingTools.co.uk can both so just that if they can pass Gecko who may be distracted by the Erskine Cup, because this weekend the semi-finals are here. The two ties are:
With three of the top six competing, the line-up is stronger than at one time appeared likely, and Talkies Tips is everyone's favourite as a big underdog representing England against Australia. 

Tictac left a comment, asking:
Hi, what are the rules for the FTL, for e.g. which bookmakers are/aren't allowed, and what is the time range in which you can take odds (e.g. are opening lines allowed)?
I would possibly be interested in joining next season, who is/isn't allowed to enter?
Thanks!
The basic rules are detailed here, but we're getting ahead of ourselves since I'm not sure what my plans are for next season yet and Tictac's planning his tactics already. With most weeks having action at the weekend and in midweek, there's a lot of work involved. Five or six hours a week is probably not too far off, and although I quite enjoy this kind of thing, it does take up a lot of time.

My impression is that at least some people have enjoyed it. The top four were not household names before, and that certainly hasn't changed! From those who run services, feelings are probably mixed. For myself, the current success of the XX Draws and the disappointment of the Value Selections is a complete 180 from last season, when the Value Selections topped the table and the Draws lost nearly 33 points. I can't speak for Matt, Stewboss, Graeme, Skeeve or Peter, but I'm guessing they were hoping to be a little higher by this stage of the season. Still a few weeks to go, but unless things change quite a bit, I'm not sure the Bounty offers will be as generous next season if there is one.

Other changes that might go before the Rules Committee could include a bonus for the top places so that the incentive to place higher is greater, and whether to continue to use Pinnacle's prices given their exit in mid season from the UK market.

Whether or not to accept Lay bets, or how to settle them will also be looked at. While the formula for settling odds-on selections was based on the real prices attainable on Betfair (recent examples from the weekend are lays of Moenchengladbach at 1.74 on Pinnacle laid at 1.79 on Betfair, Eintracht Frankfurt (1.73 and 1.74) and Augsburg (2.0 and 2.0). 0.025 has been a reasonable average over the years, but using a flat average obviously isn't ideal.

The much respected Joseph Buchdahl, author of the excellent Fixed Odds Sports Betting - Statistical Forecasting And Risk Management commented:
"a winning lay at 2.00 will be recorded as a profit of 0.98 points."
Pinnacle Sports don't offer margins of 0.5%. If -0.5 is 2.000, a typical price for +0.5 will be 1.9 to 1.92 for a high volume market.

You could never make a profit of 0.98 from a 1 unit stake betting on a football market where one option is priced at 2.000 at Pinnacle.
Saturday's timely example of Augsburg at exactly 2.0 shows that there is no right or wrong answer here. Augsburg were a Bundeslayga selection (just) priced at evens with Pinnacle as a back, yet also layed at exactly 2.0 on Betfair:
I think Joseph may think I am in the business of verifying systems, but the FTL is more concerned about relative results than actual results, although the closer to actual that they are, the better. The other issue with Lays is that they are only possible on exchanges and we should really allow for commission if we were looking for actual returns, but for FTL purposes, the Lay is an abstract calculation.
To give you some context and background here, the calculation you are referencing is from the Friendly Tipster League – it’s a season long private (28 individuals), friendly contest and is not, and is not intended to be, a verification service of any kind reporting precise numbers. I leave that kind of thing up to you!
Several established tipster services are participants, but the expectation is not that their official returns will match those of the FTL. Rather it affords them the opportunity to show-piece their talents and the merits of their selections relative to others.
You refer to the settlement of a Lay bet, and unfortunately as you know, because Lays are only available on exchanges, Lay prices can only ever be approximations derived from the published back prices.
There is thus no “correct” number, but the formulae proposed (based on empirical evidence) were accepted by everyone entering at the start of the season - although following one nightmare of a weekend when one entrant submitted close to 100 bets, many of them higher priced Lays, I realised that simply adding 0.05 to the back price for these wasn’t adequate, and the rule for these was updated, published and implemented moving forward.
Again, the Profit and Loss numbers are used solely for comparison purposes within the FTL rather than intended to reflect accurately what returns could be expected in the real world.
Indeed, Pinnacle’s prices can often be beaten, and perhaps shouldn’t be used given their exit from the UK market in mid-season, but my opinion is that they are a decent benchmark and the fewer changes that are made once a competition starts, the better.
The settlement rules are not, and never can be, perfect if we are allowing for abstract bets other than Backs to be included. Unfortunately there is no industry standard for recording bets, even for traditional Backs, never-mind artificially derived Lays. However, the demand to allow Lay selections in the competition was apparent, and a simple (if imperfect formula) was proposed and accepted for use in the FTL. The same rules apply to everyone, and are certainly (at least in my opinion), an improvement on earlier seasons, and I’m sure they will continue to be improved upon if the FTL continues in future seasons.
If anyone has any other constructive suggestions for a returning FTL let me know. I'd also like to know if the Erskine Cup idea should be continued, and also the monthly prizes. Perhaps four quarterly prizes would be better? Any sponsors interested? This season, Ian Erskine and Betcast generously chipped in but then went very quiet early on.

Tuesday, 17 March 2015

FTL Update 16.March

Another losing round in the FTL with a modest 7.40 points lost overall. The number of entries in profit dropped to 16 as Mortimer dropped into the red zone, and there are another five entries clinging on to a profit by less than two points. More than half the entries (17) are in single figures, either up or down.

The big weekend winner was BettingTools.co.uk who moved to a new high after Burnley's win over Manchester City boosted his profit to 10.13 points. He regains first place from Gecko who drops to second despite making a solid 3.81 points. Randolph and Trend to end chose to stay idle, the former presumably resting before his tough upcoming Erskine Cup semi-final, and remain in third and fourth places respectively.

Moving up one place to fifth, XX Draws made a second highest profit of 5.58 points with four winners out of eight picks, and on a good weekend for draw hunters, Drawmaster made 3.55 points, TFA (Bounty Boy) made 3.06 points, TFA Euro Draws made 2.00 points and only Draw Picks let the draw portfolio down losing 4.42 points.

The profits from draws were a pleasant surprise. In the top five leagues, there were only 10 draws (from 49 matches) which usually bodes ill but the XX Draws found the only draw in France, one of two in Germany, and two of three from Italy. TFA Euro Draws also found the solitary French draw, plus one of two in Spain, and all three in Italy (from 17 selections).  In the Lower Four leagues in England there were 17 draws (from 48 matches). Drawmaster found his four winners (from 9 selections) here as did TFA Draws with three winners from 7.

The remaining Bounty Boys all had losses. Football Elite were down just 0.40 points, Skeeve lost 1.14 points and Football Investor lost 5.43 points. Combined liabilities remain at £1,725.
The leaders in profit by 10 points or more, a group of just five now, are:
Others in profit are:
XX Unders recovered some of last week's losses with six winners from eight, and moved up four places as did Online Trader. Club Havana was the big loser losing on all 8 selections and dropping three places as did Draw Picks. Jamie A's weekend wasn't the best losing 2.96 points but the others all had modest losses, or in Fairfranco's case, a modest gain.

The down by less than 10 points group is led by Mortimer who lost 5.06 points and four places, and the winner here is Graeme's TFA Draws who made back 3.06 points and climbed two places.
No prize money on the line for any of the remaining entries but signs of a pulse from Drawmaster, and TFA Euro Draws while Sjosta (-5.94) and Football Investor struggled. Paul Watson came up with a new strategy of looking for Draw Trebles. One actually hit, but unfortunately a treble that would have made over 39 points in profit included Newport County, who had already played Cheltenham Town on Friday evening before the email was received. I'm not sure if this game was initially scheduled for the Friday night, and while I am sure it was an innocent error (always assume best intent) it's a reminder to double-check kick-off times. As always, I run this friendly competition as fairly to everyone as I can, so the bet was settled as a winning double.  
With the end of the season in sight, a reminder to have your selections in prior to kick-off for Tuesday and Friday matches, and by midnight (or close to it) those nights for other matches.  
More games coming up this midweek in the Lower Four, so another update after those results are settled.

Statwatch: Some more observations on the nine most followed leagues.

Draws in Serie A are at 34% compared with an average over the previous five seasons of 26.1%. There have already been more draws this season than last, and we have 112 matches to go. With 91 matches remaining in the EPL, the Draw count is just one behind last season's record low

This season's Highs since 2003-04:

Draws in Serie A (34%)
Away wins in La Liga (30.7%)
Home wins in Ligue 1 (48.1%)
Away wins in Ligue 1 (27.7%)
Away wins in League One (35.3%)
Away wins in all nine leagues (30.3%)

And Lows:

Home wins in Serie A (38.1%)
Draws in Ligue 1 (24.2%)
Draws in league One (24.3%)
Home wins in the Big Five (44.8%)
Lowest Ratio Home Goals to Away Goals in Serie A (1.24) and League One (1.08)