Equalising Formula

L commented on my past post about levelling up your profits (or losses) writing:

Nice stuff

Is possible to specify mechanically the risk reward ratio of bets in this case variable hedge? For example, at twice the rate sell 50% stake etc...

It is an easy calculation to determine how much you should lay to level up your position. Basically it is the difference between your green and your red, divided by the price, so if you are in the position I was in, Warriors -£600, Thunder +£2,300, Thunder price 1.6, the formula of (Difference) / Price tells you the amount to lay is £1,812.50.

Difference = £2,900 / 1.6 = £1,812.50. 

If you lay this amount, your red of £600 on the Warriors goes to a green £1,212.50, while your green on the Thunder reduces from £2,300 to £1,212.50.

It's easy to set up a simple spreadsheet to input your three variables, and output the amount to lay.

Ellis Rayner said:

If you believe 1.6 was a fair price then how about loading your greens to reflect that? 1.6 = 62.5% so based on your figures about £1,420 and £850

I'm not sure what Ellis is suggesting here by 'loading my greens'. Sounds like something my Mum used to do with my dinner. My What If? formula shows that laying £1,450 would leave a situation close to what Ellis suggests, but when you leave your outcomes significantly different with a fair (correct / true) price available, one is in second-guessing territory, which is a land I am trying to leave.

Second-Guessing

A frustrating Thursday night / Friday morning in the NBA with the Oklahoma City Thunder giving up a game-winning shot as time expired in a loss to the Golden State Warriors which cost me about £900. Well, £450 after Super Premium Charges are taken into account – not that there is anything ‘super’ about them.

One thing that I always seem to mess up on is how to close out a winning situation. As is my wont, I was laying the Warriors at what I felt to be too short a price, and my opinion appeared to have been proven right when the Thunder completed a comeback from 14 points down with 7 minutes left to lead by one with 2.3 seconds left.

At this point I am in the red on the Warriors, decently (by my modest standards) large green on the Thunder, so what to do? The Thunder at this point were trading at about 1.6.

Basketball is perfect in these situations, because there’s a time-out when the market settles down and you have time to make a decision, for which there are a few options:

A) Take no further action

B) Reduce or eliminate the current risk on the Warriors

The position I was in, using approximate numbers, was Warriors -£600, Thunder +£2300, Thunder price 1.6.

Now that’s a decent win if the Thunder hold on, but do I want to risk losing £600 in the event that they don’t? That’s a rhetorical question, because the answer is that I am not in the business (albeit part-time) of gambling, and preserving my bank is important, so the question becomes, “is that price of 1.6 value”? 

If it isn’t, then the right decision is to take the value, whether that be backing or laying, but if the price is close to value, and in my opinion it was close enough, then I want to at least eliminate the downside. So first my numbers became Warriors £0 Thunder £1,940, but then I want to at least show a profit on the event, which means giving up £60 from a Thunder win to buy £100 on a Warriors win. Continue this through to the pivot point and you end up with approximately £1,200 on each side.

Unfortunately for me, as the screenshot shows, I didn’t lay enough, stopping at the £600 mark. I've noticed this before in myself - it's almost as if I don't want to be disrespectful to the team that got me into the winning position! 

It seems like when I level up, I would have been better off leaving the bet to run, and when I don’t level up, I would have been better off had I done so. Of course it only seems that way, but assuming the price is close to fair, following a rule to always level up would take away the second-guessing afterwards. 

Still in the NBA, and NBA Tips are still under water, although the bleeding has stopped for the time being. Still down on the season by 16.76 points (ROI -26.6%) though. 

Like the Peugeot lion, John Walsh goes from strength to strength in the NHL, with a push and two winners since the last update, and now up by 10.02 points (30.36%). His NFL picks aren't quite so good yet, although he did eke out a very small profit last weekend.

Model Bettors

In my Bad Bettors post a few days ago, I wrote:

In other tipster updates, John Walsh had a quiet week in the NHL, but another profitable one. Interesting to note that after backing in Regulation all season, the last three selections have all been Moneyline bets. I am trying to ascertain why the change and will update you if I get an answer. After 30 bets, an ROI of 28% is very good.

John did indeed reply, and here are his words of wisdom on a few topics:

It's quite an honor to be in a post titled 'Bad Bettors'.

First I will address why the change from regulation betting to moneyline in my NHL bets. The simple answer is to get value. There are sections of the season in which certain teams play to overtime more frequently and of course the opposite is true. If I can get more value (from my model) then it only makes sense. My current feeling is that the implied probability from the odds haven't caught up to this in some cases. Last season was really different with the lockout shortening the NHL schedule.

I'm not defending NBA Tips methods. I don't know who it is or how he/she comes up with selections. This person may have an edge, there really is no way to tell with such a small sample size. That is one of the toughest things about sports betting, getting a large enough sample size to know if the system has an edge.

I think being a fan of the sport is necessary to be successful at betting on it. It's like the engine on a car, you need the motor for the car to work but you need the steering wheel, brakes, etc.

Betting on (American) football determining team weaknesses, changes in personnel, etc. are the other parts of the car. At the same time you can have a system that has worked in the past but if you don't know why it has, you could be in dangerous territory. You have to have the motor in the car for it to work. I've played football, coached football, watch every NFL game. The NFL network is on my television all summer. NBA Tips could be the same with a system that accounts for every factor.

Public knowledge is different now than even 5 years ago with so much information being accessible from a phone. My NFL model didn't suggest Tampa Bay was value in the game that just finished. I really thought the implied probability wasn't taking the whole Dolphin offensive linemen situation as enough of a factor. Miami had a patchwork line with a split dressing room and what sounded to me like very little leadership. Not to mention the distraction for the game. Although, the second half of this game is the most impressed I've been by the Dolphins all season considering the previous factors coupled with being down at the half on the road.

The Tampa Bay Buccaneers v Miami Dolphins game was more than a little frustrating for me. I laid the Buccaneers after they went up by 7, and the Dolphins then played about as bad as I have seen a team play for most of the remainder of the half. Usually I would be looking to lay the leader at the half, but I couldn't see anything but a runaway Bucs win. Thus it was with mixed emotions (surprise, relief and also a little annoyance that I'd missed an opportunity), that I was able to trade out for a profit when the Dolphins played surprisingly well in the second half. In the end, the Buccaneers retook the lead and held on to win.

John's NHL model would appear to be working, although a new bet type for this season (totals) was a push on Tuesday night. He mentions NBA Tips, and 62 bets is indeed a small sample, so we'll revisit their results later in the season. The bet selection process isn't working so far, and it's a strange model that finds value in bets as diverse as "first to 20 points", " highest scoring quarter" and various individual player totals, but we'll see how it all settles down. I would certainly be most impressed if these selections proved to be profitable long term. Incidentally, they recorded a profit of  57.35 points last season, but I didn't follow them until the end of the season when they must have hit a bad run, because I posted a comment in June, and never received a reply:

I am not sure where these results were proofed, but they read this blog avidly now and will let us know I'm sure. The Heat Rebound post shows how an example of how an edge can sometimes be found pre-game, and while I would love to be proven wrong, I'm just not convinced that NBA Tips have a genuine long-term edge. More than a month ago, I wrote the following, and I haven't seen anything since to change my opinion:

Fortunately, I am not following NBA Tips with any money. The error in the total number of games in a season didn’t fill me with confidence, and without meaning to appear rude, they appear not to have any idea about the purpose of pre-season games. Theses games are not played anything like a regular season NBA game. Rare is the coach who gives a darn about the result. One notable exception last year was Golden State Warriors coach Mark Jackson who said he wanted to win, and they won all six I believe. This year, with a proven winning team, he has said that he doesn’t care about winning, and has joined pretty much every other coach in using the games for establish the playing roster, easing players with previous injuries gently back into action, and trying out a few new ideas and tactics. Play tends to be sloppy, especially the first couple of games with timing off and players just not game ready.

I did try to point out via Twitter to NBA Tips that their selection of Over 201.5 for Sunday night’s Los Angeles Lakers v Denver Nuggets game might be suspect with no relevant data, and their response was, let’s say, a little naïve and of course, no supporting data.

“With regards to tonight we feel like this will be rather open in style and Nuggets will certainly help this. Nash / Gasol also aid the open D we expect tonight. We feel if our followers can get around 1.8 at 201.5 this is value especially as Bet365 have pitched the line at 205.5.”

"We feel...". Based on what? Steve Nash and Pau Gasol were indeed back, but returning from injury they are most certainly not match fit, and are most certainly not going to play for too long in meaningless games. In the event, Gasol and Nash both played for less than 22 minutes, with Nash failing to score a single point. Both players had minus ratings, -14 and -5 respectively.

NBA Tips mentioned Denver’s scoring average last season of 207 – what they may not have realized is that Denver had a new head coach on the bench – former Lakers star Brian Shaw.

The style a team plays is that of its head coach, and last season’s average total in regular season games has absolutely nothing to do with a first pre-season game.

Pre-season games can offer value opportunities, but backing Overs pre-game at a recommended 1.8 (in fact only £6 was traded pre game on Betfair at a high of 1.75 so at least no one appears to have followed this poor advice) is not one of them. The game was apparently poorly played with the two teams combining for 50 turnovers, and the point total of 172 was well short of what was required.

Emp posted a comment on my Mugs and Shrewdies post, and I'll try and get to those in the next few days, and Jesse Livermore shared this joke with us:

Two economists are walking down the street. One says: "Hey, there's a dollar bill on the floor." The other says: "Impossible. If it were real, someone would have picked it up by now."

Mugs And Shrewdies

Cobo wrote:

Thank you for addressing my doubt in length Cassini.

It's true the markets pre-off are very very accurate and difficult to beat. At least in tennis they are, and they have improved even more these past three years. This is a fact. Or as some people say, a true fact.

Regarding in-play, it's true lots of things can happen that make the price more difficult to assess. That plus fear and greed.

Anyway, if there are opportunities in the Stock Exchange, which has far more people and money trying to profit from it, there is no reason to think betting exchanges should behave differently (except perhaps having more opportunities from that lesser competition).

Of course, having advanced stats of historic in-play game action from a team/player will always help make a better decision.

Cheers!

P.S: And English is NOT my mother language, so please forgive any mistakes ;).

This part is rather casually thrown out there:

“Anyway, if there are opportunities in the Stock Exchange, which has far more people and money trying to profit from it, there is no reason to think betting exchanges should behave differently (except perhaps having more opportunities from less competition)”.

I would argue that while there are indeed opportunities in the trading of stocks, and there are similarities between stock and betting exchanges, there are also significant differences between the two. The biggest is that a stock buys you something tangible – a share in a company’s assets and future earnings – a genuine investment, whereas a bet is more like a traded option that will shortly expire either worthless or in the money.

In my experience, unless you have inside information, you are never going to beat the stock market long-term. The simple truth is that other people with far more money at their disposal will always be ahead of you and the price you can buy at or sell at is always the price that insiders are happy to trade at, i.e. a value price for them, not for you. I've written in this blog many times that you are not going to have an edge by looking at publicly available data.

What allows the common man on the outside to benefit from stocks is the long-term nature of the stock market. Stocks have yields which provide an income, and the stock market inexorably rises as a whole, as the economy grows. The best we can realistically hope for is for our investments to grow broadly in line with the market, which is why you should invest money consistently in index funds and forget about it. Trying to time your entry or exit into / from the market is futile.

Including dividends, but excluding inflation, the stock market's long-term return has been around 9.6%, accounting for inflation it is closer to 6.2%, and after a few years, whether you paid £4.81 or £4.82 for your BP shares isn’t going to be an overwhelming concern. The market for most stocks is so liquid (14 million BP traded yesterday) that the negative edge is always going to be small.

Several studies show that market timers typically get into a market after a period of solid growth, and having missed much of the gains they then exit after the market ‘corrects’ itself and thus miss out on the start of the next bull market. Fund managers love to say how they can beat the averages, but unless they have inside information – look up Raj Rajaratnam, Rajat Gupta, Ivan Boesky or most recently Steven A Cohen – they may get lucky for a year or even a few years, but long-term, they won’t beat the average.

There is no long-term nature for sports investing, and no income from dividends either because your bet has no intrinsic value. To make money from sports trading you need to be entering and exiting the markets at net value prices. You do not have the cushion of dividend income or the expectation that the value will rise in time.

What it boils down to is that you need to be better at assessing the true probability of an event than someone else. One advantage of sports betting on the exchange is that amounts bet are typically much less than the stock market sees on trades, and you can ask for a value price rather than have to take or leave what someone else offers you. Put your bet out there, be patient, and you only need one mug for the bet to go live.

I found this comment on the Betfair Forum in a thread about in-play trading:

Some of us are also making a living trading in-play using Sky Sports pictures. What enables us to profit is being able to price up the events as they are happening.

One of the attractions of betting in-play is that the playing field is more level in terms of information. Pre-off, a few privileged individuals can profit from their inside knowledge. Once the event starts, most, or almost all, of that information becomes public.

This, Feck, is probably a big reason why the pre-off markets are rather dry, especially in your chosen sport, which is more vulnerable than most to the effect of insider trading.

The chosen sport referred to is of course horse racing, but the contributor [Contrarian3] is on the same page as me. One comment on the thread was this:

In Running is where the mugs do their money at a far quicker rate and the shrewdies make theirs. 

Bad Bettors

The comment about losing more that would be expected on coin-flip selections is interesting. Is this really true, or do people win 50% of their bets, but at a price of 1.9 inevitably lose due to commission?

It was meant as a rhetorical question, because the answer is that overall bettors do fare worse than would be expected. Emp said...

Well for proof look at it this way; I dug up your results from last year.

Amateurs- -115.26, -9.61%ROI Over 1199 selections.

The spreads/commissions are certainly not large enough to eat up a 9.61% ROI and 1199 bets is a good enough sample to say this isn't variance.

What's more as you note on that very post, this -9.61% ROI has this positive selection effect that the worst betters were disproportionately likely to quit in the middle. Had they stayed, it would probably have been a lot worse.

This is a very rough proof, but I think it's enough to justify what to me is intuitively obvious, that there is such a thing as being really bad at betting without knowing anything whatsoever, and thus reversing those tendencies must necessarily demonstrate a handsome profit.

Unfortunately last season's FTL wasn't intended to be scientific and only covers football so the use of its results as a proof should probably not be allowed, There were no 'official' prices used, as the table was for general comparison, but other studies have shown that there really are bad bettors.

Poor money management is usually cited as the primary reason, but there is also this from a US targeted article:

Most sports bettors know just enough to make them dangerous, as there is a great deal of difference between being knowledgeable in the NFL and being knowledgeable in NFL betting.

Being able to name the starting offensive line for the Dallas Cowboys isn't likely to help a person win a bet.

What many sports bettors don't realize is that they're actually competing against other bettors who spend countless hours on handicapping, studying trends, injuries, and betting angles.

It's no secret that [American] football and basketball are the two sports that receive the majority of the betting dollars, just as it's no secret that most long-time sports bettors will say those are probably the two toughest sports to show a long-term profit in.

I can't help but think of NBA Tips when I read this. They clearly love the NBA, but love of the game and public knowledge isn't enough as explained above. They like to Tweet from the rooftops when a double comes in as it did last Tuesday (at 4.2), but down by 16.28 points after 49 selections suggests that they are falling well short of finding value, and updates to the P&L page are behind my own numbers.

I will now appear to be a complete hypocrite and say that on rare occasions, I do have a punt on a basketball game, for small stakes. Last night's game in Los Angeles between the Lakers and the Minnesota Timberwolves did stand out as a value bet with the Lakers very poor so far this season - no Kobe Bryant, and Pau Gasol posting career lows with Steve Nash also struggling - and the Timberwolves looking good with the two Kevins (Love and Martin) in sparkling form. I asked for 1.94, and as often is the case, I was matched, a little peeved when I checked and saw I'd missed 1.99 earlier. Not really. The downside of betting early in this sport is, as I have mentioned, that you can be left high and dry with late team news, so I tend to err on the side of caution and bet late with all the facts. The Timberwolves led by 24 at the end of the first quarter and never looked back winning by 23.

In other tipster updates, John Walsh had a quiet week in the NHL, but another profitable one. Interesting to note that after backing in Regulation all season, the last three selections have all been Moneyline bets. I am trying to ascertain why the change and will update you if I get an answer. After 30 bets, an ROI of 28% is very good.

John has an NFL selection tonight, so I shall update those later, but here is the latest TFA FTL table after the weekend. The selection of Sunderland at 8.97 to win v Manchester City and to a lesser extent, Newcastle United at 7.15 to win at Tottenham Hotspur led to some big moves.

Big winners were Murphy's Law (up 10.97 thanks to the North-East duo of Newcastle United and Sunderland), Punter's Friend (up 8.28 - Sunderland) and Webbo (up 6.28 - Sunderland, plus a lay of Chelsea). The big loser this weekend was Emp whose 37 selections lost 10.91 points, and cost him 7 places. Murphy's Law moves up 9 places to third (and £75), Punter's Friend moves up 5 places, and Webbo goes up 4 as does Fedslam who moves up to second place in the process (£125). Graham C moves up to fourth and £50.

Slipping along with Emp were Hofs Hackers and Fairfranco, but although Forza Fizzer lost 2.21 points, he retains first place and the £250 first prize that goes with it. My own value selections also remained static, despite a profit of 2.34 points.

The Football Analyst was up 0.07 points, but lost one place. His liability (max. £400) on the bounty increases to £325. The late Stoke City penalty at Swansea City cost TFA and a couple of others a winner, including Peter Nordsted whose official selections still had their best weekend yet this season +5.58 points. Peter's Premier Betting entries still prop up the table though. Football Elite has quietly moved into profit and up to fifth place, and a special mention for Punter's Friend who was at one time down 22.22 points, but has gained back 17.75 points in the past three weeks.

In total, up about 20 points this weekend, but still more red than green. As always, with 197 selections to update, the possibility of error always exists. If you see any palpable errors, please let me know. Should be a little quieter next week with the International break, but Skeeve and TFA will likely be active along with a few others. 

Homework

As I suspected, Emp suggests that any tone of patronisation in his earlier comment was quite unintended.

I think this may be cultural rather than a second language issue. English is my mother tongue for all practical purposes, though I speak two others. Does homework have a negative connotation for you?

I just said your research was impressive, why would that come across as patronising? Research is very important, and I don't think this is a very elementary observation; certainly, large numbers of blogs, sports-betters analysts and others haven't realized this, so it certainly is impressive, even though I guess you may have taken it for granted that anyone should know.

Interesting question on homework. For me the word has quite positive connotations, although it suggests an assigned school task rather than a voluntary adult exercise. The schools I attended (or rather, were sent to) both put a premium on achieving results and homework was part of my routine from the age of five, when we would be sent home with ten words to learn how to spell, some antonyms and synonyms and examples of their use in a sentence. 

I suspect that had my preparatory (primary / elementary) school not established the routine of homework, that I would have resented it at my secondary school, but it was in fact a useful excuse to avoid having to help with the washing up. My parents could hardly promote the importance of school work, and then deny me the opportunity to attend to it. It was also my life-saver on Saturday mornings when my parents would take a weekly shopping trip to Croydon which I hated, but homework provided me with a great excuse to avoid that nightmare too. I say excuse because I don't remember actually doing too much homework, but spent the time listening to sex offenders and paedophiles on Radio One, although I don't remember the word 'paedophile' from my ten daily words from prep school.

I guess the reasons that the comment "I am impressed you've done your homework" could be interpreted as patronising are a combination of possible sarcasm, the suggestion that I may not know what I am talking about, (do many Premium Charge players not have at least some idea?) and the inferred 'surprise' that I do, and also maybe that calling the research 'homework', suggests talking to a child. Anyway, a breeze in a teacup and I really didn't assume any insult was intended. Research does sound a much better word than homework though!

Emp followed up with:

I do think that with regard to both in-play as well as pre-game, you are falling a little into the difficult/impossible conflation.

It's true that very very few people will have edges against a market dominated by smart money, but that isn't the same as impossibility.

The ease with which countless bettors manage to lose well over what they would with a random coin-flip selection method and level stakes implies that it must be possible to be profitable without inside knowledge.

Indeed if NBATipsGuy is as consistently bad in the future, his tips are as good (and probably better) than any of yours or TFA's.

It's obviously essential to contemplate what your edge might be, but I think it's also a mistake to ascribe deity status to analysts simply because they are running a hedge fund, have fancy computer models and swing a big line. Plenty of them manage to lose as well.

The comment about losing more that would be expected on coin-flip selections is interesting. Is this really true, or do people win 50% of their bets, but at a price of 1.9 inevitably lose due to commission? NBA Tips are mostly made at around the 1.9 mark, although some Individual player bets are as low as 1.72 which is a recipe for disaster. You need one heck of an edge to make a profit on novelty bets backing at those odds. NBA Tips continue to lose as I expect for reasons previously detailed.

Sporting Differences, Innate Ability

Cobo asked:

Hello Cassini, Are you sure that "advanced not publicly available stats" reasoning can only be made for NBA? I don't know, it may be the same in baseball, hockey, tennis... At least I really don't know where to get info like distance a tennis player run on each match, placement of shots, etc... and that is being tracked for sure. Of course, in that case the effect anything has on the player is even bigger than in basketball. I talk about tennis as it is what I trade, but the situation may be the same on other sports. Also, as much as I completely agree with your post on numbers VS intuition (people not relying on numbers), I certainly think there are people with an innate ability to price risk and take advantage from it consistently. Needless to say, these are just a few compared to the number of "pro" tipsters giving advice. Coming back to the NBA, if they have those advanced stats they for sure should be able to track the price in play far more accurately than you or anyone else.

I don’t believe I ever said that, and I certainly never intended to imply it. There may well be other sports it applies to as all sports are different, and need to be looked at individually. I just happened to be comparing football and the NBA in response to RandomMan’s comment, and because those are two of the few sports that I actively involve myself with.

Each of the three sports mentioned, baseball, hockey and tennis have their unique characteristics. Baseball is laden with statistics probably more than any other sport, and it is a series of short plays from a limited number of states. I doubt that there is any meaningful data in this sport that is not publicly available. Hockey has similarities to football (team sport, low-scoring), but the all-important power-play is a big difference, as is the importance of the goal-tender. Not as important as an all-star player in a basketball team though, a sport with a similar number of players but very high-scoring. Tennis (specifically singles) is an individual sport, and obviously anyone in the player’s inner circle, including the player themselves, is at a huge advantage, and as Cobo says, even more key than in basketball. Tennis is also technically the same sport whether played by men or women, but everyone knows that in practice the two are essentially different sports and need to be treated as such from a betting perspective. The inner-circle comment applies to horse-racing too, and here we have a sport where connections know a runner isn’t going to win, and it's an accepted part of the sport. Just don't be too obvious about it. Remember my Two Swallows story? I do, and a good lesson.

Regarding the use of stats versus the in-play price, the NBA teams who currently pay for the data are not interested in the betting angle – they are interested in winning games and the data they collect, is collected for that purpose. Their advanced stats will tell them, for example, based on the current opposition line-up, what match-ups and plays would be optimal, but not what the prices should be when up by five with 2:23 left in the third quarter, one all-start player with three personal fouls and another having an off-night shooting bricks. This might well change if in-play betting takes off in the USA, but for now at least, a combination of understanding the game, and understanding the market, affords an outsider a good chance of long-term profitability.

The reason that in-play offers value is, as I have written many times in this blog, that the market - driven by fear and greed - often overreacts and someone with the ‘innate ability’ to price accurately can profit from this. The ‘innate ability’ isn’t actually innate at all, but something learned from studying the markets and watching hundreds of games. After a while, you instinctively ‘know’ when a price is wrong – that ‘blink’ moment I have written about. Look at the chart towards the end of most NBA games, and see the ups and downs.

Some of those movements are down to errors – perhaps it is greedy people forgetting that basketball is a game of momentum, or a nervous nelly not understanding the impact of two early personal fouls on LeBron James. These errors in pricing do not occur, or at least not with any real money, in the settled pre-game markets where decisions are more deliberate (as in after long and careful consideration). In-play, he who hesitates is lost, so you need to instinctively ‘know’ value when you see it.

Contrast the basketball chart with one from football, and you’ll see why I am of the opinion that trading football in-play is, for most of us, unlikely to be a profitable activity long-term. With prices changing in minimum increments, and often stalling, there are just not many non-Suspended price-changing moments in football to take advantage of, and when there is, we are unlikely to be at the front of the queue with several thousand people watching the games in many countries.

Emp writes, perhaps a little patronisingly awarding me a Gold Star for my knowledge:

I am impressed you've done your homework and know why home-field exists in almost every sport. It's also worth noting that this is most pronounced in basketball, because more than any other sport, refereeing is highly subjective. Very large numbers of fouls, many of which are close/either way calls, not to mention that superstars at home can go from the locker room to the court without being called for travelling.

I can at least die happy now, knowing that whatever else I achieve in life, there was always that one time when Emp was impressed. It means a lot to me.

His final line was actually quite amusing, and I’ll give him the benefit of the doubt that with English a second language, he didn’t actually mean to come across quite as patronisingly I have suggested. I like to think that occasionally I know what I’m talking about.

Tyttetrading had this to say on my response to RandomMan, and his perfect English makes the point very clear :

Hi, also, I think Cassini did explain it in and earlier post. The difference is also in the difference between trading/betting using knowledge about the sport or if trading/betting using knowledge of the markets. As mentioned in many posts, one of the reasons for the value that often appears on the draw in soccer matches, has nothing to do with the sport. It's because the bookmakers adjusts there odds to level their risk. Most bettors want to bet on a team, meaning that not much money will go on the draw. To attract money to the draw, and thereby level risk, the bookmaker can offer higher odds on that outcome. I hope you understand my point, my English not being perfect (sure that Cassini will point that out:-)).

Finally, does anyone else find it quite amusing or bemusing that Ladbrokes are now actively touting their Exchange, while Betfair are actively promoting their sportsbook?

Tipster Updates

You can follow NBA Tips1 at the UK's leading free tipping site but here are their results so far this season:

Not exactly evidence yet that an edge has been found, but it's early days.

Don't believe everything you read

For John Walsh's NFL selections, it is not much prettier, and I suspect that the NFL is another of those sports where insiders benefit from knowledge that is not publicly available. A terrible weekend just gone by:

A lot better in the NHL though. Here are John's results from the past week:

3PW And Goals

A little more on the suggestion from Jamie last week that three points for a win (3PW) might have made a difference to the percentage of matches resulting in draws, or in the number of goals scored.

A little more analysis shows that while there has been a slight increase in goals since 1982, there is no evidence to suggest that this can be attributed to the 3PW change.

Football has changed considerably in the past 30 years, and if we compare the 10 years prior to the points change with any period after, we are also including other changes in the game. It’s interesting to see that despite these changes, many numbers have stayed relatively constant.

The starting point for this post was the conversation about why the draw was (almost) never the favourite outcome, and the 2PW to 3PW change is a red herring.

Before the change, the previous 10 seasons saw the combined goal expectancy per game in the top English division at 2.57. Over the past 10 seasons it is 2.65.

One noticeable change though is that while the total expectancy remains much the same, the ratio of home goals versus away goals has changed. 1971 to 1981, the ratio was 1.58 to 0.99 whereas from 2004-13 it was 1.53 to 1.11.

Home goals were regularly scored at a rate greater than 3:2 (i.e. 60%) in Division One, (11 times in the last 21 seasons), but the Premier League has seen this happen just once, in 2009-10.

Not surprising then that home wins have declined from 49.5% to 45.6%, and away wins climbed from 21.8% to 27.1%. That 27% has held solid over the past 20 seasons.

Back to the draw and how it correlates with the goal totals. Looking at the past 42 seasons, in 1973-74, the goal expectancy was at a low of just 2.4, and the draw occurred in a high of 32.25% of matches. In 2011-12, goal expectancy was at a modern record high of 2.81, and the percentage of drawn matches was just 24.47%, but interestingly the correlation is highest during the eleven seasons of Division One following 3PW – for the statistically minded at 0.83.

During the entire Premier League era, there is no significant correlation although in the early days, there was. The years from 1992 to 2005 saw the correlation at 0.71, but then game the outlying season of 2005-06 when just 20.3% of matches were drawn, despite the average goals per game dropping below 2.5 for the first time since 1974. In other words, the complete opposite of what would be expected.

2005-06 was unusual though, in that while the easy wins (by 3 or more goals) were slightly below their long term average, wins by 2 or more were at an all-time (Premier League) high – 146 games.

Thus the number of ‘close’ games was less, but what was remarkable is that while the ratio of one goal matches to drawn matches averages about 1.42, in 2005-06 it was at 2.04 - a true outlier.

I mentioned in my email to XX Draw subscribers the unusual situation this season where the goals per game average has jumped up in Serie A, but dropped in the EPL, yet there are more drawn games in the latter. The reason is that there have been a number of games in Serie A which have seen a high aggregate of goals, and thus less chance of a draw (although there have been three 3-3 draws already) while in England, just one team has scored more than four goals, (Manchester City 7-0 v Norwich City at the weekend) and just that one game has seen more than five goals.

The conclusion is that while the goal expectancy clearly affects the probability of a draw, the correlation over a season may well be masked by either a high number of blow-out wins, or as in 2005-06, by an outlier.

Tipster League - Fizzers And Sparklers

After a record number of selections this week, and the Monday night games in the books, here are the latest standings. Not much time for a detailed update, so highlights and lowlights only.

The new overall leader is Forza Fizzer who showed that last week wasn't a one-off fluke and added a solid 6.14 points to overtake Emp who made a small loss. 10 points clear in the race for £250.

The big winner (+11.76) was The Football Analyst, who once again performed well in the first weekend of the month. Graeme moves up 6 places to 16th, and reduces his bounty liability by £100 to £300. Newcastle United's win over Chelsea was the big winner for Graeme.

After a week off, Fedslam was up over 8 points and moves into profit (and prize money) and up 8 places. Early days for Hofs Hackers, but results so far are promising +5.47 from the 8 selections to date. He sits in second prize spot looking at £125.

No one else moved by more than three places - the XX Draws (Under) slipped three places after dropping 0.26 points. It was a weekend of two halves for yours truly. Saturday saw 3 winning draws from 7 selections, with 5 going under, and three winners from three for the Value Selections. Sunday and Monday, not so good. No draws from 6, only 2 matches staying under, and all three value selections (all in Italy) failed to win.

At the bottom, Premier Betting made 0.5 points profit, but not enough to move out of the basement. Their big winner was the Everton v Tottenham Hotspur game which stayed under 1.5, and also gave Peter Nordsted a Drawmaster winner.

For the table as a whole, it was a good weekend, up 29.84 points, and eight entries are now in profit, including Football Elite who are finding their feet again after a challenging start to the season.

As always, let me know if you see any errors in the table. With 183 selections (some duplicates of course) to check, from at least 12 leagues, the possibility of error is very real.

The weekly John Walsh NFL and NHL updates will have to wait.  

No Place Like Home

NBA teams win approximately 60% of their home games, but in the early days of this season, that percentage has shot up to 77.5%. Early days of course, and home openers can motivate teams to heights that may not be attainable once the season is under way and reality has set in, but even so it is notable. Miami Heat has been the road team for two such openers, and lost them both, having not lost back-to-back games since January. 

Added to the fact that Miami are the current champions and the team to beat was the return of Derrick Rose in the Chicago Bulls game after a year out. In Philadelphia, the game at the 76ers saw the official retirement of Allen 'practice' Iversen which was followed by the home team scoring 19 consecutive points and going on to win the game. 

These unusual situations are likely part of the reason for the high number of home wins, as studies have shown that referees, albeit subconsciously, favour home teams, and when a crowd is pumped up, it is reasonable to assume that this influence would only be enhanced. (Between 2003 and 2011, referees called an average of 22.15 fouls on away teams per game and only 21.13 fouls on home teams). When all teams have played two or three games and their home openers out of the way, it is certain that the home win percentage will drop down to that long-term 60%, and normality will be restored. Just a few reasons why early season betting can be troublesome.

RandomMan asked:

So major European football leagues offer value (sometimes) but NBA doesn't. Why is this? Surely all statistics are as readily publicly available on both sports.

As I mentioned in the previous post, analytics in the NBA is far more advanced and arguably relevant than in football. SportVu for example uses "six cameras installed in the catwalks of every arena, and constantly tracks the movements of every player on the court and the ball. The data collected provides new stats based around speed, distance, player separation and ball possession". Having only five players and the absence of 'midfield' play makes this data a lot more relevant, and isn't publicly available, and I suspect that as a result of the higher relevance of inside information, the money betting pre-game in the NBA is for the most part a lot 'sharper' than the money sloshing around in the big football markets. There are relatively few players whose presence or absence will significantly move football markets, but this is not true in the NBA. Fortunately, the game still has to be played, and this is where value can be found, but in my opinion, tipping pre-game in the NBA is an exercise in futility.

Shiny Shoe Tips

Emp asked:

"Do you only trade in-play on the NBA or do you also have pre-match bets that you keep on?"

I do occasionally have punts on the NBA pre-game, but rarely. For me and my limited resources, the only way to consistently make profits from this sport is to take advantage of pricing errors in-play. I was discussing the topic of finding a pre-game edge with NBA Tips 1 yesterday on Twitter, and as I said then, "if a price looks to offer value, it's because someone knows something you don't".

The problem with betting NBA pre-game is that all the basic statistics are out there for everyone. What is not public knowledge is the advanced analytics that according to the Washington Post last month, three-quarters of teams have full-time staff working on, and by definition, inside information. In a sport where a team has only five players on a court at a time, a star player likely to miss a game through injury or coach's decision makes a big difference to the odds.

There was a Miami Heat v San Antonio Spurs regular season game last season, where the Spurs rested their 'big three' - Parker, Duncan and Ginobili. My phone rang, a well meaning friend anxious to give me the 'breaking news' and looking for a share of the certain profits. Except of course that by the time my friend heard the news, it wasn't news. The prices and lines had long since moved on, and anyone backing the Spurs in good faith prior to this decision was left with a poor deal. (In the event, a 'weak' Spurs team actually came very close to beating the Heat and covered the updated spread).

Joseph Kennedy is reputed to have said "that he knew it was time to get out of the market when he received stock tips from a shoe-shine boy".

Does anyone actually follow free tips from an unproven tipster? I don't know if the NBA Tips1 tipsters polish shoes during their day jobs, but I suspect their main income is not from betting or the sport of basketball. Self-proclaimed as the UK's leading free tipping site, although how much competition there is in this area is debatable, the results so far this season are a little disappointing:

I already mentioned their error in how many games are played in a season and their willingness to offer (losing) tips on meaningless pre-season games. Pre-season games are not played to win - they have one purpose which is to get the team ready for the regular season. To their credit, they did stop that, and they are clearly very enthusiastic about the NBA (as we all should be), but what are their tips based on? Based on what do they feel that their tips offer an edge, or are they for entertainment purposes?

There are few meaningful statistics available, and even if there are, there's the inside information problem as mentioned earlier. Last season's statistics are of interest, but for many teams, not relevant because players have moved on and new ones have come in. I could be very wrong, and be writing a big apology at the end of the season, but I am not seeing any evidence that at the moment, they have much idea of what it takes to win in the NBA.

Sponsorship Announcement

Fairfranco had a question about the (former) FTL table, correctly answered by Jamie and Emp. The column headings IP and LW do indeed refer to Implied Price and (table position Last Week). The +/- column shows how many places someone moved since last week. Whether the Strike Rate (SR) proves to be of any interest remains to be seen.

Graeme, The Football Analyst, has capped the bounty at £400 which is fair enough. I was actually expecting the trickle of of new entries spotting the value in taking Graeme on with a close to 20 point start to be more of a flood, but then TFA did liken himself to Barcelona which may have frightened off a few Cambridges, Alfretons and Lutons:

"Putting my systems into Cassini’s friendly tipster league would be like Barcelona playing in the Bsq Prem for a season"

Graeme did also suggest that as potentially the main sponsor of the FTL, he should have naming rights. It's just not a catchy name, with too many Ts and Fs. "TFA's TFL" or "The FTL sponsored by / in association with TFA"? WTF! But with up to £400 coming our way, I guess I can accommodate the proposal, so the FTL will henceforth be known as The Football Analyst Tipster League. With such large sums of money at stake, the "Friendly" went out of the window long ago anyway.

For those concerned that the impartial approach of this blog might be threatened, rest assured that editorial content will be unaffected, and I will continue to call it as I see it. If the TFA selections perform more like Barnet than Barcelona, you will read it here. Speaking of which, Barcelona did slip up again last night with Oldham Athletic losing at Notts County.

And yes Graeme, there is a minimum number of selections required - 64 for anyone joining before the end of September, and 56 for those in October. Plus no sitting on profits, with a minimum of five selections per month. With 22 entries to track already, I think I'll take Graeme's suggestion one step further and close it to new entries for this season. At this time of the year, as many of you know, much of my focus switches to the NBA which opened its 2013-14 season last night. The best sport for in-play trading by some distance.

Although my interest is mostly in the Western Conference, where 'my' Golden State Warriors (if you see a team live more than once, they are yours!) look very good to improve on last season's play-off run and win more than 47 regular season games, I will also be keeping an eye on the Brooklyn Nets in the East. Their Russian owner Mikhail Prokhorov, is bankrolling more than $102 million in player salaries, (and projected luxury tax payments totalling $86 million) a sum higher than any other team in the league. They are the Chelsea of the NBA, albeit with an untried coach (part-owner Jason Kidd) but assisted by a former Nets head-coach in Lawrence Frank. I'm not sure the Nets are at home to Mr. Failure this season, and with Derrick Rose back and looking strong for the Chicago Bulls, and a strong Indiana Pacers team, Miami Heat face some tough competition if they are to make it to their fourth consecutive Finals.

Fizzer Fizzles

Overall, a losing weekend for the FTL tipsters combined, but some success stories within the total loss. The big winner of the weekend was Forza Fizzer who found five winners from seven selections, made a profit of 11.13 points, and leapt from 13th into second place overall, and first place (£225) in the race for the cash. Fizz had these decent priced home winners:

BURNLEY 2.58, YEOVIL TOWN 4.56, COLCHESTER UNITED 4.85, OLDHAM ATHLETIC 2.89, STEVENAGE 3.25 

Emp held onto his overall lead adding another 4.66 points, and coincidentally also had Burnley and Yeovil Town in his selections. He also had Reims to win at Olympique Marseille at 8.18, and win they did, 3-2.

Fairfranco moved up to third overall (£112.50) gaining 5.58 points, and the idle Skeeve hangs on to third cash place and £67.50. Provisionally in fourth cash spot (£45) pending payment, is new boy Graham who opened his account with a 0.56 point profit. The only other entry in profit is my Cassini Value Selections, but they had a disastrous Saturday with all three losing without scoring and Toulouse the worst losing 0-5. Sunday was better with Bologna beating Livorno, although it could hardly have been worse!

Of the entries in the red, i.e. most of them, Punter's Friend made a move off the bottom moving up four places with an 8.5 point profit, and Murphy's Law also moved up four places with a smaller 0.72 point profit. Football Elite made a small profit, as did the Drawmaster, but the XX Draws (and Unders) as well as the Bundeslayga system all had small losses. Late goals again proved an irritation with three of the nine XX Draw games losing out to them, although Paris St Germain did score a late goal to make up for that a little.

Fedslam sat out this week, debutante Jamie Ayton had a weekend to forget, and Rubicon dropped four places. Webbo lost nearly five points, Scatter Gun dropped just one and then we have the big boys Peter Nordsted and his Premier Betting and The Football Analyst. Graeme actually made 0.7 points on the weekend but still drops one place with a bounty liability of £400 currently.

Pete's Premier Betting had another losing weekend. The Official bets, using the prices quoted by Peter, lost another 5.5 points, while the cash eligible entry lost 5.18 points. Sunday was actually a good day for them after a wipeout on Saturday, with Tottenham Hotspur v Hull City the low scoring game expected, and Sunderland defeating Newcastle United. The tide may have turned.

Rather embarrassingly, the Amateurs are now performing better than the Professionals, although as that simply means they are losing less, it's not too much to shout about. Football remains one of the harder nuts in betting to crack, but someone losing his nuts right now is my old friend Steve over at Daily 25 who had this to say last week:

How do I even describe how I’m feeling right now. Depressed, angry, pissed. I decided to get up at 5:30am this morning as I had $32,000 worth of bets on the League 1 and 2 games and knew it was a pivotal moment for my betting career. I’m pretty sure I have never lost that much money in just 90 minutes and I’ve had some big nights in Vegas.

The $ in question are Australian, but that's a lot of money to be staking on lower league football. I know when Steve followed the XX Draws that he didn't get the best odds, which is crucial in my opinion, and these were selections in big markets, and I can't help thinking that in less liquid matches, the value must be borderline at best given the stake size.

John Walsh's hockey picks have gone as cold as the ice they are played on. Since the last update, he had four consecutive losing selections but the rot came to an end last night as the Dallas Stars hung on to beat the Buffalo Sabres 4-3.

After three losing weeks, John's NFL selections returned to profit this week, although still down on the season overall.

Back to football and goal expectancy, a couple of comments on yesterday's post. Jamie had this to say:

Well that's not really a provable statement, but if we let ϵ>0...

Lets go with the alternative statement of when is a football match likely to end in a draw, which then leads to either low goal expectancy (for 0-0 scorelines) or it suits both teams as you suggest.

Thus, a draw doesn't usually suit both teams.

Is this because a win is worth 3 points to the winner and a draw is worth 1 point each? - i.e. game theory.

Or because this is a sport we're talking about?

Maybe a model based on how much a draw would suit both teams would give a higher draw probability in some cases.

An in-play model might work with teams who play a specific style - i.e. more likely to sit back and settle for a draw with x mins to go.

Nick added:

As Jamie suggests you would think the points incentive of a win over a draw would effect the draw price. But the year they changed from 2pts to 3pts in England (1981/82), there was 121 draws, actually 3 draws more than the previous season's 118.

In a way it proves Cassini's "mathematical proof" on goal expectancies. If back then there were people running their BBC Micros or Acorn Electrons to get goal expectancies to derive match odds, their models wouldn't have been any less accurate, at least for draws.

It's an interesting angle, and while in the immediate period after the introduction of 3 points for a win, the number of draws dropped, they have since returned to the same level. As Nick said, there was an immediate drop in draws as might be expected with more goals:

In the five seasons before the switch, there were an average of 133.0 draws per season in what was then the First Division; in the five seasons after, there was an average of 113.4. By way of comparison, there were 108 in the Premier League last season and if you extrapolate that to take account of the reduction in the number of top-flight teams from 22 to 20, you get a figure of 131.

I think that in a match where the draw would suit both teams, this is reflected in the goal expectancies, weighted by just how much a draw would suit both teams. An end of season match where a draw would see both teams survive relegation for example would be a draw that would "really" suit both teams, whereas an early season Norwich City v Cardiff City might suit both teams, but maybe not enough to stop them trying to the end for a winner that would give them an extra two points.
  
Whether three points for a win possibly reduces the likelihood that teams will come to a tacit agreement to play out the final minutes of a tied game for a draw is an interesting thought, and it certainly seems that way at times!

If three points for a win means more goals, then the number of draws would be expected to fall. There's a reason why the Bundesliga sees less draws than Ligue 1.

Incidentally, the origin of the two points for a win, one for a draw comes from 'challenge' matches, where a winner would take all, and in the event of a draw, the pot would be split.

Why The Draw Is (Almost) Never Favourite

While we await the conclusion of the weekend's matches and the official prices before the latest, eagerly awaited FTL table is published, here is a post from the Betfair Forum and my contribution to it. The original poster (the astronomy minded Swift-Tuttle) posed this question:

Does anyone have a mathematical proof (or could refer me to one) of why a football match is unlikely to end in a draw?

To put it another way, of the 3 results possible the draw is never favourite. Why not?

Please don't mention those matches in Italy where the Mafia are involved and the draw is a firm favourite. I'm discounting those.

My thoughts on this topic were:

The original poster asked if there was any mathematical proof why the draw can never be the favourite outcome in a match, excluding special cases where a draw will suit both sides such as are seen in late season Italian league matches or in the final game of round-robin group matches.

While it is not a mathematical proof in the true sense of the word, and unlikely to win me a Fields Medal (I’m too old to win it anyway), the answer lies in the fact that any credible football pricing model derives the match odds from goal expectancies, and the only combination of Home Team goal expectancy v Away Team goal expectancy which will result in the draw having the highest probability is when goal expectancy is very low.

If your model uses a simplistic method along the lines of “in the last 20 matches between a top team and mid-table team, 10 matches have been drawn” then you might put the draw probability at 50%, but such a model is fundamentally flawed and no serious punter would ever use such a model.

The probability of a draw is clearly highest when the expectancy of goals is low. Of the under 1.5 goal outcomes, one-third (the 0-0) results in a draw. Add another goal and one-third of the six possible outcomes results in a draw, but add one more goal, and the draw probability drops precipitously. The draw outcomes from the Under 3.5 goal markets total two from ten.

So unless the goal expectancy in a match is two or less, the draw will never be favourite, but the closer to zero the goal expectancy is, the more probable the draw is. The relative lack of interest in the draw as a betting outcome does, in my opinion, make the draw a value bet when the market’s expectation of goals is higher than it should be.

Results from my model show that in 1,823 such matches over four seasons in the top five European Leagues, 565 matches have resulted in draws, an implied draw price of 3.23 which, with the possible exception of the low-scoring Ligue 1, is usually an easily beatable price.