Dragon's Lair

Green All Over is all over for another calendar year, as I set sail for some family time where finding time to blog and invest will not be a priority. With parents well into their 80s, the truth is that there will not be many more Christmases for me when they will be around, but there will still be time for a trip to a foreign land on Boxing Day though, (well, Cardiff), weather permitting, with the next Cassini generation.

So a Merry Christmas and a Happy New Year to all of you, and I look forward to provoking, teasing, irritating, entertaining, boring, educating, infuriating and (perhaps) amusing you in 2013.

Sports Hedge Fund

From Veryan Allen's Hedge Fund Blog comes this piece on sports investing. Some incorrect generalisations, but an interesting read:

Sports hedge fund? Unlike stocks, bonds, real estate and commodities, sports results do not depend on economic factors. Sports bets offer arbitrages and mis-pricings like "efficient" markets. Find alpha in stocks, bonds, commodities, currencies or sports? They offer different sources of performance so ALL OF THEM. Diversification of uncorrelated returns is the way to go. Focus on building return streams.

Michael Jordan, Wayne Gretzky, Babe Ruth, Pele and the greatest sportsperson ever, Donald Bradman, were just lucky flukes like hedge fund managers Jim Simons, George Soros, Warren Buffett, Jesse Livermore and Munehisa Homma? Buy every stock because it's "impossible" to find good stocks in advance. There are no good sports people just lucky ones? No such thing as talent and dedication?

Any pension fund not investing in sports is breaching fiduciary responsibilities. How many teams would win if managers just randomly chose players from the phone book? Skill is necessary to win in any business. Passive fans say there is no such thing as investment skill! There's no sporting ability as well? Selectors, coaches and scouts waste time because sports talent is just luck? Investing is like sports; having competitive edges others don't have.

Name a sports team that won by picking anyone off the street. Index funds do that and GUARANTEE most of your money gets tied up in losers. John Bogle thinks if I golf with Tiger Woods or play tennis against Roger Federer I have a 50% chance of winning. That is the absurd random walk theory behind "passive". Eugene Fama theorizes a 5ft tall person is as likely to slam dunk a basketball as a 7 footer. Even if you weigh a 100lbs a great career may await you as a sumo wrestler or defensive linebacker!

Investors are supposed to be satisfied with the "average" return of index funds but you don't see "average" sportspeople on TV. Could you watch a sport where every amateur player competes? Media ratings would not be good just like the poor returns of holding "every" stock.

Unlike the sports media, the financial media focuses on market "averages" despite their irrelevance. Imagine Wimbledon including every owner of a tennis racket. Most matches would end 6-0, 6-0, 6-0 but would it change the tournament winner? Efficient market hypothesizers say if I run 100 meters enough times I'll randomly walk it in 9 seconds at some point! The first sub 2 hour marathon could be by you. Good luck betting on stochastic stupidity.

$7 billion will be bet on the Super Bowl, almost all without thorough systematic analysis. Larger amounts will be bet on the soccer World Cup. The favorite is usually overpriced in most sports.

Another anomaly, among many, is geographic bias. Many gamblers favor the team whose name they geographically most closely identify with, even when the owners, players, coaches and managers have no such locational origination. Those who invest or gamble based on emotion or patriotism are likely to lose money over time. Amazing how many throw money at teams from "their" city or country regardless of the odds. It is nice to arbitrage them though.

Curious the contrast of how unhedged betting on equity or credit markets is commonly regarded as "investing" but sports bets are "gambling". With skilled analysis, future winning teams and players can be identified, as can winning stocks. Short selling the losers is even better. Sports offer no beta, just like stocks(!), but there is plenty of absolute return available if you know what you are doing. You won't always be correct of course but all that is required is a small forecasting advantage. The odds reflect the crowd's perception of winning probability NOT the actual probability. Variant perception - the crowd must lose over time to those with more information and sophisticated analytics.

Sports produce a vast array of statistics, which with the right tools can be datamined for PREDICTIVE information. Take horse racing. I know quants who have taken serious money (USD 100 million+) out of only Tokyo or Hong Kong horse racing. Most bets are made based on the lucky number of the jockey or the feng shui of the stables or another irrelevant metric.

Such illogicality allows the rare skilled, disciplined bettor to arbitrage the many unskilled and irrational.

In every sport, teams build up a database of results. Drilling down, each individual player or horse builds a career track record. Just like a stock, if you evaluate the data closely enough you will be able to make better bets than "random" would imply and arbitrage the prices of those who set the odds and spreads.

Nowadays with sports betting in reasonable size easy to implement and with significant global capacity, I would expect sports hedge funds to emerge. Before I entered finance I managed a private sports betting hedge fund. It will be fascinating to observe investor reaction to what, to me, has always been clear; making stock picks and sports bets is the SAME underlying investment process. Putting money on the Dallas Mavericks or Real Madrid is structurally isomorphic to betting on stock or bond prices.

Whether stocks or sports, it is a skilled quantitative and fundamental evaluation that enables accurate bets to be made after elimination of institutional rigidity and local or national biases.

Develop an informational or analytical edge, make bets in many areas and arbitrage the emotional sports betting crowd. Maybe someone will set up a sports betting hedge fund. There is plenty of alpha out there, globally.

You Little Ripper

I love the smell of hampers in the morning, and I love Australians! In my inbox today is evidence that whatever we might say about them, some are just downright decent blokes... ;) Steve writes:

G'day mate,

This should shut you up. :)

http://www.daily25.com/as-promised/

Thanks for a great year of draws. Never have I watched matches and been so bored and yet happy at the same time when a game ends 0-0.

Steve is one of the majority who typically like to see goals in their football. I used to be like that, but the XX Draws have changed me, and the 'perfect' game is now a 0-0. Anyway, for those too lazy to clink the link to the post, or for stickybeaks who want to take a squizz, here it is in full:

Not one to say something and not follow through, here is a gift page created for the master of the draws, Cassini.

http://kickstartkristmas.com/Gift/688

A few weeks ago I mentioned jokingly that I would buy Cassini a hamper for the excellent returns from the XX Draws, he took it upon himself to make it a reality and last week I said, and I quote,

XX Draws added another $2,200 profit and if he can make even $1 profit next week then a small (very small) hamper will be on its way.

Now I am $50 poorer and Cassini will soon be a few pounds heavier. I will leave the page up for others to contribute to and leave a message to say thanks for a great year, whatever the final figure is, I will let Cassini choose the Hamper that he wishes and send it on Christmas day, so he should receive it in time for a toast to myself on New Years Eve.

A few weeks ago I also started a gift page for Jonno from Sportpunter.com. As I am an actual friend of his, I managed to trick him into revealing what he wanted for Christmas and created a page for him too. An amazing amount of $250 has so far been donated and we are just $50 short of getting the gift. If you want to say thanks to Jonno for his many years of hard work and what I can only assume is an overflowing bank account if you have followed his tips, then go to the address below and donate what you can.
http://kickstartkristmas.com/Gift/575

Sadly this year there will be no gifts for Skeeve and FB Elite, but let's hope that changes in the new year.

Note: Kickstartkristmas.com is a site I created a few weeks ago to help me get what I really wanted for Christmas. If you want to donate to my gift you can see it here
http://kickstartkristmas.com/Gift/680
Merry Christmas

What a gentleman, but while I would hate to come across as selfish, I encourage all of you to ignore the second and third links and focus on the first.

Seriously, I appreciate Steve's spirit of giving and if everyone reading this post gives a dollar (Aussie) to one on the three, then there will be much happiness across the globe on Christmas morning. Perhaps any excess contributions could find their way to a charity somewhere. Just a thought. 

I Smell Hamper

It's so close, I can almost smell it. After a weekend with four Classic XX Draw selections, (three winners and an 87th minute 1-0) and five Extended selections (another two winners), surely it won't be long before my boughs of holly decked halls are filled with the aroma of such gourmet specialities as caviar, 5J ham, Champagne, confectionery, crackers, cheeses and wines selected for their quality, provenance and of course, flavour. Always eager to assist, I have located the Harrods hamper page here, although of course Harrods is only a suggestion, and a John Lewis selections would be just fine.

Being serious again for a minute, here is the latest FTL table for your perusal. There is still one result to come from Monday's matches, but the positions will not change:

Noteworthy gainers this week were Neil, up to 6th, XX Draws (Classic) into profit and up from 18th, Back Lay The Draw, up to 14th, while Premier Edge were probably the biggest losers dropping from 5th to 9th. Twelve entries are now in profit, but the overall net returns are still well in the red. Entries in italics represent entrants who have given up for whatever reason, and the (Pro) qualifier identifies professional services.

There will be no updates to this table for a couple of weeks as I take a well earned break for the holidays. The XX Draws will continue to flow, although the Bundesliga has now stopped for its winter break, with France, Italy and Spain breaking after next weekend, so the flow of winners might actually be a mere trickle.

It's All Over. Well, 80% Anyway

Although most of you will not be surprised to read that the majority of bets on the Under / Over markets are placed on the Overs, you might be surprised about how lopsided the bets on Overs are. A recent (November 2012) study from the University of Zurich finds that:

...betting volume is highly concentrated on the over bet, accounting for over 80% of the total betting volume.

The study used data primarily from central European based tipico, and studied whether bookmakers, knowing that the more 'desirable' bet was on overs, (most punters apparently find it a more attractive proposition to bet on goals) would adjust the price (offer worse odds) in an attempt to maximise their profits, a strategy that had been suggested by Steven Levitt in his famous 2004 study Why Are Gambling Markets Organised So Differently From Financial Markets? 

The conclusion of the new study is that bookmakers do not distort the odds offered, since:

Bettors can easily compare the odds listed by several diff erent bookmakers and fi nd the best odds through a number of websites such as oddsportal.com or betbrain.com, which increases the bettors' price sensitivity. Thus, small price changes tend to have a large impact on the betting volume and eventually on the bookmaker's profi t. If a bookmaker increases the price (lowers the odds) of an over bet, sentimental bettors would switch to a competitor. On the other hand, if a bookmaker lowers the price (increases the odds) of an over bet, he gains additional sentimental betting volume, however, at the risk of substantial losses. We find that bookmakers do not distort their odds in either direction. They o ffer odds that reflect their best prediction of the true outcome probability and add an equally distributed commission, even when bettor sentiment leads to a highly asymmetric volume distribution.

Thanks to Scott (PT) for the link, and Helmut Reul @SportAnalytik for Tweeting it also.

Still plenty of games to come this weekend, but a few FTL contenders are done for the weekend, so I'll update them here. First, Peter Nordsted continues to play down his recent form with:

Well we have fallen a little flat over the past two weeks with none of the selections ending level. This should be expected though as we have had a spectacular run of results recently. Let's hope this weekend we can return to some form.

By my reckoning, Pete has had one winner from three selections for the past two weekends, and continued the trend with Stoke City v Everton this week, so the statement that 'none of the selections' have ended level is not quite correct. Drawmaster continues to perform well, and one winner from three each weekend should see Peter being a little more positive in my opinion.

The Stoke v Everton match was also an XX Classic Draw selection, one of two on Saturday with the other being the 1-1 draw between Reims and Liile.

With the Extended selections finding one from four (Granada v Sevilla 0-0) the hamper is on! Three more matches to negotiate today, but I am salivating already. Or perhaps at my age, drooling is the correct term.

Football Elite found another winner to go with the value Norwich selection already discussed, which was Fortuna Dusseldorf who beat Hannover '96 2-1. One money back draw (Reims v Lille 1-1) and one loss on Granada v Real Sociedad 0-0 rounded out the weekend for FE for an overall small profit. Interestingly, three of the four FE selections were XX Draw selections, and two of them finished level.

Back the Lay The Draw selection found one winner (Nijmegen v PSV Eindhoven 1-1) and after being down as much as 19.7 points, this strategy is now down just 5.65 points. No winners for Talkbet (Jon) or Little Al, but Neil had two winners and two losers, and increased his profits, while Premier Betting also had two winners from two selections, with another to come today.

Premier Edge still has five more to come, but at 3-6 so far, need a strong Sunday to keep the run going. The Bundeslayga system enjoyed Schalke's defeat to Freiburg, while Football Formbook also needs some wins on Sunday to finish the weekend in profit. No selections again from Griff, Tony or CKL.

One comment I've not addressed yet, which was this one from the exotically named Giovannina Bäcker who suggested:

Hey Cassini, Why bother with premium Charge and not just making new Account on your Son / Dad / Wife / Brother / Sister and so on... of course you need to make it a bit smarter (not same address, not directly close your betting, start betting slowly in meanwhile on other account...) - i think if you want, you can avoid at least some major junk of it..

If only it were that simple, but my betting style has become very specialised over the years. Nor do I have a brother. For straight punts, I can of course take my business elsewhere, but Betfair is the only viable option for the specialised markets where I make my bigger profits, and it would take Betair and its algorithm all of five minutes to match my new account fingerprint to my old one. Someone has said, the worst that can happen is that you pay the Premium Charges that you would have paid anyway, but I suspect my accounts (old and new) would be closed, and that is a price I am not willing to pay. With the consensus being that liquidity is dropping across all sports on Betfair, I remain hopeful that the Premium Charge will be modified at some point, or if BETDAQ would get their act together and turn a few markets in-play, that would help!

I read on the Betfair forum that the football markets now total around 80 on a typical EPL game (Norwich City v Wigan Athletic for example) which seems excessive to me, and only likely to spread the money too thin. Several markets are just re-wordings of other markets - Norwich to Score? Wigan to keep a Clean Sheet? and while I know some people like these fringe markets, (First Half Corners Match Bet? - Really?) - for me many of them are a distraction and a waste of time.

And finally, according to Blogger, this blog is closing in on half a million hits, just three months shy of its fifth anniversary. It will be taking a break at Christmas, as will its owner, but 2013 should arrive with over 500k views.

Spam And Prices

I may have to reconsider this blog's position on comments. As many of you know, posting comments has always been wide open, but of late I seem to be getting spam along the lines of  

Facebook Hot Girls, Most Popular Singer Girls, Top Most Famous Girls in the World.

Now is a good business, the girls are very good. I would like to invite you to visit our website too

Make Earning on your Every panels you purchase, Many types of Earning Systems

Get your own home based job in data entry, copy pasting, clicking and different more jobs

It is interesting to read that you like to bet on the "Draw" in football matches.I like betting on football but i always back a football team to win.I like getting the odds in my favour though by using a bookmakers free bet bonus to bet with though!

Fascinating stuff (not), and all the above in the last week or so. It's becoming a daily annoyance, so if I have to tighten the restriction on comments, I hope you will understand that it has nothing to do with attempting to limit the exchange of thoughts, opinions and ideas, but a necessary step. One might hope that Google's anti-spam filters would be a little more robust, because the other side of the coin is that I find some genuine comments in my spam folder.

WhyAlwaysMe enjoyed my Sinking Below post, writing "Superb post Cassini and really informative". I can't help but agree.

He then follows up with the following (posted three times, no doubt because it vanished into the aforementioned spam folder):

I too am interested in your XX Draws project but am struggling to get a handle on your data.

Do you have an average back price for the draw over the 1175 selections please? I looked in the prospectus and found a figure of 3.45 over 361 selections.

Also taking the 3.45 for a minute, is that a bookie's price or Betfair? If it is the latter is it adjusted for the 5% commission? I don't see Betfair as a viable betting proposition unless you plan to trade, if the 3.45 is not adjusted for comm the margins become wafer thin.

I am impressed by the detail and seriousness with which you approach xx Draws, however my concern is are you really beating the market? Your 3.11 break even price suggests that you are which is good going for sure. However if the 3.45 is pre commission it does look a bit tight. It would be good to clarify too if these prices are near kick off one's or early prices.

Thanks for the great blog Cassini.

As I have noted before, I started recording the draw prices on January 1, 2011, and as explained in the prospectus, projections for results prior to this date are purely that - projections. There are no guarantees, but the draw price is historically very robust, and the more data I collect moving forward, the more accurate the numbers become that I can back-fit.

The prices are taken from Betfair, primarily because it's convenient, and the draw price is usually as good as the best bookie price. Not always though, and the other issue is that the draw price often moves between my recording it and kick-off. Today's QPR v Fulham game for example saw me record a price of 3.6, while Peter Nordsted found 3.68 available at Pinnacle. One 'problem' with using Betfair prices though, is that they fluctuate. Again looking at today's matches, and the Norwich City v Wigan Athletic game was a Football Elite selection with a best price of 2.25 with Bet Victor as of Friday. Two days earlier, and Neil had picked the same selection at 2.29. Early on Saturday, and Norwich could be backed at 2.4, as I tipped in my previous post. Now that the game is in the books, I can probably reveal that FE stated in his comments:

I didn’t expect to be having this one as I assumed Norwich would be around the 1.9 – 2.0 mark I have them at but 2.2+ looks worth taking IMO both in terms of current form and long term trends.

As I mentioned yesterday, I had Norwich priced at 1.97, so the 2.4 was very surprising to me also, with its 21.8% edge.

No one includes commission for comparative purposes for the simple reason that not everyone is on the same commission rate, but yes, if you are on 5% commission, you need 3.22 to get that net 3.11. The individual needs to factor in whatever your personal rate is, and you also need to consider whether you can get your bets in early and wait to get filled, or you only bet just before kick-off and have to take what is there. In my prospectus, I emphasise the importance of getting the best price possible, because at the prices draws trade at, there's a big difference between backing at 3.45 or one tick higher at 3.5.

When it comes to football betting, margins are going to be thin, and ROIs are going to be low. There's no get-rich-quick formula, but if you have a selection process that has just a small ROI and plenty of selections, your bank will grow slowly but surely.

Another older comment that I haven't responded to was this one from Danny Murphy in November:

Great blog Cassini. One thing I have wondered with your XX selections is this - are they all equal as regards testing? Has XX Extended U1.5 been tested over a few hundred trials for example?

In the prospectus I could see the historical XX draw stats but I am not so clear on the records of the other categories.

As you commented above it's very hard to make money betting on football and perhaps your table is a good illustration of that.

The U1.5, U3.5 and Half-Time 0-0 (HT00) markets are all new this season, and have not been tested over hundreds of trials. My theory was that if the U2.5 and Match Odds draw (MOX) markets are profitable, then these should also be profitable, but I am not finding this to be necessarily true, although the sample size is still small. One concern is that these are far more illiquid markets than the popular U2.5 and MOX markets, but even so, it's quite bizarre that the Extended selections occupy four of the top seven spots in the table, yet the fifth market - U1.5 - languishes in 29th place. The FTL table does show this and other discrepancies, but over time with more selections, I still expect the Classic, Extended and Bundesliga bets to cluster closer together, with the thinner markets trailing the more liquid ones.

The Extended HT00 results illustrate the importance of not jumping to conclusions too early. After 58 selections, this category showed a profit of 27.76 points (47.86%), yet the next 29 selections yielded just one winner. At the time of writing, 137 selections, the ROI is down to the 5% level, which is far more sustainable. Don't try this bet in the higher scoring Bundesliga though - 32 selections, 6 winners and an ROI of -38.59%!

All Roads Lead To 1.4

Much has been written this season about how Liverpool are always priced too short, and today's game v Aston Villa sees Liverpool recommended as a lay at 1.4 by Neil. Al opts for the draw, while Premier Edge go for 2 or 3 goals in this match, (2.13 BETDAQ) but don't involve themselves in the Match Odds market.

For me, there is little value in the Match Odds prices for this match, but I do have a different opinion on the Unders markets. While our Match Odds numbers are in synch, we arrive there by different routes. Betfair has the Over 2.5 favourite at 1.83, while I have Under 2.5 priced at 1.8 based on recent form, although for matches with this 'profile' as they say, my Over 2.5 price is 1.83. Incidentally, my price on the 2 or 3 goals market is 2.23.

With Neil and Premier Edge both in profit in the FTL, I thought I might share their selections this week. First up Neil who goes for:

Liverpool v Villa - Lay Liverpool @ 1.4
Norwich v Wigan - Norwich @ 2.29 [I have Norwich at 1.97 - Value!, and Norwich are out to 2.4 right now]
QPR v Fulham - Draw @ 3.56
Stoke v Everton - Stoke @ 3.24 [3.15 - No Value]

and from Premier Edge on a run of 17 winners from the last 23 selections comes no less than 14 more picks this week:

Liverpool v Aston Villa – 2 or 3 goals (2.13 at Betdaq) [2.23]
Manchester United v Sunderland – Under 3.25 goals Asian Goal line (1.943 at Pinnacle)
Norwich v Wigan – 2 or 3 goals (2.12 at Betdaq) [2.08]
QPR v Fulham – BTTS: No (2.35 at 32Red)
Stoke v Everton – 2 or 3 goals (2.12 at Betdaq) [2.08]
Mallorca v Athletic Bilbao – 2 or 3 goals (2.0 at Paddy Power) [2.27]
Granada v Real Sociedad – Over 2.25 goals Asian Goal line (1.943 at Pinnacle)
Sevilla v Malaga – Sevilla -0.25 Asian Handicap (1.877 at Pinnacle)
Sevilla v Malaga – 2 or 3 goals (2.0 at Paddy Power) [2.1]
Zaragoza v Levante – Both Teams to Score: No (2.0 at Ladbrokes) [1.83]
Tottenham v Swansea – Under 3 goals Asian Goal line (1.787 at Pinnacle)
Valencia v Rayo Vallecano – Both Teams to Score: Yes (1.8 at BetVictor) [1.75]
Real Madrid v Espanyol – Both Teams to Score: No (1.8 at Paddy Power/Stan James) [1.44]
Barcelona v Atletico Madrid – Over 3.25 goals Asian Goal line (1.952 at Pinnacle)

There are also three Bundeslayga (also in profit) selections this week:

Results Highlights

Al is still bitter about the US election: 

It's good no emotions are used in your draw selections. Unlike your love of Obama which is based purely on emotion and no logic.

The bet was made because the price on Obama was value if you ignored the mainstream press and looked behind the numbers. The signal and the noise in fact, as Nate Silver calls it in his new book. It was a rare opportunity and all the sweeter that it was Obama and not Romney in the winner's circle. Keep drinking the Kool-Aid Al.

A Little Misleading?

A couple of interesting details in Betfair's latest results.

Part of this increase has resulted from many customers being introduced to the concept of trading through our 'Cash Out' tool, which makes it simpler for customers to lock-in profits or limit losses during a sporting event. The tool is one of the ways that we are bringing the advantages of the exchange to a wider audience and it has been particularly popular with recently acquired customers.

For recently acquired, read 'those who don't yet know any better'. And on 'risk' revenue they reported:

Risk revenue increased by 47% in the period, reflecting the higher volumes and an easy comparative in the first quarter of last year, when margins were weak following losses in our telephone betting channel. Margins in August and September were affected by adverse sporting results, but October and November have seen strong margins.

It would be interesting to know what those adverse sporting results were. I've had a few adverse sporting results myself over the years!

Betfair employees might be a little nervous, reading under 'cost cutting' that:

These efficiencies will include how we source services from external suppliers as well as the number of people we employ. This will involve some difficult decisions, but we are confident that our employees see the need for the business to become leaner and will embrace being part of a more efficient organisation.

Other nuggets:

Our trading team further manages risk through exposure management, including hedging into the Exchange, and customer management, including limiting the stakes of certain customers.

Betfair also offers products such as casino and arcade games in which customers are playing against the ‘house’. Revenue is generated through an inbuilt margin, with the odds favouring the operator. Over a short period of time returns can be volatile, but in the long run the operator should win.

Reduced activity by a significant number of customers could have a material adverse effect on our operations, financial performance and prospects.

We closely monitor the behaviour of customers and have a team focused on their management and retention. We aim to ensure we provide a service and platform which grow value for both the customer and Betfair, and significantly reduce the risk of customers leaving.

It could of course be a coincidence that just a few weeks after entering the 50% tax zone, and with the subsequent significant decrease in activity on my account, I received an email from my new 'account manager' - I never heard from the 'old' one - asking for my opinion on a number of topics. It was actually rather annoying that they only bother when it's too late, but while I won't be closing my account, I will certainly be spending less time on Betfair, not that my £30k of Premium Charges over the years will make much of a dent in their numbers, but if enough people reduce their activity, then as noted above, they might notice.  

Ups And Downs In The FTL

A little tardy this week, as running the numbers for the midweek selections took priority, although we'd have all been better off if I hadn't bothered. Three losses in the French Ligue, and all three saw 3 or 4 goals in them. The only other entrants with midweek selections were Football Formbook with a loser, and Football Elite with a loser. The table below includes the midweek games. Premier Edge was the big winner this weekend with a third consecutive winning weekend moving them up to fifth, and Football Formbook was the big loser dropping down to last (30th) place. Backing the Lay The draw selection is finally hitting some winners, (4 from the last 6 and up to 19th), something Football Elite would like to see fairly soon after the worst run that I can remember from that service which now languishes in 25th spot. Here's the table:

I need a profit from the XX Draws this week. Anything will do apparently, because Daily25 writes in his review of last week:

XX Draws added another $2,200 profit and if he can make even $1 profit next week then a small (very small) hamper will be on its way.

We shall see. As you all know, the numbers decide the selections, so there's no emotions to them, although with three midweek losers, it's looking like I may be going hungry this Xxmas. I'll try praying.

tyttetrading had a couple of comments, one was:

Hi Cassini, regarding the FTL standings. When betting, its very normal to say that you need at least 500 selections to prove your edge. I am not a member, but do you have this number of selections over the years?

I still have some weekend and midweek results to enter, but as of right now there have been 1,175 selections (Classic and Extended) with 378 winners, 32%, which means a 3.11 price is required to break even.

His second one was:

Hi Cassini, I was reading through your service page, since I wanted to look through the stats for your service. However, the "results" page is blank, and I can only find the results for the 2011-12 season (which is only a sample of around 150 picks). Do you have a more samples or a link to more extended results?

The 2011-12 results page actually covers 225 selections (from 1.1.11), but I have updated the results page with the 58 Classic and 136 Extended selections for this season to date. For this season to date, the Extended results are good, the Classic not so good, but overall we are in profit with 61 winners from 94 games, 31.4% and 3.18 needed to break even. It was a good decision to expand the service to include the Extended selections.

Sinking Below

I mentioned in yesterday's post that I would be ignoring BigAl's provocations, but at least one person has requested me to sink below him, at least for now. Pat Cusack commented:

I was looking forward to your response to Big Al. He raised, what appear at first glance at least, to be valid points.

Your spats have been very interesting in the past, so go on don't rise above him just yet.

Keep up the great blog

I'd hate to disappoint anyone, so for those who may have not have seen it, BigAl's comment was this:

I hope you don't mind me asking about your bankroll growth.

Your bankroll growth in 2005 was around 1400%, in 2006 approx 2500%.

Simple extrapolation since then suggests average annual growth of little more than 100%. [A subsequent comment noted that "Just realised, the extrapolated average bankroll growth figure over the last few years that I mentioned is misleading. The average will be much much lower than the 100% I suggested but I suspect a man of Cassini's intellect will understand what I meant".]

Obvious factors are the increased charges from Betfair, and possible relative lack of liquidity in some markets compared to desired bet size.

Many professionals will regularly dip into their funds to provide an income stream but your comments in your post clearly guard against this in your own circumstances.

It's a massive slowdown in % terms and as far as I can tell you have increasingly more bets on football as per the FTL, and these are in highly liquid markets.

Furthermore, rightly or wrongly, you have often advocated staking approx 2% of bankroll per bet so your stakes should be increasing.

It's probably fair to say that bankroll growth from a small bank several years ago is always likely to be far higher than in the present day with a much bigger bank but nonetheless your figures seem slightly curious unless you largely deviate from many of the strategies you preach.

Perhaps you have had a lot of bills and weddings to pay for along the way.

I think most readers will agree that when you are dealing with small numbers, percentage figures are meaningless and often very misleading. I actually don't know what my bankroll growth was in 2005, so I'm surprised that BigAl does. All I know is that in 20 months from April 2004 it grew from £98.50 to four figures, shrunk back down to £20 and ended 2005 at £1,329. Regardless, the rate of growth was strong, but from such a low starting point, not really that meaningful.

The first year I kept full records was 2006, and the year started with a bankroll again in four figures, but second time around, a few important lessons had been learned. Growing this to £32k+ in 12 months was impressive, although £18k of that came in the final three months of the year, and in hindsight was perhaps a little fortuitous. The start of 2007 was calamitous, with more than £5k being lost on day one, a lesson for thinking that I could do no wrong, and a decision from which my bankroll didn't recover until August of that year. I tell this story to make it clear that bankroll growth is far from steady. 

Another thing I should point out is that Betfair's default maximum exposure limit is £5,000 and I have never requested that this be raised. One lesson of my 1.1.07 disaster was that even with a decent sized bank, I am not comfortable losing that much money. (Other lessons learned that day were that when trading football in-play, I do not have an edge, and certainly not one that can justify a risk of £5,000, and that when I have a bad loss, I get very tired and have to take a nap). 

BigAl mentions a couple of things that I have addressed previously, but will clarify again. Al says - "many professionals will regularly dip into their funds to provide an income stream but your comments in your post clearly guard against this in your own circumstances". I am not, never have been, and never will be, professional in the sense that I need any of the profits to cover living expenses. I like to think that I have taken a 'professional' approach to sports investing, i.e. keeping decent records, trying new ideas, maintaining discipline etc., but I have never 'needed' to withdraw money. However, with a self-imposed limit of £5,000, there's no point in leaving significantly more than that in a non-interest earning account, as it is money that will never be used. I typically maintain between £15k and £25k in my account, enough to absorb a few large hits, but not too much. Most of the withdrawn money has ended up in more traditional investments, but it has on occasion been used for a decent vacation. None of it has been used to pay for groceries or utilities. It is withdrawing money from a betting account for these reasons that I caution against. I have no problem with withdrawing money once your bank has essentially outgrown itself, but a betting bankroll should not be needed for food and lodging expenses. The need to make money means you make less than optimal decisions as I have commented on in the past. 

The second clarification is on the 2% staking level that BigAl states I suggest. As regular readers of the blog will know, this is my suggestion for punts. In-play trading is where significant edges can be found, and I have never suggested, or meant to imply, that a self-imposed limit be applied when a big value in-play opportunity presents itself. As some screenshots have shown in the past, my exposure often gets close to the £5k limit, which with my average bank size is clearly a lot more than 2%, and I have often spoken of the low-risk / high-reward nature of laying large amounts at short prices when the market has overreacted. In summary, the chances of you finding an edge pre-game are a lot less than they are in-play - at least in the less popular or understood sports, so a limit of 2% makes sense for the former, with a more flexible limit for the latter. It works for me. Any FTL bets do not exceed 2% of the bankroll, and several of those markets are anything but highly liquid, where getting even 0.2% matched is often a challenge. The disappointing, and out of line with expected, returns on those markets mean that's a good thing!  

Now one might say that with a verifiable edge, why don't I remove the £5k limit and risk ever higher amounts? One reason is the already stated observation that I am not comfortable risking more than this, but there is another factor - the Betfair markets I am most interested in are not huge. On the Los Angeles Lakers v Utah Jazz game last night, an hour before tip-off and a total of ~£24k had been matched, and the amounts available were:

Although this screenshot is taken pre-game, it's not dissimilar to what you might find during a break in the game. While the clock is running, the amounts are usually less, and the gaps wider, and so it is clear that £5k is almost always more than enough relative to what the market will offer. To his credit, BigAl does recognise this possibility in his comment. One other note here is that my strategy is often to lay at sub 1.1 prices, and at these prices, £5k is usually more than enough. 

BigAl also mentions the increase in charges, and since September 2008, I have been paying the 20% which will clearly make a difference, although nothing compared with the impact of the 50% charges introduced in October. In fact, the mere looming of this charge has meant that profits on Betfair are lower than ever, as I have moved some activity elsewhere. I was facing my own fiscal cliff for much of this year, before proceeding over it. 

One final thought is that all months or years are not the same. I have written before that my best time of year is typically October through March when I make 68% of my profits, and also that edges can disappear overnight. Back in 2006, my top earning sport was golf, and it's no surprise that I don't mention it much these days, because I have found profits harder and harder to come by, and other sports have became more attractive. Trading a golf tournament for four days and handing over 50% is not an appealing idea. 

And markets do not remain inefficient for ever. The NHL used to be a strength, but the markets on Betfair are now all but dead, and there's no NHL anyway these days! 2006 was the last time I made four figures on that sport. Baseball was top in 2007, and peaked in 2009 with five figures in profits, but again the markets have weakened. The NBA had a strike last season, all these are factors that discredit the idea that bankroll growth is steady. It is anything but, and the need to constantly find new edges in different markets, if not sports, is clear. I may have to turn to horse racing soon, at which point I shall know it is time to move on and get a life.    

BigAl calls it a massive slowdown in % terms, and of course it is - 1000% growth is unsustainable, but having reached my limit relative to the size of the markets, and after adjusting for increased charges and other outside influences, returns have been relatively steady over the years in real terms.    

Conservative Study Finding

Intelligence Institute Study shows Fox News viewers have an IQ that is 20 points lower than the U.S. National average.

Study shows that the Americans who watch Fox News have an average IQ of 80, whereas the national average is 100. Researchers were not "shocked" by findings.

The results of a 4 year study show that Americans who obtain their news from Fox News channel have an average IQ of 80, which represents a 20 point deficit when compared to the U.S. national average of 100. IQ, or intelligence quotient, is the international standard of assessing intelligence.

Researchers at The Intelligence Institute, a conservative non-profit group, tested 5,000 people using a series of tests that measure everything from cognitive aptitude to common sense and found that people who identified themselves as Fox News viewers and 'conservative' had, on average, significantly lower intelligent quotients. Fox Viewers represented 2,650 members of the test group.

One test involved showing subjects a series of images and measuring their vitals, namely pulse rate and blood pressure. The self-identified conservatives' vitals increased over 35% when shown complex or shocking images. The image that caused the most stress was a poorly edited picture of President Obama standing next to a "ghostly" image of a child holding a tarantula.

Test subjects who received their news from other outlets or reported they do not watch the news scored an average IQ of 104, compared to 80 for Fox News viewers.

Lead researcher, P. Nichols, explains, "Less intelligent animals rely on instinct when confronted by something which they do not understand. This is an ancient survival reaction all animals, including humans, exhibit. It's a very simple phenomenon, really; think about a dog being afraid of a vacuum cleaner. He doesn't know what a vacuum is or if it may harm him, so he becomes agitated and barks at it. Less intelligent humans do the same thing. Concepts that are too complex for them to understand, may frighten or anger them."

He continues, "Fox News' content is presented at an elementary school level and plays directly into the fears of the less educated and less intelligent."

The researchers said that an IQ of 80 is well above the score of 70, which is where psychiatrists diagnose mental retardation. P. Nichols says an IQ of 80 will not limit anyone's ability to lead happy, fulfilling lives.

The study did not conclude if Fox News contributed to lowering IQ or if it attracts less intelligent humans.

P. Nichols concludes that he wasn't shocked by the studies' results, rather how dramatic their range. "Several previous studies show that self-identified conservatives are less intelligent than self-identified moderates. We have never seen such a homogeneous group teetering so close to special needs levels."

Dyadic Interactions

My old friend Scott emails me some fascinating links, lately often related to the US Presidential Election and Nate Silver, and most recently pointing me to a study of Basketball Teams as Strategic Networks. I love reading this stuff, and Moneyball was an excellent book as I have written before (and the movie was surprisingly good too), but while any background information on a sport is interesting, I'm still not convinced that in dynamic interaction sports such as football and basketball, there is much to be learned from a gambling perspective. A dyadic interaction sport such as baseball is completely different in nature, with each play commencing from a set state, of which there are a finite number. In baseball for example, the historical probability of a team winning from each state is known, so that, for example, in a tied game going in to the bottom of the ninth inning, the home team has gone on to win 63.9% of the time. Combine this data with the pitcher and  the batters he will face, and you have meaningful and useful data for trading.

It's not the same in most other sports. Football is just too fluid, played on a large playing surface, and it's only with a penalty where the dyadic interaction 'kicks' in. Basketball has its free-throws for which a player's percentages are available, but the game is so fast that profiting in-play by recognising a certain pattern of build up is, at least for me, not a realistic proposition. The benefits of these studies to managers and coaches are clear, but to the trader, not clear at all. There's a reason why chess matches don't draw crowds of 100,000. It's the unpredictable nature of football, and to a lesser extent, basketball, that keeps us all watching.

Betslayer stopped by to say: 

Still the best betting blog in town, just a quick one I am rapidly approaching the cliff edge of the 2nd Premium Charge....how have you coped with the 50%? Are you finding ways to adapt? I must admit I was horrified to get 20% but its actually made no difference to my profits.

First of all, congratulations on reaching this milestone, but in my case it's been more of a millstone. My optimism that I would be able to deal with the 50% rate has faded, and my Betfair activity is now much reduced. Unless I'm pretty sure I won't be paying the Premium Charge, there is no point me using Betfair for punts, and as most readers are aware, the most lucrative markets are in-play ones, and only available on Betfair. BETDAQ are either not turning markets in-play or have feeble liquidity. The result is that most punts are now made elsewhere, and my in-play trading has become very selective. There's an opportunity cost to trading, and while it was worth it to me to give up discretionary time to play for £100 or so a day, it is becoming apparent that investing the same time for approximately half of that amount just isn't worth it. I have other projects and activities that I would rather devote my time to, and this was always about the challenge, not the money. Fortunately, giving up the day job was never an option. While the Super PC threshold may seem a big number, when you divide it by an average salary, it's not that many years before you are going to hit it. 

The seven day Premium Charge cycle is jut too short as well. This scenario is deflating to say the least - you have a good win on a Sunday, £2k perhaps on the NFL, but then you lose the same amount on the Monday game. When the PC is taken on Wednesday, your balance is now down £3k from the Sunday high. If you weren't feeling bad enough about that Monday loss already, it's a kick in the groin when you get that PC deduction. To climb back to the high now needs a £6k win. Of course, were the win and loss each 24 hours earlier, i.e. in the same PC week, you would be £1k better off than under the first scenario, and that is both ridiculous and very irritating. I have heard that Betfair may be looking at a three month cycle, and this would certainly help, but the charge, and the way in which it was introduced with no opportunity to reset for established accounts, both leave a bad taste.

There were also a couple of comments from our old friend BigAl. It's perhaps something of a surprise that he still reads this fine blog given some of his previous comments and abuse, and I'm sure he will understand if I choose to rise above his provocations moving forward. 

Growing The Mighty Oak

With Newcastle United handily beating Wigan Athletic last night, the table is set, and we now have over 1,000 selections in the bag. Sadly backing everything would show a loss to level stakes of a little under 5%, but 13 of the 30 are in profit, although with just four selections, the Bundeslayga system should probably be ignored for now.

Neil closed the gap on Premier Betting and Premier Edge ended in the green after their Newcastle picks both won. Most of the others I covered yesterday, but one big anomaly is the fate of the XX Extended family of bets. With four in the top six spots and a relatively large number of selections, it's strange to see the Under 1.5 bets in 30th and, rather embarrassingly, last place. The Classic picks are all grouped fairly closely, but the Bundesliga selections are evenly spaced out from 2nd down to 28th, which is also odd.

I mentioned Betfair Banter's plans to move towards Draw finding, and Stu commented:

Betfair Banter may be about to board a sinking ship. This season's Premier League results have included 33% draws, considerably higher than the 26% average of the last four seasons. (Highest was 29.2%). There doesn't have to be a "correction" of course, but I would not bet against it.

The most EPL draws in the last 10 seasons is 111 from 380 matches in 2010-11, 29.2% as Stu said. The lowest in the last 10 seasons is 2005-06's 77 (20.3%) with he average (mean) being 98 (25.8%). With 49 draws in the bag already, another 49 would see the average hit, and with 231 matches remaining, that's a strike rate of 21% which is almost certain to be reached. The long term average from here on in would actually leave the season total a little under the high watermark of 111. If Webbo is expecting the draws to continue at 33%, he's likely to be disappointed, but what is key is not the strike rate, but the returns on the draws. 33% of draws at 3.0 isn't as notable as 20% of draws at 6.0.

In reference to my last post, Chris asked:

Hi Cassini, where did you get Table 1 if you don't mind me asking?

Not at all. It comes from a 2011 study by Alexis Direr of the Laboratoire dí Economie dí OrlÈans (LEO) and Paris School of Economics, rue de Blois:

This article investigates the degree of efficiency of the European Football on-line betting market by using odds quoted by 12 bookmakers on 21 European championships over 11 years. We show that systematically picking out odds inferior to a threshold delivers a rate of return of 4.45% if best odds are selected across bookmakers and 2.78% if mean odds are used. This amounts to backing overwhelming favourites whose probability of winning exceeds 90%. Our results only exploit information contained in odds, are robust to the use of real-time data and different sample periods and hold under risk neutrality and expected utility preferences for realistic degrees of risk aversion. Transaction costs reduce profitability but only for small stake bets.

I have the pdf if you can't find the study on the web.

And finally, yesterday was one of this blogs busiest days ever. The reason? Ian Erskine sent out an email which he does fairly often, but this one contained a link to his latest blog post, and a teaser for myself that:

(Cassini if you get this email you will want to read todays post)

I almost fell out of my chair, and not because of the missing apostrophe in "today's"! The reference to myself was in a paragraph about starting banks. Ian wrote:

Over time we will break down the myths of what is required but my starting point today is bank. I bet long before Betfair and lost fortunes. My first deposit into Betfair was £50. My initial starting cash to get where I am today is less than £1000. You can believe that or not believe it. I will give you a further example Cassini who operates Greenallover which is an entertaining blog started with I believe £98 and has never made another deposit, went down to £20 odd quid and now pays 60% Premium charge, focusing mainly on American sports betting and trading (see Cassini nice plug, if you wish to pay back the favour by means of a few articles for FTS it would be welcome, I ask publicly as I did email you but had no reply).

A couple of minor corrections - one is that the starting bank was £98.50 (the deposit of £100 was charged a processing fee of £1.50 and I remember it well) and I pay 50% Premium Charge, not 60%, but the point is correct. A small bank is all you need. If you have patience and an edge, a small bank inevitably becomes a large bank. Patience means 'don't make withdrawals', and stake sensibly.

My £98.50 was deposited in April 2004, and at the end of August 2005 had shrunk to £93.11 via the £20 that Ian mentioned. While that 16 months wasn't exactly lucrative, it didn't matter. Studying the markets was a hobby, the money was nothing.  As with most apprenticeships, the financial reward is secondary to the benefits of the learning experience. Once you've found an edge, the only thing between you and steady profits is you. Note that edges don't last forever though, markets are smart, so the ability to adapt is essential.

2012

Ian - I shall be in touch regarding returning the favour. I must be the rudest blogger in town ignoring emails, but in my defence, I have a lot going on most of the time.

And Then There Were 30

There are still some games to come this weekend, so no table update yet, but as things stand right now, it's been a poor start to December for the XX Draws. Two of the four Classic selections this weekend lost out to late goals (86' and 89') and the 86th minute was also cursed on Saturday with two Extended selections becoming losers at this time (West Ham United v Chelsea and Nice v Paris St Germain). The latter was particularly galling. When a game is goalless after 75 minutes, it's hard not to expect more than a win on the U3.5 market, although the HT0-0 was already in of course. One draw from four in the Bundesliga for a small loss overall, and I have decided to add the Bundeslayga system to the FTL starting this week. That makes it a nice round 30 entries, although some of the bigger losers appear to have given up. Some of you may have seen a post I wrote on this system earlier in the weekend, but I took the rare step of deleting it because it was misleading - one of my formulas was wrong and exaggerated the returns! Here is how this system came out on its official debut this weekend:

Not a bad start, and as FTL entrant Tony appears to have given up with his lays, this system will fill that void. The Bundeslayga system combines the 'wacky' nature of the Bundesliga with the findings of studies that show opposing odds-on favourites (excluding very short priced favourites which are profitable to back) is a winning strategy.

Pete Nordsted's Drawmaster selections are all done, and he made a small profit with one winner from three selections, the draw in the QPR v Aston Villa game.

Peter Nordsted's Premier Betting had a good midweek and weekend, and is now up 4.72 points on the season.

Whatever happens in the Newcastle United v Wigan Athletic game tomorrow night, Neil is in profit for the weekend. The only question is by how much - he has Newcastle to win.

And newly into profit after a stellar six out of six winners midweek is Premier Edge. The midweek form continued into the weekend, where the winning run extended to nine, before a couple of setbacks were met. Whether or not the final games of the weekend see them end in profit or loss, it's certainly been a good week for them. They also have Newcastle to win tomorrow night.

Little Al couldn't find any draws, and drops back into the red by a little, and the Free Under / Over Soccer picks had another losing weekend.

And then there is Football Elite who extended their losses with just the one winner and one 'no-bet' from five selections.

After a profit in midweek, Talkbet Jon lost his gains back this weekend, and remains in the red, although off his historic lows, and backing the draw in Ian Erskine's selected matches finally produced a couple of winners this week - after 19 losers we hit two in a row, and three in four, as tends to happen with these things!

As mentioned earlier, no selections from Tony, and none from Griff or CKL either.

Back to draws for a moment, and the attraction for betting on these for those who don't need a high strike rate has reached webbo at Betfair Banter, who has seen the light and has now devoted his life to draw finding, or at least the next few weeks.

Newcastle seem a popular pick tomorrow night to beat Wigan, and from the perspective of my own Elo ratings, this is an interesting game. Newcastle and Wigan both overachieved last season, and at the start of this season were ranked 9th and 8th respectively. It's no surprise that going in to this weekend, this over-achieving has caught up with them both, and now Newcastle are ranked 15th with Wigan even lower at 18th. My numbers for this game have Newcastle at 1.9 so the 2.1 available suggests some value. A more sweeping look at other EPL games matching this 'profile' shows that the Newcastle price should be around 1.935.

Away from football now, and a strange new market appeared in the NBA last night - Portland Trailblazers @ Cleveland Cavaliers featured a +0.5 market, which of course makes absolutely no sense as the draw is not an option. Your Premium Charges at work.