Saturday 24 August 2019

Relatively High

While the Basic Unders System doesn't generate as many selections as its Basic Overs sibling, their ROIs this season are fairly close.

Unders is at +12.9% from 138 selections, while Overs is at +11.5% from 278 selections. 

Basic Overs is currently in a slump, seven losses from seven over the last four days, with a first losing month of the season looking highly probable. 

It's the 9.5 total that is causing the problem. 

I apply a similar strategy in baseball to that for the the NBA. When the total is (relatively) high, my expectation is that the public will be deterred from backing Overs, and Overs can thus offer value.
If you've looked into the idea of backing Overs in the MLB, you'll know that where to draw the line regarding what qualifies as a 'relatively high' total can be a tough call. 

Until this season, the 9.5 total was profitable, but markets tend to correct, if slowly. The average total in games when the road team is favorite is 9.1 this season, so 9.5 is no longer 'relatively high'. Compare with the average of 7.9 runs in 2014 and 2015. 

In 2016, the ROI on the 9.5 total Overs was 22%. In 2017 it dropped to 6.3% and in 2018 to 0.9%. The trend was clear, and after losses standing at -13.8% before the all-star break it was clear a change was needed. 

The basic Unders system, on the other hand, is running hot, with ten wins from the last eleven selections, the latest being the 14 inning game between the New York Mets and the Braves which ended 1-2 with 12 of the 14 scoreless.
College Football returns today, although Week One has no Power Five or Group of Five Conference games that look likely to qualify for the Small Road 'Dogs System that has performed so well this millennium with 18 years of winners exceeding losers. 

Thursday 22 August 2019

Cornered Tigers Pounce

As most of you will have seen by now, the hottest favourite ever in an MLB game lost last night, just 10 days after losing as a -430 favourite.

The official database price on the Houston Astros last night was -470, but I saw it said that it touched -600 (1.167) at the MGM sportsbook in Las Vegas:

The move to -600 appears to have been driven by a fairly substantial bet at -575 (1.174)
I did caution against trying to buy money yesterday, but maybe not everyone reads this blog. 

The Astros never led last night, losing 1-2, so anyone following my Unders recommendation will be happy today. 

My other selection yesterday, the Washington Nationals, also had a comfortable (11-1) win at -150 (1.667), although 1.72 was available on the exchanges.

Having set a new record yesterday at -470, we don't have to wait long for a possible new record, with the same Houston Astros again currently listed as huge favourites at -500 to beat the Detroit Tigers in game four of their series, although there appears to be some reluctance (hardly surprising) to back at that price and it may well drift to around -440. 

Wherever it ends up, this is the fifth -400 or shorter line this month, and the third in four days, beating 2004's three in five days. Exciting times, unless you are an Astros fan perhaps:
Note the different prices quoted (above are from BetLabs), one of the constant challenges with multiple and dynamic markets, although not a big deal on outlier selections such as these that, on average, happen once or twice a season, or five times a month...  

Wednesday 21 August 2019

Astros Close in on -500 and NFL Week One Preview

With a current line of -470 (1.213), in from -420 (1.238) yesterday, the Houston Astros (Justin Verlander) are in line to be the hottest favourites in database history (i.e. since 2004) when they play the Detroit Tigers today. They've traded at 1.2 on Betfair, so -500 isn't an impossibility when the line is updated! 

The current record for the shortest price was jointly held at -460 by the Boston Red Sox who lost in 2-5 in 2004 to the Tampa Bay Rays (or Devil Rays as they were then known), and the Cleveland Indians in 2017, who beat today's underdog Detroit Tigers 2-0. 

This is only the 21st time a team has been as short as -400 since 2004, and the fourth time for Justin Verlander, which is a record over Clayton Kershaw who has been this favoured just the three times.

Of the 20 completed matches at -400 or shorter, the 'dog has won five, with two others going to extra innings, so as we saw as recently as the 11th of this month, a short price doesn't guarantee anything.

I should point out that from the National League, the big favourite has an unblemished record. 

While the sample size is too small to be confident in anything, the evidence suggests that favourite backers are not buying money.

The 88 matches at the -350 or shorter price do show a profit on the favourite, at least on the Money Line, with an ROI of 2%, but a loss on the Run Line, but again, if looking only at National League teams, the ROIs are 14.4% and 5.1% respectively, while they are in the red for American League teams. 

A safer bet appears to be Unders. From the 88 games, Unders has an ROI of 12.1% spread evenly across both leagues.

The T-Bone system (ROI this season ML 11.7%, RL 10.7%) looks to have a qualifier today with the Washington Nationals at -150 (1.667), while we also have an Overs (ROI 14.2% YTD) selection

Speaking of 'dogs, with the NFL season fast approaching, some of you may have seen the talk on Twitter and no doubt elsewhere on where the value may be for the Week One matches.

The premise is that the off-season is disruptive, and that the public bias is to rate teams based on old form or data meaning that favourites may be over-rated and 'dogs under-rated.

For the past five seasons, Week One 'dogs have a 54.8% record ATS overall, but in Divisional matches, for which there are five this year, the number is 81%. 

In the AFC, the record is 90%, with the one loser (New York Jets @ Buffalo Bills in 2017) failing to cover by a single point.

Divisional opponents are faced twice every regular season, and are the NFL equivalent of a football six-pointer - a win isn't just a win, it denies your divisional rival a win, so the stakes are higher in these games. Add in that the teams tend to know the weaknesses and strengths of a divisional rival better than they do a team they might face every few years, and it's not unreasonable to treat these games as a separate category.

Readers should be familiar with my Small Road 'Dogs strategy which over the past five seasons has a winning record of 54%, but which in Divisional games is 55.8%, and then if you look only at NFC games played on grass, you're at 80.8%, but that's perhaps for another day. 

Monday 19 August 2019

On The Inside

One of the more interesting Twitter accounts I follow is A Lucky A Day and you can all thank him for being the trigger behind today's post.

While most people talk about the betting market for an event in singular terms, Lucky challenged this idea writing in response to a comment about "beating the market":

He is, in my opinion, totally correct. Each sportsbook has its own market, with prices driven by supply and demand, which vary from book to book, resulting in differing lines and prices. 

Because there may be several markets for the same event, spread across multiple jurisdictions, each with their own regulations and rules, it seems quite reasonable to say that there is not a single market. Soft books and sharp books also offer different markets, in that who is allowed to be active varies.

As Lucky further explained:
There are several different prices depending on the market place. Because there is not free movement of money across these markets and the terms/regulations of each market are different I think the concept of "the market" is flawed.
If there was one single market, then clearly the avenue of arbitrage which many people seek to exploit would not exist. 

The only pure markets are those on the exchanges, where supply meets demand, and anyone can play. The days of significant arbitrage opportunities between them are also long gone, with a change in prices on one exchange being almost instantaneously reflected on others.

Lucky's point on the market letting you play is also critical. Not everyone has access to every market, so beating a market that you don't have access to might be academically pleasing, but from a financial point of view, it's pointless. It's a waste of time, and time is money. 

Coincidentally, I had an exchange with Lucky last night in which the topic of 'insider information' was touched upon.

As readers of this blog will know, I do not get involved in horse racing following the personal revelation almost 40 years ago, that insiders possess rather vital information such as that a horse will not win

As with all sports, this is far less of a concern the more prestigious or valuable the event, but at the lower end of the scale, insider information is a risk. You would certainly expect insiders to make the most of their inherent advantage, but that advantage means the outsider is at a disadvantage. 

The traditional term 'insider information' comes from the world of finance and is defined thus:
Insider information is a non-public fact regarding the plans or conditions of a publicly-traded company that could provide a financial advantage in a securities market.
Substitute "horse" for "publicly-traded company" and "betting" for "securities" and you have the horse racing equivalent. If the stable, and those close to the stable, know that their horse is not being run to win, that is "insider information".

A further explanation of "insider information" states:
Insider information is a non-public fact regarding the plans or condition of a publicly-traded company that could provide a financial advantage when used to buy or sell shares of that or another company's securities. Knowing about a company's significant, confidential corporate developments, such as the release of a new product, could provide an unfair advantage if the information is not public and only a few people know about the developments. Insider information is typically gained by someone who is working within or close to a listed company.
Again, it's not difficult to modify this to fit the horse racing markets, or any other individual sport market for that matter. However you frame it, "insider" means "someone on the inside, or someone with privileged information", and definitely not this:
By definition, insider information is the opposite of public. It's private. 

If someone with no horse racing connections is in their parent's basement and has information on a horse race, it isn't insider information. It makes no difference to the definition of an "insider" how efficient the market is, an insider is someone on the inside. 

Similarly, information that is publicly available is public, even if it might be hard to discover. 

Tuesday 13 August 2019

Curiousities and Carryovers

The Houston Astros were the sole T-Bone qualifier last night, but the game was postponed due to rain, which means they are a qualifier today, but be careful. The Astros @ Chicago White Sox game is a double-header, so the qualifier is the first game (4:40 EDT) with Zack Greinke pitching, not the second game with Gerrit Cole on the mound. This second game will be a T-Bone selection in the unlikely event that the Astros lose game 1, but as we saw on Sunday, anything is possible.

The postponed game was also a basic overs selection, but with a quality starter, backing Overs is generally not the best idea.

The only bet yesterday ended up being the Overs in the Boston Red Sox @ Cleveland Indians game and this squeaked in by one run scored in the bottom on the ninth, showing why you want that extra inning on your side.

I've mentioned the importance of having a logical rationale to systems, and one Tweet yesterday was this:

AL teams that have had six straight games of 6 or fewer hits have gone a perfect 9-0 since 2004. ACTIVE on the #Rangers tonight vs. #BlueJays.
My response was this:
The Texas Rangers lost 4-19.

Somewhat related, and there's a similar suggestion for a system today that suggests opposing teams who scored more than 15 runs in their previous game. 

The premise is correct in that the system is not based on the team performing poorly in the next game, but that it takes advantage of the public bias influenced by the recency effect, that the hot hitting will continue.

In database history, the record of opposing these big hitters is 164-167, not a huge sample, but profitable with an ROI of 3.8% on the Money Line.

When the opposition were the team on the wrong side of the high score in the same series, the ROI climbs to 4.5%, and when the team were embarrassed on the road, the ROI jumps to 12.2%, or 20.8% in night games. 

The Texas Rangers qualify today at around -129 (1.77). 

Monday 12 August 2019

Blind Squirrels

I briefly mentioned the Houston Astros in last night's post, and the fact that their price of -450 (1.22) was the shortest for a road team going back to at least the 2004 season.

The Astros are currently favourites to win the World Series at a best price of 3.79, and were playing the Baltimore Orioles whose haplessness was also mentioned in a recent post and who are currently the second worst team in baseball. Cue one of the biggest  upsets in MLB history, with Yahoo!Sports reporting it like this:

The Houston Astros are a very good baseball team. The Baltimore Orioles are ... not a very good baseball team.
Still, over the course of a 162-game season, every blind squirrel finds an acorn. And even the Orioles (39-78) can beat the Astros (77-41). Sunday’s comeback win by the Orioles marked one of the biggest upsets in MLB history
According to ESPN, the Orioles were listed as +420 underdogs by some sportsbooks. Some of it might have had to do with Houston’s 23-3 rout on Saturday and the fact that ace Justin Verlander was starting for the Astros.
Per The Action Network, it was the biggest moneyline upset since the Washington Nationals upset the Minnesota Twins in 2007 at +395 odds.
The article concludes by saying that the result...
...takes nothing away from the fact that the Orioles and Astros are on opposite ends of the MLB totem pole, but it goes to show that in sports — but especially baseball — anything can happen.
As has been previously mentioned in this blog, baseball is a sport where it is unusual for a winning team to win more than two-thirds of their matches, and similarly for a losing team to lose more than one-third of games, although the Orioles did manage this last season. 

Coincidentally, the last team to go outside of the above boundaries was the Houston Astros in 2013, winning just 31.5% of their games in their first season in the American League (AL) after moving from the National League (NL). 

Trivia Fact: The Astros are the only team to have played in the World Series as a member of the AL (2017) and the NL (2005) 

Thus it is rare for a team (0.87%) to be priced at -300 (1.33) or shorter, and extremely rare (0.05%) to be priced -400 (1.25) or shorter. 

As the table below shows, the frequency has increased in recent seasons, and I have broken down the numbers still further for the ultra-shorties, together with the number of matches with such prices:

ROIs from small sample sizes are pretty much meaningless, but if you are seriously interested in when these hot favourites typically offer value, look at teams playing at Home in a Divisional game versus opponents they have just beaten in the series.

In the 2016-19 four season period, the 36 selections are +15.3% (Money Line) and +17.2% (Run Line).
Interesting that the same Baltimore Orioles have brought about the most recent three of the four upsets. Note also that there are probably going to be several more selections this month and next, as the regular season reaches its climax.

Sunday 11 August 2019

MLB Favourites, In General

I'm not sure Betfair are being too selective with how they are targeting the above offer, nor do I think they are being too honest about the 5% commission given my Premium Charge status, but at least they suggest that "other charges may apply". Indeed they do. 

Having opened my account in 2004, this might have been interesting at that time, but 15 years on, there's not quite so much to learn, nor are the markets as giving as they once were. 

So no traveling for me this month as was previously the plan. The sick dog appears to be on the mend after surgery although, unlike her owner, she really needs to put on a couple of pounds. Holiday plans have now been rescheduled for late October, which means I have the rest of the baseball season pretty much uninterrupted. 

Unfortunately the Baseball Overs System had its worst day of the season on Friday, with none of the five selections successful. 

For the three systems I play combined, it was the busiest day of the season so far with 14 bets (including Money Line and Run Line), but an overall loss of 1.26 points. It happens. 

Yesterday was a lot quieter, but again, we were out of luck.

Even though consecutive losing days are a rarity, for anyone following along, perspective is always helpful, and 32 points or so in the last month and a season long ROI of 12.3% brightens the overall picture.

I received an email regarding baseball betting, espousing rationale for why value has tended towards favourites in recent seasons:
Handicapping baseball is so different than handicapping any other sport because of two factors. It is the most random of the major sports and because the predominant wager is the moneyline. The key to beating MLB is to find reliable favorites. When paying for selections, you certainly can feel "ripped off" when the service releases a bunch of 160 favorites - especially when they lose.

However, a serious, thorough investigation into the results reveals that betting favorites, in general, is the way to make money in baseball. The reason is simply that it is MUCH more satisfying to win with a 150 underdog than it is to win with a 150 favorite. In fact, when a handicappers wins with a 150 favorite, it is embarrassing to promote this success because a 150 favorite is supposed to win. When a handicapper wins with an underdog, he is blasting it all over social media.

Also, there are sports monitoring services that do not accept favorite of more than 200 and there are handicapping websites that do not allow their handicappers to release favorites of more than 200. The reason is simply that customers are much more forgiving if a handicapper loses with underdogs than with favorites. When a customer pays for three plays and they are all 150-plus favorites and they go 1-2, the customer is likely to never buy picks from that handicapper again.
All of this has tilted the value towards favorites.
I guess the term 'handicapper' sounds better than 'tipster', and I'm not sure who finds it embarrassing to generate a long term profit from short priced selections. Anyone following the free basic T-Bone System is up 5.6% on the Money Line, 6.8% on the Run Line since 2016.  

I don't follow any monitoring services but the idea of not accepting selections because they have a higher probability of winning is nonsensical. 

A selection can offer value at any price and the the favourite-longshot bias means that, on average, bettors have a tendency to overvalue "long shots" and undervalue favourites.

Readers of this blog will know that over the past few seasons, the supposedly "permanent" value on outsiders observed over many years, has moved to favourites:
"In contrast to the consistently observed favourite–longshot bias found in racetrack betting markets, it has been shown that gamblers in the market for Major League Baseball games reveal the opposite behaviour. This paper updates the previous study with ten years of additional data for the 1990–99 seasons. The strength of the reverse favourite–longshot bias is virtually identical to the original paper. The result suggests that, contrary to most reported inefficiencies in gambling markets, this bias appears to be permanent."
Nothing in betting markets should be permanent, although some biases are certainly stubborn.

The -200 (1.5) price is coincidentally one that I use for my 'shorties' system, the results of which are regularly updated here, and it does show the value over recent seasons:
All favorites are not the same of course, and returns from the hotter favourites in baseball exceed those of the weaker favourites.

This chart might be interesting to some of you, showing the outcomes from betting favourites from weak to strong, with the seasons broken down into four year intervals to show changes in the markets over time:
Note that there are no Run Line prices available for 2004 to 2006, and that this is correct to the end of Saturday.

Coincidentally, given that we are discussing favourites, the Houston Astros today became the shortest ever (since 2004 anyway) priced road team at around -430 but they lost 7-8. I saw that the price touched -500 at some places, but in a game like baseball, you really are playing with fire at those odds. Fortunately this match was an Overs qualifier for me, but as the chart above shows, there's a point at which backing the favourite does not offer value. 

The writer of the email states that:
...a serious, thorough investigation into the results reveals that betting favorites, in general, is the way to make money in baseball. 
It's the "in general" qualifier that makes all the difference of course.

Finally, while no one seems to leave comments on blogs any more, I do get emails from time to time. One such email today from Rob was appreciated, and he wrote:
Love all the stuff on your blog and that you are so helpful to like minded punters.
Just looking at your draw state you look to back draws when there is less than 25% difference between the teams. Using today as an example Leicester 42% win chance Vs Wolves 29.4% win chance. Am I right that this would qualify as approx 13% between each teams win chance %?
Is it really that simple ?
The numbers for win probability I had were Leicester City at 0.435 and Wolverhampton Wanderers at 0.269, but the way I calculate the difference between them is by dividing the raw difference of 0.166 (0.435 - 0.269) by the higher probability, e.g. 0.166 / 0.435 = 38%.

It's not complicated, but an easier strategy using just the eye is to back the Draw in EPL games where every raw price (Home, Draw and Away) starts with a 2 or a 3, i.e. 2.0 to 3.99. 

Since the Pinnacle era, this has an ROI of over 3%, but when an outlier of a season comes along as was the case last season, you're going to take a hit. 

Incidentally, Draws in the EPL this season are currently hitting at a 30% strike rate, the highest since 1998-99! 

Tuesday 6 August 2019

New Season Draws Closer

With one of my dogs not having eaten in six days, my travel plans are currently on hold, and while that comes at quite a financial cost to me, the lucky beneficiaries are you readers, as I find myself with some unexpected time on my hands.

Football Investor tweeted the above yesterday, and as probably the world's leading expert on the Draw in football, of course my interest was piqued. 

Unfortunately Mr Investor was a little disingenuous in that he doesn't actually say what the profitable strategy is, but merely shows the results of this strategy, but the selections and results look very similar to the Draw strategies that I have suggested over the years.

When you are looking for value with the Draw, you need to look for matches that are between teams with similar win probabilities.

Footy goes back to the 2012-13 season, which was the first to feature Pinnacle's prices, and he looks at the English Premier League, the top two leagues in Spain, and Serie B, all of which have been extensively covered in this blog previously.

Investor's selection criteria appear to change from league to league, and there is nothing wrong with that since each league has its own personality as I have written before.

For example, last season he had 230 selections from Serie B, 90 in the EPL, 47 in La Liga and 21 in La Liga 2. 

Serie B was last reviewed in June and included these numbers showing the profitability of the Draw in 'close' matches, i.e matches where the 'true' difference between the two teams is 25% or less.

2016-17: ROI of 9.7% from 107 selections

2017-18: ROI of 24.8% from 105 selections

2018-19: ROI of 11.4% from 94 selections

These ROIs are very close to Footy's 8.8%, 17.6% and 9.9% returns respectively. Although the more selective criteria I use does generate a higher ROI, the higher number of selections Football Investor uses generates more points of profit. 

The Draw in Spain's second tier La Liga 2 last season was reviewed here back in June, and in 'close' matches had an ROI of 6.6% from 102 selections. Footy Investor's 21 selections had an ROI of 19.7%, but ROIs from such a small number of selections are almost meaningless.

Certainly there is value to be found in backing the Draw in competitive matches. David Sumpter's Big 6 idea worked well when the six teams were comparable, but we're in an era now where such matches see a 1.26 home favourite or a 1.5 away favourite. We may look back and talk about the Big 2 era. 

David's idea worked well, not because the six teams were the best, but because the majority of Big 6 matches met the following criteria:
It turns out that when two well-matched teams meet (i.e. the probability of a home win is only slightly bigger than the probability of away win) then draws are under-priced. When matches are skewed so there is a strong a favourite (i.e. the probability of one team or the other winning is larger than the other) then draws are over-priced.
Stick with this principle, and whether you look at the differences between win probabilities, or exclude hot favourites, you will improve your chances of making a profit.

Worth mentioning that one season is a small sample with a lot of variance, and when a season such as last season's EPL comes along, with Premier League records broken for average goals per game, most Away wins, and fewest Draws, even the best of Draw strategies will struggle. 

Sunday 4 August 2019

The Hottest Month

July was indeed a record month (going back to 2004) for hot favourites in MLB, with a combined 37.55 points from Money Line and Run Line bets.
Followers on the T-Bone system also saw record profits, with the Money Line and Run Line returns almost perfect with 18-2 and 16-4 records respectively.
With 37.40 points already banked, the 2012 record of 50.58 points is well within reach.
As for the Totals Systems, the Overs won 63% of matches while the Unders had a quiet month with just six selections, but all of which were winners. Note the correlation with the Money Line outcomes on these selections also:
July also features the Aestivation System which added another 11.35 points to an already full pot.

Total for the month: 95.7 points from 291 bets and about 10 minutes a day.

It'll be hard for August to improve upon these numbers, although it is typically a good month for hot favourites, but for now at least, the simple strategies are paying off hugely. I hope some of you are following along.

Saturday 3 August 2019

NFL Weak One and November Rain

I was literally on the plane when I saw A Lucky A Day's comment last weekend, and while I wasn't planning on spending my time this weekend on sports betting, one does what one can to keep one's readers happy. It's my last weekend before three weeks away, although plans are in a state of flux right now with one of my dogs getting sick while I was away for work last week, and currently hospitalised. If you've never owned a pet, it won't make sense that there is a possibility of cancelling a three week trip because of a sick dog, but some of you might understand. 

Anyway, the premise from Plus EV Analytics is that markets are inefficient early in the NFL season, and that backing big underdogs is a profitable exercise. As the positor concedes, one season doesn't prove anything but in 2018 the results were:
A +300 NFL 'dog is approximate to a team getting 7.5 points, but the selections above are teams getting 8.5 points or more, and backing these seven teams against the spread (ATS) would have produced a 5-2 winning record.

How about in previous seasons?

Looking at the past five seasons (2014-2018), the ATS record is 18-12, while over the past ten seasons (2009-2018) the ATS record is 40-38, which is a slender 0.1% ROI using Pinnacles -105 as our base. 

In Divisional games, the five and ten year records respectively are 8-2 and 18-11 which is a little more promising, and for Divisional games played on an artificial surface, the numbers are  3-1 and 10-4. 

To reset for a moment, since the current 32 team format was introduced in 2002, teams receiving more than 8 points have a losing (48.8%) record in the regular season.

Week One is, as might be expected, the most inefficient with a 17-9 ATS record for the big dogs. In Divisional games only, the record is 8-2 and in Divisional games on an artificial surface, the record is 100% from just the five games.

Here is a breakdown by week: 
It's interesting that the middle weeks of 8, 9 and 10 show an undervaluation of the underdog, especially on a grass surface, and the only rational explanation I can think of is that the bad weather is starting to impact the playing surface which drags the favourite down to the underdog's level!