Sportpunter

Previously we looked at how the Sportpunter AFL model had gone betting head to head. In this article we will look a little more in depth about how it goes betting on the line or with a handicap.

Traditionally, unlike Americans, Australians had always bet on any sport on the head. It makes perfect sense too; punters want to bet on whom they think is going to win as opposed to who is going to win by a certain margin. But more recently line betting has become more in vogue. This is because many of the head to head betting matchups are already almost over before they begin, with two many one sided matches making the betting uninteresting.

So line betting has made its way into Australian betting culture, and there are other advantages with it as well. Returns can be less varied as most odds are set to around 1.85 to 2.00, so wide fluctuations are less likely to occur. AFL is primed for line betting because the distribution of points throughout a game is relatively normally distributed.

Since 2005, we have been recording all the line bet odds, which isn’t as long as head to head betting, of which we had a lot more years of odds for. Throughout this time, head to head betting had made $5,330.07 profit at 5.3% ROI, whilst betting on the handicap has made twice as much: $10,716 profit for 8.2% ROI.  Many would say that this is a great sign that handicap betting is the way to go as not only is it more profitable but also more consistent.

So let’s have a look at how well the model had gone on line betting. The total results below differ slightly to that above as some matches in the database were added to the results post match, but are for obvious reasons not included in the betting history, but included in the analysis. As shown below, results have been very good. Not much can be talked about with regards to probabilities and odds with the exception that probabilities of over 75% didn’t do too well. Similarly, and interestingly, overlays greater than 50% lost. Sure there were only 8 bets, but this is strange. It may be more interesting to look at the actual lines to see if we can find a pattern that could possibly explain this.

Shown below is the home team handicap for betting on the home team and the away team. Home team betting has done very well, and those in the know will know that home ground advantage is much underrated in AFL, so there are no surprises here.

What is interesting however is that the model has failed to make a profit when betting on the line on the away team when they are favourite. In other words, when the line is zero or lower, 64 bets have been made for a loss of 24.9% ROI. This is interesting, and although the sample size is small, could well once again have something to do with the fact of the biased home ground advantage in the sport.

If this pattern continues to exist, then there could be a case for not betting on the line for teams that are favourites when away. Using this information could well enhance your betting profits even greater than the %ROI as shown above.

Either way, it goes to show that betting on AFL is a seriously profitable sport to bet on.

Sportpunter are getting subscriptions for their AFL Model now. Click here to register.

The AFL season is set to start again this Thursday and with my personal favourite model, we will be predicting from day 1. The tigers play Carlton in the traditional first up game, and we hope (or maybe we don’t) that it will a little closer game. IT may well be the year of the tiger, but any resemblance to Richmond is surely coincidental. Many people have put their stamp on the bulldogs to bark, and as punters we do indeed hope that the dogs – in another form – will.

But first up we will have a look at how the model has gone betting on the head to head, and determine exactly where the profits and losses can be further analysed.

Shown below is the analysis of Sportpunter’s AFL model from day 1, with exceptional results shown.

Profits are generally recorded throughout all probabilities and all odds, which is great too see. And just like any good model, higher overlays result in higher profits, with a return of 27.6% ROI from 250 bets for overlays greater than 35%.

It seems therefore that the model is right on the market with regards to head to head betting. But has the model done better betting on the line? We will look at that next.

Previously we looked at how the big outsiders have done well in NRL since 2002. Simply betting on all teams where the odds are greater then 4.00 would have netted a profit in excess of 19% ROI. But when does this occur? On home teams or away teams?

Shown below is a breakdown of how home and away teams have gone for underdogs (Odds greater or equal to 2) and for odds greater than 4.00

As you can see, simply betting on the home team when they are the underdog has netted a very nice 8.6% ROI from nearly 500 bets. Away underdogs has lost slightly. But interestingly, home teams with odds greater than 4.00 have lost slightly whilst the away team has gained. Of course, there are only 59 times when the home team has odds in excess of 4.00, as it is obviously a reasonably rare occasion.

Shown below is how the Sportpunter NRL model has gone since 2002.

As for how our model has gone, it is no surprise that it has done exceptionally well with the bigger odds. Shown below is how it has gone since 2002. Probabilities under 50% have made 24% ROI from 1517 bets. Betting at odds of 3.00 or more has made an incredible 27.4% ROI. Overlays of 40% or more have made 31% ROI from 254 bets.

The bias in the NRL is simply incredible. Simply by changing the way you bet, the possibilities are endless for NRL given the massive bias and a very strong model. Of course there might not be overlays on all outsiders, and the model shows that instead of making 19% ROI over the big overlays and odds, the model makes a lot more profit, and is independent of the bias, which in itself is a very important thing. In other words, if the market adjusts, the model will, and should still do well.

To sign up for the 2010 NRL Model, click here.

The NRL season is about to start this Friday. It’s looking to be a great season with Parramatta slight bookie favourites over last years premiership team Melbourne Storm. Sportpunter’s NRL model has done exceptionally well over the years, making over 10% ROI since 2002 on 900 bets.

The next few days we will be giving you all the analysis about the NRL model with the anticipation of the new season about to start. But to start off with, here is some amazing statistics about how one would have gone blind betting teams since 2003.  Blind betting basically means that we will back each team no matter what to win $100. How would we have gone in certain circumstances? The results are shown below.

The first table, as shown above looks at betting home teams or away teams. A 1.3% loss was made blind betting the home team, whilst a 5.2% loss was made blind betting the away team. Hence, the home ground advantage is probably a little underrated by the bookmakers and general public.

This is also shown in the results for when hosting a team travelling interstate or betting on a team that is travelling interstate. A greater profit, or perhaps I should say, a smaller loss, was made for the host team to a travelling team.

Interesting statistics, but not all that useful. What is useful is the statistics below which look at how we would have gone blind betting teams at certain odds.

Amazingly as shown above a profit was made betting all underdogs. That is blind betting all teams with odds of 2.00 or more, made a profit of 2.2% ROI. This is quite substantial, and the second of the two tables above looks at this race a little more closely.

Betting on teams with odds of 2.00 to 4.00 basically broke even, which obviously means that betting on teams with odds of greater than 4.00 has been amazingly successful. In fact, if one had bet every large underdog with odds of 4.00 or more, you would have made a profit of 19.4% ROI from 292 bets. On average that’s 36 bets per year.

The bias in the odds in massive. And we will be looking at certain circumstances where this might well be even more highlighted. What happens about big outsiders when home teams or when the opponent is travelling interstate. We will look at this in the next article, as well as an indepth analysis of Sportpunter’s NRL Model.

To sign up to the Sportpunter’s NRL model in 2010 click here.

With Davis Cup coming this weekend, I thought it was a good opportunity to find out how well the model goes. Many people would say that the Davis cup is a completely different game compared to normal tournaments, and if you have been to several countries which host it, I’m pretty sure that you will agree. Although this is the federation cup – which is the female equivalent to the Davis Cup – take a look at this clip which shows Maria Sharapova being heckled in Israel.

Ok, this is extremes, but the Davis Cup isn’t far behind as far as crowd fanaticism goes. Surely the home ground advantage is worth more in the Davis Cup than it is on any normal match. So how have we gone in the past blind betting as well as model betting on the home and away players?

I have data for only 168 matches, so not a lot should be read into it, but the finding are quite interesting. The home player averaged odds of 1.53 whilst the away player averaged odds of 2.72. Betting to win $100 on the home player means than $92k was bet for a profit of $1.6k for a 1.7% ROI. Whilst betting on away players $23k was bet for a profit of $319 or 1.4%. So, despite a lot more being bet on home players, largely due to very short odds, both home and away bets seem to be doing pretty good.

As far as the model goes. 63 bets were made for a 5.6% ROI betting on the home players, whilst 1.0% was made betting the away players over 23 bets. A lot less bets on the away players, but at least it shows that the model is on song at getting the home court advantage correct when it comes to Davis cup betting.

For all the tennis predictions, click here.

We previously looked a college basketball totals betting as well as focusing on unders betting. Today we will look at how well we have done suggesting bets on the overs.

The data below show our betting history for overs betting.

Similar properties exist to that of the analysis of unders betting. Almost 1/3^rd of the betting is done on the overs, and a smaller but still very good 3.4% ROI is made. Overlays greater than 30% resulted in no profit, although with a sample size of only 5 bets, nothing substantial can be made from this.

However what is more interesting is that for overlays of 15% or more, a 3.3% ROI loss was recorded. What does this mean? Well it means that higher overlays for overs betting does not necessarily result in higher profits. The reason, as discussed before could well be due to team changes and injuries. A team that changes with an injured team is more likely to score less than normal, and if bookmakers adjust for this, it could see how model with a larger overlay on the overs.

Betting with an overlay between 5% and 15% for overs resulted in a 6% ROI, which is very good.

So the conclusion to be made from betting college basketball? Be careful about the larger overlays. I wouldn’t recommend not betting them at all, but rather restricting the bet size.

Sportpunter’s free College Basketball predictions are shown here

We previously looked at College basketball totals betting and where and when it is profitable. Overall a 5.5% ROI was made, but interestingly some of the larger overlays failed to return a profit. Is this the case for over and unders betting? We will do the analysis and for unders it is shown below.

Over 2/3rds of the betting is done on unders betting and results are greater than that of overs. A 6.3% ROI is made over all, which is incredibly good considering the large turnover from over 2500 bets.

What is most interesting is that higher probabilities along with associated higher overlays still didn’t make a profit. We are talking about a small sample size again of 100 bets, but it would seem the model has certain variables that are not accounted for, which could explain the large overlays not being successful. These could be, amongst other things, injuries or returning players.

The current College Basketball model does not factor players into account, and gamblers might well want to be aware of any changes to teams. Even so, a decent profit is made betting the predictions as we give them. I would suggest smaller bet amounts for larger overlays.

We’ll have a look at the unders bets tomorrow.

Sportpunter’s free College Basketball predictions are shown here