Enjoy this sample article, which was first published in Smart Betting Club Issue#55 from November 2010. For more articles like this and the latest info on the Fink Tank be sure to take up a full Smart Betting Club membership.
Powerful Football Ratings: Fink Tank
Service Name: Fink Tank
Fink Tank is a free football predictor that is provided by Detech and sponsored by the (English) Times newspaper. It’s named after Daniel Finkelstein who writes regular columns in the Times, but was originally created by Henry Scott and Alex Morton.
We last wrote about the Fink Tank’s predictions towards the end of the last football season and have been blogging on our progress following the predictions this season via our blog https://smartbettingclub.com/blog/category/fink-tank/
As a quick summary, Fink Tank use shots on target as the basis for their prediction model and weight this based on certain factors such as home bias and other factors. With goals generally at a premium in football matches, a crossbar or dodgy penalty can mean a team undeservedly losing a game. A better judge of a team’s dominance and ability is apparently the ratio of shots on goal a team has and allows against it.
The reason is that there will be more shots on target than goals in a particular game, which provides you with more data by which to assess a team’s true strength. The more shots a team manages to create, more goals it is likely to score, so the theory goes.
Fink Tank provide their predictions in the form of estimated probabilities for the home win, draw and away win. You can then compare this to the available odds to assess whether a particular game offers value.
Taking this approach we previously assessed Fink Tank’s performance from 07 to 09 in the English Premier League.
Unfortunately the ratings did not prove too profitable, but there were a few caveats including there being a number of games missing and a shocking away performance in 07. Ultimately, though the ratings did ‘outsmart’ the performance of just backing home/ draw/ away matches blindly. This performance was also based on average bookie odds before kickoff, when bigger returns might have been possible by shopping around.
Fink Tank Value Betting Update
Our approach in testing the ratings this season has been to simply back an outcome whenever the odds available are better than what the odds should be according to Fink Tank. This means a mix of short priced favourites and long shots.
As we have had to track everything manually, we have been tracking the English Premier League only. We have been recording the best odds realistically available (No two-bit bookmakers).
Performance up to the end of November is as follows:
ROI: 25% ROI
The 2010/2011 season total runs as follows:
As you can see, the ratings have managed to pick out some good value, with a number of big priced winners making all the difference.
So far this season there have been a handful of big priced away team winners, which is quite the opposite of previous seasons. Thankfully, the Fink Tank ratings have ensured the value bets fell on the right side of these.
Highlights include Wigan beating Arsenal away at 13.59, Blackpool beating Newcastle away at 9.0, West Brom beating Arsenal at 18.50 and Newcastle beating Arsenal at 14.04. You might spot a north London trend there with Arsenal being at the receiving end of the shock results.
The home/ away split is as follows:
ROI: 1.2% ROI
ROI: 48.60% ROI
So far, the returns are bucking the long term trend. Indeed backing the away side blindly would have produced a small profit so far this season. What the Fink Tank ratings have done is spot the value in those selections.
In fact, it is the big priced home and away value picks that have carried the Fink Tank Predictions this season. Perhaps the bookmakers are over compensating after getting the strong home bias wrong last season. In 2010/2010, you could have made a profit blindly backing all home games in the Premier League.
Higher value, higher profits
If you were to have restricted yourself to games offering strong value, your profits would have been even better this season. Greater value means that there is a bigger difference between Fink Tank’s prediction for a home/ draw/ away win and the odds on offer at the bookmakers.
For example, restricting your bets to games with more than 10% value would have returned 48 points with an ROI of 57%.
The sample size is on the low side here, but it does help underline the importance of value betting. The size of the value appears to indicate the degree of error in the bookmaker’s pricing.
It’s also worth pointing out that value is not the same was probability. For example Fink Tank’s estimation of the probability of West Brom beating Arsenal away was just 8.2%. If the same match were to be played 100 times, West Brom would have been expected to win just 8 of those games. However, as the bookmaker odds were implying a probability of just 6.9%, there was value on offer (34% difference).
Not all of the big value picks have been at long odds, but its probably fair to say that there has been some good fortune this season in converting the big value bets into winners.
How to find value bets
It can be a little confusing learning how to convert Fink Tank’s probabilities into value bets, so we have created a Fink Tank value bet tool for you.
You can download this spreadsheet here: https://smartbettingclub.com/FinkTankValue.xls
Get the odds.
The first step is to record the odds on offer from bookmakers. You can use any method to do this, but we prefer to use www.oddsportal.com
Click on the league to view the summary for the forthcoming week.
This will display the summary for all those games including the average odds available. We suggest you create an account with oddsportal so you can remove the odds from any bookmakers you have no intention of using.
The odds will look a little something like this:
1 = Home win
2 = Away win
X = Draw
For speed, we record these average odds to quickly find out which way the value lies, but then click on each game to get the best odds once we’ve decided which bet we’re taking.
We can see that Wolves are 4.44 to beat Blackburn for example.
You don’t have to record the odds from all the upcoming games, this is just the method we’ve been using for tracking purposes. You can easily just record the odds for any particular game that you’re interested in examining.
Get the ratings.
The next step is to get the ratings from http://www.dectech.org/football/index.php These will be presented as follows:
We can see that Wolves are given a 24.9% chance of beating Blackburn away for example while Arsenal are given a 72.7% chance of beating Fulham away.
Use the Value Tool.
Download our Fink Tank Value tool from here: https://smartbettingclub.com/FinkTankValue.xls
The tool automatically works out the value for you, but at the moment you still need to manually enter the odds and the predictions. The odds can be copied and pasted from oddschecker, but for ease of explanation, we’re going to just type these in.
The first two columns are optional (Date and Game), but useful for keeping your records.
Next you enter the home odds, draw odds and away odds available. For speed, we’ve been entering the average odds from oddsportal, but there are many different methods you can use here.
FT H, FT D and FT A refer to Fink tank’s predicted probabilities for Home, Draw and Away respectively.
We have entered the information for the Birmingham Spurs game here:
Now the value tool will spring to life and tell you if there is a bet here.
The % Home, % Draw and % Away display implied probability from the bookmaker’s odds. This is automatically generated, allowing you to compare Fink Tank’s predictions with the bookmaker odds. If Fink Tank’s predicted probability (on the left) is greater than the implied probability.
Fink Tank (FT) Away win probabilities are slightly bigger than the probability implied by the bookmakers, giving us a value bet. Fink Tank make Spurs a 46.6% shot, while the average bookie odds are implying a 46% shot. There’s not much in that, but it seems to be a better bet than backing Birmingham or the Draw.
The next column will automatically display this for you:
It will turn green with a number 1 if that match outcome represents value.
If you do have a bet, the next step is to input the actual best odds available. According to Oddsportal, you can actually get 2.20 in many places.
You now need to enter the next few columns manually.
Here we’ve entered the pick, the price, the bookie. The finals columns are automatically generated. They show Fink Tank’s predicted value probability and the value on offer.
We can see that backing Spurs at 2.2 is value, but only 3.70% value even taking the best odds on offer. Considering that Birmingham are home under dogs, you might find enough reasons to skip this bet.
If you wanted to scan the whole league, we find it easier to enter all the odds & probabilities in one batch then assess the bets. We’ve done this for the forthcoming fixtures below:
There is some big value on offer in some of these games and we believe this is important keeping in mind the dominance of the elite and non elite top half clubs in the Premier league.
For example, Chelsea taking on Everton is an Elite vs bottom half match which hugely favours Chelsea. Indeed Fink tank give Chelsea just a 15.9% chance of victory, but by taking the best price of 10.98 you are getting good value. If Everton were less than 10% value then you might wish to think again.
Value betting is about dealing with probabilities. No team has a 100% probability of winning, what you are trying to ensure is that you are adequately rewarded when the unexpected does happen.
There can be some good value at relatively low prices though, such as Wigan at 2.72 which offers nearly 14% value according to Fink Tank.
The Fink Tank ratings have provided a good account of themselves so far this season, picking up some huge value bets along the way. No doubt there has been some luck in converting value into actual profits, but there’s no denying that using Fink Tank would have helped you get on the right side of those upsets. It has certainly helped that this is turning out to be one of the most open Premier Leagues in recent years, so that must be a consideration. We unfortunately have no data for Fink Tank’s predictions for other leagues, but imagine their accuracy would be in line with what we have been able to test.
At the very least, we believe that Fink Tank’s ratings provide a useful filter from which to refine your bets further. For example, you might want to take into account the home underdog effect and be wary of opposing elite teams unless there is clear value. Derby games might also be avoided unless there is clear value. As the ratings take no account of form, they could be used well in conjunction with a stats tool such as Form Lab Pro or the excellent free http://simplesoccerstats.com/
According to our private testing, Fink Tank ratings are not as well calibrated as commercial offerings such as Winabobatoo, but considering that Fink Tank’s ratings are freely available, we have heartily recommend members involve them as part of their betting arsenal.