Inflection points are the points at which the curvature or concavity on a curve changes from plus to minus or, from minus to plus. *Translated into layman’s language* – the turning points on the profit/loss graph where profits turn to losses or, where losses turn to profits.

### Inflection Points in Football Betting

If you have ever **calculated your own odds** you will certainly have noticed that bookmaker prices often do not show the ‘true’ picture.

In other words, their odds seldom adhere to ‘true’ mathematically calculated values *(the statistically expected values)*.

In the majority of games, odds are either higher than mathematically expected or lower…

**Why Is This So?**

A **bookmaker’s aim is to make a profit** and they price their odds to ensure that sufficient action is taking place on both sides of a bet, with enough profit retained whatever the outcome.

In addition, bookmaker betting odds are **often aligned to public opinion** to guard against a disproportionately large amount of money being placed on just one side of a bet.

Thinking this through with a mathematical mind, it seems very plausible that there *must be* certain odds clusters which are regularly under-priced, whilst other odds clusters are habitually over-priced.

The evidence to support this ‘theory’ can be easily spotted in the **HDA Simulation Tables** using visualisations *(see example below)* in the form of Profit/Loss curves based on the data within these tables.

*For example, the EPL:* Between 2009-2014, if you had gambled unemotionally and systematically on *all* English Premier League matches to be home wins *(at average bookmaker odds)* and placed a constant stake of 100 units per fixture, then at the end of the fifth season, your losses would have accumulated to **-3,157** units.

Even higher losses would have been incurred backing all home teams to win if you would have limited your betting to the zone of odds between **1.49** and **6.53**. Within this zone your losses would have totalled **-7,711 units** *(198 plus 7,513)*.

However, if your strategy had been based on home wins at odds below **1.49** and above **6.53**, after five seasons your profit would have been **4,554** units *(7,513 minus 3,157 plus 198)*.

### Inflection Points for Home Wins for the EPL

The above screenshot shows the Profit/Loss curve for the EPL if back bets *(at average bookmaker odds)* were placed on all 1,900 matches between 15/08/2009 and 11/05/2014 *(at 100 unit stakes)*.

You can see from the graph that the first real turning point is at home odds of **1.49**. The P/L curve reaches a peak of 198 units profit before starting to fall. The erosion of profits continues until home odds of **3.49** are reached (P/L value: -7,084 units), where the curve turns for a spell before dropping again and reaching its final inflection point at odds of **6.53**.

This type of graph is of great help in visualising profits and losses and in this case, the curve shows that in this odds group *(60% of the matches)*, bookmakers under-price home team wins. Bookmakers make a long-term profit, whilst bettors lose more than they win.

The interpretation is that ‘home win’ bets must be quite popular on EPL matches at odds between **1.5** and **3.5**, and bookmakers react to this high demand by reducing prices. Of course, this also means that the odds on other sides of the bet *(draw and/or away win)* must then be increased.

### Working Out Inflection Points using the HDA Tables

Here’s a handy little tutorial:

### How to Use the Knowledge of Inflection Points

#### (1) check your own betting pattern

Do you recognise your own betting patterns within the most common odds clusters; those which show a falling Profit/Loss curve? *(For example, backing the home team to win at odds between 1.50 and 3.50 in the EPL).*

If so, perhaps reconsider your strategy and avoid the tranches of odds used by the bookmakers to make their profits. Of course, no matter how hard you try, betting in the zones where bookmakers habitually reduce prices (odds) is asking for long-term losses.

It is very rare that people succeed in any walk of life by swimming against a strong current and you can safely assume that the bookmakers know their job and have made a living from manipulating figures for centuries.

#### (2) don’t even try to “beat” the bookies

Swim along with them. ‘Play’ the market the same way as they do. Start looking at strategies which are not in conflict with the market, but in rhythm with it.

#### (3) be aware that each league has different inflection points

Customs and habits of people vary from country to country. Every nation has different cultural settings, and everybody has certainly noticed regional differences expressed in tangible goods such as food or housing.

Unfortunately, differences in betting patterns and the subsequent reaction of bookmakers when setting their odds cannot be spotted without some mathematical calculations. Customer habits, especially in the betting market, remain well hidden from the bettor.

Whichever league you prefer betting on, identify the odds clusters which are utilised by the bookmakers to make their profits – and then work around them.

#### (4) use the market inflection points for your own benefit

Concentrate on developing your personal betting strategy by taking the market rules into consideration.

For example, in the EPL, you may wish to look at a ‘back the home team’ strategy which embraces home odds under **1.50** or, alternatively, over **3.50**, or even better, over **6.50**.

#### (5) find a strategy and identify matches for producing long-term profits

Using the **HDA simulation tables** and finding the various inflection points will provide you with knowledge of odds clusters which produce a long-term profit or loss when backing.

If you already use our **Value Bet Detector** for calculating odds for individual matches then the knowledge of profitable odds clusters will help you to pick matches which are worthwhile re-calculating.

Hello Soccerwidow!

First of all, this is a very useful site and congrat for the tables!!!

I tried to understand the calculations, but I have a problem.

In the “inflection point” article you wrote that if we had gambled unemotionally and systematically on all English Premier League matches to be home wins (at average bookmaker odds) and placed a constant stake of 100 units per fixture, then at the end of the fifth season (2009-14), our losses would have accumulated to -3,157 units.

In the other article, which shows some screenshots from the HDAFU tables, there is a screenshot which shows the “Backing home team to win” (by team at home) in the English Premier League from 2009 to 2014 seasons, the total profit is +7395 units.

I don’t understand the difference of the total unit numbers.

Thanks in advance for your answer!

Regards,

Falconer

Hi Falconer,

Many thanks for your query.

Yes, the original article was based on average odds but the HDAFU tables are always based on the highest bookmaker odds.

Right now we are converting all of the HDAFU tables to highest odds at point of kick-off, which is the basis for the

current batch of summer league simulations.The winter leagues will also be based on highest odds at the end of the ante post market when they are reissued in May/June, when the major European leagues conclude.

Hope this helps.

Hi, interesting article.

I have a question about this in combination with value betting.

Once we have defined inflection points for home wins in a league, is it not a sound strategy to only bet on home games in a profitable oddscluster, on the condition that the odds have value as well?

So, basically two conditions:

1. Odds have value

2. Odds fall within a historically profitable oddscluster

Or do the inflection points solely rely on the complete dataset? I.e. all home wins?

I guess this is (maybe not so advanced) maths, I am really curious about your opinion on the matter.

Hi BM,

The inflection points identified in this article are based around five full seasons of data.

If each of the five seasons produces profit in the chosen inflection point cluster then ‘value’ is inherent in that particular group of matches.

With system betting such as this, we look for the one cluster group in each league with the largest profitability and if it shows profit in all five seasons then it is paper tested for inclusion in the portfolio.

I hope this answers your question and thanks for taking the trouble to write.

Is it not possible we may not find any profitable infliction points in the table eg A league over 5 years? Is it possible any profitability has been created by chance and may not continue?

Yes, it is possible that some leagues do not show clear inflection points. How bookmakers structure their odds (prices of bets) depends on the market (demand from the bettors). This varies from country to country as well as from league to league.

To your second question… no, the curves have not been created by chance.

Hi,

While its true that this strategy would have won money in the years you mention. It would have lost money on the Premier League 2014/15. There was only one home win with odds above 6.50

Hello Jay,

Yes, thanks for pointing this out. It allows me to remind everyone that no matter what your homework reveals about a forthcoming season, everything must be recalibrated at the end of that season. Its statistics become the first year back and the old fifth year back is discarded from the analysis, so that you only ever calculate using the last five seasons’ data.

This also hammers home the point that some systems will fail in some of the leagues you choose. But if you have a large enough portfolio of potential bets in a season (we are talking at least 500 but preferably more), then the law of large numbers will help the accuracy of your statistical analyses.

Losing systems should then be supported by winning ones, with enough profitable systems to guarantee an overall profit. In the end, it becomes just a numbers game where accuracy of the historical data used in your analyses is paramount to success.

Hi

So lets say im currently in the 14-15 Season having the Profit/loss Graph of the previous 5 seasons. Would your really choose the below 1.49 odds since the curve is pretty much jumping around 0. wouldnt the 3.49- ~4 be better?

Apart from that, wouldnt i be able to choose the Matches within the inflection points and not bet on the 1×2 but the corresponding AHC around the 50% Probability. Instead of Picking the 7.50 away win against ManU id take the +1.25 AHC or instead of the 1.20 Home win for Chelsea id take the -1.5 AHC. Maybe variance could be kept low this was?

Do. you think there is a possibilty this whole inflection Point topic happens by Chance? I have a variance Simulator and Even After 10.000 simulations/Matches with a 5% edge over the real outcome, variance could still kill your profit, and with a -5% edge variance could make you believe you are a good Player when clearly you are not

Hello Kai,

This article is one of our early inflection points pieces. Things have moved on since, and I would recommend reading the latest articles such as:

2016-17 Winter League CampaignIn response to your final point, although the inflection points change year in, year out, it is often the same bet type which dominates a particular league, sometimes for many seasons in a row. I don’t think it’s chance that home wins are dominant in the American MLS, or away wins in the Japanese J-League, both of which have been the case for several consecutive seasons now.

Consistent trends such as these are also present in almost every league we have looked at.

By filtering the data set you should end up with a targetted set of games, carrying odds usually well in excess of the zero odds for that bet type in that league. In our User Guide article, you will see that zero odds for the chosen bet type reduced from 3.804 for the league in general to just 3.006 for the filtered set. The inflection point odds there were between 3.33 and 3.63.

The benefit received from odds of 3.006 at a stake of one unit is 2.01 (rounded up); from 3.33 it is 2.33; from 3.63 it is 2.63. The inflection points odds therefore range from 15.92% to 30.85% above the benefit of the zero odds, and there is only a small chance that a hit rate variance in excess of 30% within our chosen odds range will strike the system down.

Five seasons’ statistics are enough for a single league. In the EPL, for example, this equates to 1,900 individual matches. Here, we are dealing with recent history, which is more relevant to the season being played out at the time. 10,000 matches would represent more than 26 seasons, which is far too unwieldy a list when looking to forecast results that are going to happen tomorrow.

The data set is by far the most important thing to get right when analysing past performance in a league.

I hope this is of some small assistance.