Concepts behind Sports Betting Strategies

Homepage
How to Win in Over/Under Markets?
Tracking Market Movements in Sports Betting
Why Betting Accounts Get Closed or Restricted?
Are Betting Tipsters Any Good?

Such binomial distribution is actually ideal for 50-50 propositions such as the Asian handicap or point spread markets, wherein the odds of each side winning are almost even money, or slightly shorter once the bookie factors in his betting margin. However, you will see sports bettors betting on all kinds of different odds with all levels of stake values, for instance, match betting in tennis or 1x2 markets in football.

You can depend on something referred to as the T distribution in these circumstances. And a student’s T test for the statistical significance. We’ll explain how you can use T distribution for gauging the actual performance of a sports betting tipster.

A T distribution can be likened to a bell-shaped regular distribution. In fact, for all intents and purposes, it’s actually the same thing when the number of sports bets are over 30. The T test carries out an investigation into the chances of profit from a series of bets that could have been purely a matter of chance.

Please keep in mind that the smaller the chances of something happening, the more likely that the profitability of a sports bettor may be due to something else, for instance, his betting skills. What T test does is that it simply compares the observed returns of a sports bettor with that of a theoretical expectation (assuming it to be chance related) that’s defined by the specific market he/she is betting on.

Normally, it would be a loss equal to the margin of the bookmaker, or break even in case the sports bettor takes the trouble of finding the best prices with the help of some odds comparison tool. Thereafter, the resultant T score is analysed for determining if the difference is actually statistically significant or not.

It’s intuitively obvious that the higher the profitability is, the higher would be the T score, and the betting history would be more statistically significant as well. Putting it in other words, it would mean that the higher would be the chances that skill has contributed to the sports bettor’s winnings. This T score is generally directly proportional to the excess average return scored by a sports bettor, over and above the expectations.

In the similar manner, the longer is the history of an equal return, the higher will be the chances that it’s due to something other than the matter of chance. For instance, if two sports bettors score 120% returns on their investments, with the first one achieving it with 10 bets, and the second one with 1000, whom do you think is more likely to be a skilled sports bettor?

Whenever you’re in doubt, you must consider tossing the coins once again. Getting 6 or more tails from 10 coin tosses is far more likely to happen than getting 600 or more tails from 1000 coin tosses, assuming it’s only chance in play. If at all you get more than 600 tails, you would have all the reasons to suspect that the coin is biased in some way.

Similarly, it’s much easier to conclude that a sports bettor having a fairly long profitability record, has enjoyed so much success due to his skills than due to chance. Mathematically speaking, the T score is actually proportional to the square root of the total number of bets.

Betting odds have a comparatively less intuitive influence. Achieving 120% ROI from betting odds of around 1.25 is a much higher indicator of skill compared to the equal amount of profit scored from betting odds in the range of 5.00. Betting on sports outcomes with lower probability (longer odds) is quite risky inherently (assuming that the stake amounts are the same), as the result of such bets is highly dependent on random variability.

Putting it another way, the returns have a higher volatility. 21 or 19 winners scored at 5.0 odds will deliver returns of 105% or 95% respectively. Compared to that, 81 or 79 winners at 1.25 odds will return a 101.25% or 98.75% profit over the turnover respectively. Whenever you bet at longer odds, you assume more amount of risk to get higher rewards.

We can better understand the impact of betting odds with the help of standard deviation in the profits and losses over the betting history. You can approximate the standard deviation for the levels staking, using the following expression:

S = square root of (r (o – r))

Here, ‘r’ is the actual return scored by the sports bettor, and ‘o’ means the average odds over the betting history. The standard deviation witnessed in profits and losses from bets placed at 5.00 odds or higher, is over 8 times higher compared to when bets are placed at 1.25 odds. Presuming that the expected returns (based purely on luck) are breakeven or 100%, the T score can be figured out using the equation below:

T = square root of n (r – 1) / s

Here ‘n’ represents the total number of bets.

As a result, the T score in case of equivalent returns over the same betting history, will be over 8 times smaller when bets are placed at 5.00 odds than when they’re placed at 1.25.

It is clearly evident that higher yields achieved when you bet at longer odds, as can be typically seen in markets such as horse racing, aren’t necessarily an indicator of having better forecasting ability. Having the same amount of luck will deliver you a higher percentage of returns.

Therefore, it’s fundamentally misleading to make comparisons between betting histories which factor in only the percentage of returns, something which is commonly seen in case of ranking tipsters. The T score actually provides a measure of quality of risk-adjusted returns (in excess of the expectations) when you take the betting odds into consideration.

The last step is conversion of T score into probability, or ‘p’ value, considering that a profit history can only arise due to matter of chance. People using Microsoft Excel can employ the TDIST function in this regard:

TDIST (t, degrees of freedom, tails)

Wherein,

t = T score,

degrees of freedom = number of independent data pieces = (number of bets – 1)

tails = in this argument can be 1 (in one-tailed T test) or 2 (in two-tailed T test)

As our primary objective is to figure out if the profit is actually statistically significant or not, we’d go with the one-tailed T test. As an alternative, you can also look up some online calculator for inputting these values.

The table provided below gives examples of some T scores as well as the corresponding p-values, over betting histories of 100 bets, and ROI of 120%.

50 / 0.26 / 39.72%

25 / 0.37 / 35.45%

10 / 0.62 / 26.98%

5 / 0.94 / 17.56%

4 / 1.09 / 13.89%

3 / 1.36 / 8.83%

2.5 / 1.60 / 5.63%

2 / 2.04 / 2.19%

1.75 / 2.46 / 0.78%

1.5 / 3.33 / 0.06%

As is evident, whether the profitability of a sports bettor will be viewed as something to do with skill or merely a matter of chance, is greatly impacted by the average odds at which that bettor places his bets. A 120% return achieved from 100 bets at 10.00 odds or higher, is clearly an indicator of it being due to luck.

On the other hand, if a sports bettor achieves the same kind of returns when betting at odds-on prices, his profitability is more likely to be due to his skills.

Resultantly, whenever you’re comparing different betting histories, for instance in case of tipsters, you can’t just do with only an analysis of their percentage returns. You’ll also need to factor in the odds which were in play, as well as the total length of their records.