Grex's SAM System - 124 bets -1% yield

[Grex]

I have been around on this Forum for a long time – and have published various systems with varying degrees of success 

Previously, I have usually tried to develop my own statistical system for predicting the results of matches. However, this system is different in that it “piggy-backs” on the predictions of SAM (Sports Analytics Machine) – developed by Ian McHale, a professor at the University of Liverpool. The predictions apply only to the English Premier League. As SAM’s predictions are published on the BBC website, I hope that I am not infringing any copyright rules by using them.

Basically, my system takes SAM’s predictions for 1-X-2 and compares the projected probability with published odds on Bet365 in order to calculate odds “value”. Then, by applying cut-off limits to the “value” and to the odds themselves, the system selects games on which to bet 1-X-2.

SAM also predicts the most likely score-lines, and I have included these in the system as a test.

I have followed the system for a very limited number of games in the EPL this season – there are gaps in the coverage, when I failed to get access to SAM’s predictions. The performance is good, but obviously with limited data.

The table below shows the results to date, and the selections for Saturday’s games. I stress that this is still very much in the testing phase, and I recommend that nobody risks their hard cash until the system demonstrates profit (if any) over a longer period.

I will update the table after Saturday’s games, and include selections for the games to be played on Sunday and Monday (SAM’s predictions are not yet available).

Wish me luck 

Edited April 30, 2018 by Grex

[harry_rag]

Good luck. Have been struck by a sense of deja vu, have you posted a similar system before, or proposed doing it? 

[Grex]

Thanks, Harry. I recently mentioned the possibility of starting a new system in my old "HOMES" thread, but I have never used SAM's predictions before.

By the way, I apologise for the poor definition of the graphic in my first post - I will have to work on this - possibly, using another image hosting site.

 

 

Edited November 17, 2017 by Grex

[harry_rag]

Cheers, I was thinking of a thread started by the Curlyone1 called "Sam Says" in Glory Hunts last year. Nice to know I'm not imagining things!

[Grex]

It's only people of my age that imagine things - it's called senile dementia 

One clarification..... in the red "BET" column, "NB" = No Bet - so, there are only 5 selections for Saturday

[Grex]

A good start for the 1-X-2 selections - 52% ROI. A marginal profit for the "correct score" system - only one winner out of eight. I suppose this is to be expected with odds of 5.5 to 8.5. It will be a roller-coaster ride - but let's see how it goes.

Summary of results in the table below, plus a selection for Sunday. Unfortunately, SAM's predictions for Monday's game are not published yet, so I will post again on Monday morning.

 

Edited November 18, 2017 by Grex
English !!

[jtw1]

Nice start Grex , expected nothing less !!

[Grex]

It doesn't look as if BBC is going to publish the SAM predictions for the Monday game 

So, I post the update including Sunday's Watford v West Ham game - the correct result, but not the correct score. I have adjusted the font - so, I hope it is easier to read.

The ROI on 1-X-2 system moves up slightly to 62% - but still VERY early days.

[Grex]

Selection for the weekend, including West Ham tonight 

[Grex]

Not much joy with the 1-X-2 selections - into negative territory - overall ROI -3% 

Better for the correct score selections, with 3 out of 10 correct - overall ROI +58% 

Let's see if BBC publish the SAM predictions for the mid-week round of games.

[Grex]

It looks as though BBC are NOT going to publish SAM's predictions for the mid-week games  so, we will have to wait until the weekend for the next "exciting" installment.

[Grex]

Selections for the weekend..... including United to beat Arsenal  (unlikely)

[Grex]

On 01/12/2017 at 3:44 PM, Grex said:

Selections for the weekend..... including United to beat Arsenal  (unlikely)

Oh, ye of little faith 

[Grex]

A good weekend.......

1-X-2 selections into positive territory again - 18 bets - overall ROI  +10%

Correct score selections - 29 bets - overall ROI  +76% 

[Grex]

Selections for Saturday - a lot of home wins - let's hope they oblige 

I will post Sunday's selections tomorrow morning based on up-to-date odds.

[Grex]

Status quo maintained for the 1-X-2 selections (ROI +10%)

No winners for the correct scores 

Sunday's selections below..... personally, I hope the last one fails 

[vikki37]

6 hours ago, Grex said:

Status quo maintained for the 1-X-2 selections (ROI +10%)

No winners for the correct scores 

Sunday's selections below..... personally, I hope the last one fails 

Called off I like your hope. Same here. 

[Grex]

The status quo for 1-X-2 continues - 25 bets - ROI +10%...... unfortunately, the last selection won 

Only one winner for the correct scores - 39 bets - ROI +49%

I don't think BBC will publish the SAM predictions for the mid-week games - so, until next weekend 

 

[thecurlyone1]

Sam has made predictions for tonight 

[justanotherpunter]

Can you provide a link to where the predictions are published on the BBC website - I seem to be wearing blinkers when hunting on the BBC site for this.

Thanking you in advance if this is available

[Grex]

9 hours ago, thecurlyone1 said:

Sam has made predictions for tonight 

You were right, Curly One - but, unfortunately, I was travelling and missed them 

Hi, Justanotherpunter - you can find the predictions at the bottom of each analysis page on BBC Sport, e.g. for tomorrow's Newcastle v Everton game, the link is http://www.bbc.com/sport/football/42247118 

1-X-2 selections for Wednesday evening are few - but they do include the first "X" selection. 

[Grex]

Arsenal's draw rescued what would, otherwise, have been a disappointing evening.

1-X-2...... 28 bets - ROI +14%

Correct scores...... 46 bets - ROI +39%

[yesitsme]

Are people tending to put bets on just 1x2, just correct score, or both? And level stakes, or vary it based on the odds? 

Thanks 

[Grex]

Haven't got a clue what people are putting their money on (if any).

I did say in my first post "I recommend that nobody risks their hard cash until the system demonstrates profit (if any) over a longer period"......... and this still applies. It is still VERY early days.

 

[yesitsme]

Understood. 

I'm toying with it, albeit tiny stakes (1 or 2£)

[Grex]

1 hour ago, yesitsme said:

Understood. 

I'm toying with it, albeit tiny stakes (1 or 2£)

You have been warned 

But, if you are going to go for it, I suggest you bet on the 1-X-2. Although SAM uses unknown algorithms to predict "correct scores", they are still an absolute lottery - OK, it's providing a good ROI at the moment, but with odds of 7.0 - 8.0, there could be very L - O - N - G losing streaks 

[Torque]

Sound advice Grex. Very roughly, the correct scores could conceivably see 40 or so losers in a row.

[Grex]

Thanks, Torque.

A losing streak of 40  would kill most bettors.

Presumably this figure is based on random betting. Perhaps SAM's more informed analysis will reduce the risk of having such long losing streaks - but it should still be avoided. 

[yesitsme]

1 hour ago, Grex said:

You have been warned 

But, if you are going to go for it, I suggest you bet on the 1-X-2. Although SAM uses unknown algorithms to predict "correct scores", they are still an absolute lottery - OK, it's providing a good ROI at the moment, but with odds of 7.0 - 8.0, there could be very L - O - N - G losing streaks 

Cheers - don't worry, I won't blame you if I lose money! 

I agree. The correct score looked tempting at first, but hasn't won in several bets. The 1x2 ought to be more consistent. 

[Torque]

4 hours ago, Grex said:

Thanks, Torque.

A losing streak of 40  would kill most bettors.

Presumably this figure is based on random betting. Perhaps SAM's more informed analysis will reduce the risk of having such long losing streaks - but it should still be avoided. 

As I understand it, it's based on any set of bets with around a 15% probability of success which is roughly in the middle of the probabilities of correct scores priced at 6 to 7. The fact you are only betting when there is perceived value in a correct score price relative to SAM's calculations might help to bring down any losing streaks a little, but as an example even if you're betting correct scores priced between 6 and 7 when they should be 5 you will still likely encounter a run of 30 losers or so. From that perspective, it's probably best to be cautious and factor in a losing run of 40 when deciding what bank percentage should be invested in each correct score bet. For correct scores priced 7 to 8, a losing streak of about 50 could occur and again that would need to be taken into account so that the right stake is used.

Edited December 14, 2017 by Torque