Anytime Goalscorer System

[harry_rag]

4 out of 7 yesterday (Lukaku, Vardy, De Bruyne and Sterling)

119 winners from 361 bets (32.96%) for a profit of 20.72 points.

[Torque]

45 minutes ago, harry_rag said:

4 out of 7 yesterday (Lukaku, Vardy, De Bruyne and Sterling)

119 winners from 361 bets (32.96%) for a profit of 20.72 points.

Wonder what the odds were for De Bruyne scoring three. I'm surprised you don't look at braces and hat-tricks - more losers obviously, but when the wins come they can be sweet.

[Torque]

Would be interesting to look back over your scorers so far to see how many got a second or third goal.

[harry_rag]

43 minutes ago, Torque said:

Would be interesting to look back over your scorers so far to see how many got a second or third goal.

And, in a similar vein, how many scored the first goal, but I don't record that at the moment. In terms of numbers of goals scored (from my overall data sample, matched or not) I can say the following:

From 1487 players, 562 scored one or more goals (37.97%). My actual bet strike rate lags that at 32.96%.

469 scored exactly 1 goal, 77 scored twice, 13 managed a hat trick and 3 have scored 4 or more goals.

I'll check this out later but my opinion (based on years of betting in these markets) is that you are likely to find value in the anytime price more often than you find it in the brace or hat trick markets. You probably get more "mug" money in the multi goal markets, i.e. the sort who just see a "big price" and equate that to value, allowing the bookies to apply a bigger margin.

[harry_rag]

59 minutes ago, harry_rag said:

From 1487 players, 562 scored one or more goals (37.97%). My actual bet strike rate lags that at 32.96%.

469 scored exactly 1 goal, 77 scored twice, 13 managed a hat trick and 3 have scored 4 or more goals.

My average fair odds are 2.67. If they were accurate then you'd have expected 37.45% of the players to have scored one or more, or 557 so we're certainly in the right ballpark on that basis. Taking poisson as a guide, the expectations for 2 or more or a hat trick look like this:

Brace: 8.12% or 120 versus actual number of 94

Hat trick: 1.22% or 18 versus actual number of 17

I'd say the figures for 2+ illustrate the expected shortfall while the hat trick figures are a small sample related blip. Let's compare that to a real example from today's games:

Harry Kane's fair odds to score would be 2.62 versus best bookie odds of 6/4 (not unusual for Kane to be hard to get at a value price; I struggle to get him at fair + 10% edge on the exchanges).

His fair odds for a brace would be just shy of 11/1 versus best odds of 10/1. The hat trick is 76/1 versus 66/1.

Those figures aren't as emphatic as I'd expected but the true odds as a % of the fair odds does decrease (95.45 for anytime, 91.7% for brace and 88.2% for hat trick). I'd say Sky Bet are unusually out of line here with the multi goal prices and the next best prices of 7/1 and 50/1 are more typical and make the point more.

I should probably have looked for a better example but the point is, it seems harder to find apparently value prices in the multi markets compared to anytime.

[harry_rag]

49 minutes ago, harry_rag said:

I'd say Sky Bet are unusually out of line here with the multi goal prices and the next best prices of 7/1 and 50/1 are more typical and make the point more.

Trimmed to 9s and 50s now!

[harry_rag]

10 points on Sorloth (Cadiz) at 23/10 Uni 

(Anything over 3.1 acceptable.)

[harry_rag]

5 minutes ago, harry_rag said:

10 points on Sorloth (Cadiz) at 23/10 Uni 

Sorry, he plays for Sociedad rather than Cadiz.

[harry_rag]

I wouldn’t want to bet against the Kane hat trick now! 

[harry_rag]

10 points on Maeda (Celtic) at 2.672

[harry_rag]

10 points on Salah at 3.35

[harry_rag]

2 in Italy, 10 points each on Simeone at 2.84 and  Deulofeu at 2.72

[harry_rag]

1 out of 5 since the last update (Sorloth)

120 winners from 366 bets (32.79%) for a profit of 3.72 points.

[harry_rag]

Matched for 10 points each on these so far in the 2:00 games.

De Bruyne at 3.3, Ings at 3.55, Vardy at 2.38 and Hwang Hee-Chan at 3.2

[harry_rag]

Vardy the only winner from yesterday's 4 selections.

121 winners from 370 bets (32.7%) for a loss of 12.76 points.

[harry_rag]

10 points on Cabral (Fiorentina) at 2.68

[harry_rag]

10 points on Watkins at 21/10 (Hills) and on Lukaku at 2.52 and Richarlison at 3.05

[harry_rag]

And 10 on Pulisic at 3.25

[harry_rag]

1 from 5 since the last update (Richarlison).

122 winners from 375 bets (32.53%) for a loss of 33.67 points.

[harry_rag]

10 points on Petagna (Napoli) at 13/5 PP

[harry_rag]

Liverpool	Wolves	Diaz		back	bf	10	3.05
Arsenal	Everton	Saka		back	bf	10	3.05
Arsenal	Everton	Martinelli		back	bf	10	3.1
Arsenal	Everton	Nketiah		back	bf	10	2.46
Brentford	Leeds	Toney		back	bf	6.3	2.62
Man C	A Villa	Jesus		back	bf	8.34	2.56
Man C	A Villa	Foden		back	bf	10	3.2
Man C	A Villa	De Bruyne		back	bf	10	2.92
Chelsea	Watford	Havertz		back	bf	10	2.3

[harry_rag]

3 from 10 since the last update (Nketiah, Martinelli and Havertz).

125 winners from 385 bets (32.47%) for a loss of 50.69 points.

[harry_rag]

10 points on Abraham at 3.25

[harry_rag]

No joy with the last bet so 125 winners from 386 bets (32.38%) for a loss of 60.69 points.

Overall sample is 584 scorers from 1544 (37.82%) and I'd be well in front if I could bridge even half of that gap (e.g. an additional 10 winning bets). The question I need to wrestle with is am I losing because I've been unlucky (strike rate of matched bets is lagging that of all selections) or because the market is clever enough to turn a theoretical profit (all selections at target odds showing just shy of 8% yield) into an actual loss.

Some studying to be done during close season. I might have to bite the bullet and combine my two data samples (e.g. add the actual bets to the overall data so I can compare where I do and don't get matched).

[harry_rag]

On 5/27/2022 at 3:49 PM, harry_rag said:

No joy with the last bet so 125 winners from 386 bets (32.38%) for a loss of 60.69 points.

That equates to an ROI of -1.62%. Had I been fully matched on every selection (25 bets were only part matched) I'd be 117.63 points down with an ROI of -3.05%. Even then, at average matched odds of just over 2/1, it would only take 4 more bets to have won for me to be level. So not a million miles away.

There are 11 players who I've been matched on more than 6 times, accounting for 87 bets (22.54% of all bets). Of those only 3 were profitable; Nkunku +65.13, Mahrez +59.28 and De Bruyne +20.20. The 3 worst returns were Firmino -42.36, Lacazette -23.62 and Foden -17.37. Total return from that top 11 is -3.29.

[harry_rag]

Last thought for now; bets placed with bookies are out-performing those place on the exchange. Frustrating for a system where I was hoping the main focus would be bets placed on BF.

28 winners from 72 bets showing profit of 109.4 points to 10 point level stakes, with an ROI of 15.17%

Most profitable bookie the most likely to limit stakes to less than a tenner! 

[harry_rag]

Just a few tweaks having done a post-season review. I'm going to need the strike rate to improve for this thread to be profitable but these changes should slightly reduce the number of bets I get matched on but improve the return from those bets.

1. A change to the spread "base" number I use. Say one firm goes 23-26 and the other 21-24, the market price would be 23-24. Originally I used the mid-point (23.5 in this instance), latterly I've been using the sell price (23). From now on I'll be using the average of the lowest sell price and the mid-point (21 and 23.5 giving 22.25). Reason being greater accuracy and it's a truly dynamic value (in that it will always change if either firm moves it's price whereas the sell price could stay the same).
2. For all bets where my target price is higher than the best bookies price (so I'll be trying to get matched on BF) the target price will have a minimum 10% edge over the bookie price. Currently true for most bets but there's a small subset where it ends up being less.
3. For all bets where the best bookies price is >2.5 the target price will be capped at a maximum 10% edge over the bookies price (currently it could be higher than that) - see posts on 6th July looking at AE.
4. Also for bets where I need to try and get matched on BF the target price will be adjusted to include commission so that I'm always getting my desired price after commission rather than before.
5. Where possible I'll have a look at the stats for the players I'm backing and record all bets as being A, B, C and U where A = stats make it a positive bet, B = neutral and C = negative with U = unclassified where I don't have time to check. I may ultimately have to accept that this isn't a profitable system if followed blindly so it will be worth seeing if my subjective opinion adds any value or not!

Edited July 6, 2022 by harry_rag
Point 3 added after further analysis

[Smiles Tennis]

I think putting some subjective opinion in will help as long as you are clued up with the players you're betting on. You might know that a certain player is not in form, or just back from injury etc, and people might be backing them regardless but if they're up against a tough opponent and need games to get back in form then could be one to swerve for eg. 

Interesting to follow. 

[harry_rag]

10 points on Muriqi (Kosovo) at 3.15 (B - stats neutral)

[harry_rag]

10 points on Kramaric (Croatia) at 3.35 (U)