Soccer h2h analysis

Since going live on the 14/9/09, sportpunter’s soccer predictions have gone very well indeed. Head to head betting has been the main avenue of success with 7.2% ROI made on 3451 bets. Totals betting has pretty much broken even in the same time frame, whilst asian handicap betting has lost marginally.

So why not do some analysis into the areas with different combinations on certain variables. This might well help us make more educated bets, not bet on certain matches given certain circumstances and also help me play with the model to produce better results.

So first up in this 3 part series, is the analysis of head to head betting. Whilst not as in depth as totals and asian will be, some will say it doesn’t need to be given the great results. The summary is given below:

	Prob	#Bets	#Won	%Won	$Bet	$Profit	%ROI
0	0.1	34	1	3%	$623.66	-$500.41	-80%
0.1	0.2	453	52	11%	$16,805.12	-$3,356.96	-20%
0.2	0.3	838	165	20%	$50,147.00	$1,906.43	4%
0.3	0.4	756	248	33%	$59,659.08	$10,980.07	18%
0.4	0.5	628	242	39%	$63,689.76	$224.78	0%
0.5	0.6	415	185	45%	$59,765.49	$327.38	1%
0.6	0.7	191	109	57%	$39,813.69	$7,634.84	19%
0.7	0.8	81	66	81%	$20,297.35	$6,001.79	30%
0.8	0.9	46	35	76%	$16,103.92	$526.60	3%
0.9	1	9	8	89%	$4,595.03	$88.81	2%
	TOTAL	3451	1111	32%	$331,500.10	$23,833.33	7%
	Odds	#Bets	#Won	%Won	$Bet	$Profit	%ROI
1	2	380	236	62%	$72,990.08	$5,278.45	7%
2	2.5	522	229	44%	$66,091.76	$2,143.26	3%
2.5	3	492	190	39%	$50,778.77	$5,299.75	10%
3	3.5	348	126	36%	$30,687.02	$5,695.46	19%
3.5	4	319	99	31%	$22,340.90	$3,684.09	16%
4	4.5	255	59	23%	$19,230.48	$919.02	5%
4.5	5	227	36	16%	$16,484.11	-$2,267.86	-14%
5	6	271	53	20%	$17,456.93	-$164.70	-1%
6	10	561	80	14%	$32,747.12	$4,706.04	14%
10	100	76	3	4%	$2,692.93	-$1,460.18	-54%
	TOTAL	3451	1111	32%	$331,500.10	$23,833.33	7%
	Overlay	#Bets	#Won	%Won	$Bet	$Profit	%ROI
0	0.075	511	185	36%	$27,793.94	-$710.45	-3%
0.075	0.1	425	151	36%	$29,669.53	-$1,034.00	-3%
0.1	0.15	597	223	37%	$44,953.42	$4,472.91	10%
0.15	0.2	513	190	37%	$49,199.68	$5,591.83	11%
0.2	0.3	609	180	30%	$68,067.43	$3,726.22	5%
0.3	0.4	324	84	26%	$42,444.75	$966.18	2%
0.4	0.5	169	37	22%	$21,878.11	$1,045.73	5%
0.5	0.75	220	44	20%	$31,451.66	$2,880.24	9%
0.75	1	63	11	17%	$10,129.65	$1,540.50	15%
1	100	20	6	30%	$5,911.93	$5,354.17	91%
	TOTAL	3451	1111	32%	$331,500.10	$23,833.33	7%

As shown above, profits were made in vitually all fields. Big outsiders with probability less than 20% yielded a loss, as too did odds of over 10, and this might be looked at closer. In particular would be interesting is how big outsiders faired when away and home. I would imagine that home based big outsiders (although this very rarely and probably never occurs), would give a positive result.

In fact betting at odds of 4.5 or more only resulted in a profit if 1.1% ROI. Still a profit, but it would seem the best results occur with odds lower than this amount and realistic opportunities for a win. Are outsiders bet more than favourites so that mug punters try to win big? The favourite/longshot bias might well be playing a significant factor here, in that the shorter the odds the more likely of a potential profit, irrespective of any of Sportpunter’s suggested bets. We’ll have a look at that in the future, but for now, it seems that betting on all ranges of odds and probabilities is successful, moreso on favourites.

Higher overlays resulted in greater profits which is great to see. A good model should clearly show this. The 91% ROI profit for 100%+ overlays is interesting, and on closer examination, there are a few daggers or maybe nuggets in a few bets. I note that following bets made with high overlays and strong probabilities:

Date	League	Team1	Team2	Odds	Prob	$Bet	Result	$Profit	Overlay	Check?	Error?
23/10/2009	wales1	Rhyl	Cefn Druids	7	82.2%	792.16	0	-792.16	475.3%	X	X
7/11/2009	wales1	Port Talbot Town	Caersws	4.5	67.3%	579.28	1	2027.5	202.7%	X	X
26/10/2009	greece2	Ethnikos Asteras FC	PAE AO Kerkyra	5.25	56.1%	457.64	0	-457.64	194.5%	X
22/11/2009	ecuador1	Manta FC	Liga de Quito	3.5	76.4%	670.14	1	1675.36	167.5%	X	?
29/11/2009	russia1	Rubin Kazan	Kuban Krasnodar	3.2	65.7%	501.09	3	1102.4	110.2%	X

Clearly the two wales bets have the odds around the wrong way, and hence these would not have been normal bets. The Greek bet seems legit, the Russia bet is possibly legit, whilst I cannot find odds on the Ecuador bet either confirming or denying its legitimacy.

These will come up from time to time in the program, so its important to look out for them. Either way, in 3500 bets, only 5 bets are suspect which is good. The profit for this according to the betting history would decrease the profit by $1235 only. So we still have great results for higher overlays.

Interestingly overlays less than 10% lost around 3% ROI. There might be a case for having a minimum overlay of 10% or probably better 7.5% considering the large profits made in the next section 10-20% overlays.

Either way, these are great results for head to head betting, and we will look closer at totals betting in the next post.

To sign up to the soccer model, please click here.