PlopPlop System Rating History 2000/01 to 2008/09

PlopPlop System Ratings are created from a database of over 20,000 English Football matches. 9 full seasons of the Premiership down to Coca Cola League 2 and 4 seasons of the Blue Square Conference Premier. Utilising many of the variables described in the Data section a rating is generated for each of the teams involved in upcoming fixtures. It is the difference between these two ratings that we are interested in, a positive differential suggests a Home Win, a negative differential, an Away Win.
Obviously we cannot divulge exactly how the ratings are calculated but what we can reveal is that they are based on complex statistical modelling, complimented by elements of modelling score distributions given by the Poisson distribution.

PlopPlop System Rating requires that each team in an upcoming fixture has played 6 matches. This stipulation gives the complex statistical models enough data with which to make a judgement about each teams ability. Removing fixtures where either team has played less than 6 games reduces the data from 20,442 matches to 17,602.

For ease of comparison the rating is banded into 10 groups; each containing 10% of the matches in the database. Below are the minimum and maximum PlopPlop rating within these bands.

PlopPlop System
Rating (1-10)	Minimum	Maximum	# Bets
1	8.00	34.00	1,760
2	5.00	8.00	1,760
3	3.00	5.00	1,760
4	1.17	3.00	1,760
5	-0.67	1.17	1,761
6	-2.33	-0.67	1,760
7	-4.17	-2.33	1,760
8	-6.33	-4.17	1,760
9	-9.33	-6.33	1,760
10	-28.83	-9.33	1,761

The most important thing is to understand whether segregating the matches via the PlopPlop System Rating drives a profit or not. As mentioned in other areas, if profit exists we would expect it to be observed at the extreme of these ratings. Namely a rating of 1 and a rating of 10.

NB: All of the following financial information is settled to a £1 level stake bet. For example, if there are 100 bets, this equates to a stake of £100. Where a positive rating exists (0 and above) returns are settled to a £1 home win. A negative rating is settled to a £1 away win. Returns are settled at the average odds available for the match. This ensures the results are realistic and achieveable.

PlopPlop System
Rating (1-10)	# Bets	Strike Rate	Returns	Yield (%)	Avg Odds
1	1,760	64.8%	£76.49	4.3%	1.64
2	1,760	53.1%	-£98.14	-5.6%	1.82
3	1,760	49.8%	-£132.17	-7.5%	1.91
4	1,760	49.2%	-£90.65	-5.2%	1.98
5	1,761	45.8%	-£163.72	-9.3%	2.51
6	1,760	30.4%	-£75.38	-4.3%	3.35
7	1,760	32.2%	-£87.20	-5.0%	3.15
8	1,760	32.7%	-£135.02	-7.7%	2.98
9	1,760	35.0%	-£132.82	-7.5%	2.79
10	1,761	47.4%	£168.87	9.6%	2.43
Grand Total	17,602	44.0%	-£669.73	-3.8%	2.46

In line with expectation we observe profit at the extremes of the ratings, a yield of 4.3% where we anticipate a home win and 9.6% for an away win.

In order to home in on the matches where we observe profit, only a rating of 1 and 10 are retained.

PlopPlop System
Rating (1-10)	#Bets	Strike Rate	Returns	Yield (%)	Avg Odds
1	1,760	64.8%	£76.49	4.3%	1.64
10	1,761	47.4%	£168.87	9.6%	2.43
GrandTotal	3,521	56.1%	£245.36	7.0%	2.03

Making this change reduces the total number of matches to just over 3,500 and has us realising a yield of 7.0% overall at average bet odds of just above evens (2.00). The next step to improve the situation is to look at performance within certain odds ranges, first up is home win bets, rating = 1:

PlopPlop System
Rating (1-10)	Odds Band	# Bets	Strike Rate	Returns	Yield (%)	Avg Odds
1	0.00 - 1.25	97	83.5%	-£0.18	-0.2%	1.19
	1.25 - 1.50	499	73.1%	£4.73	0.9%	1.38
	1.50 - 1.75	616	60.4%	-£18.57	-3.0%	1.61
	1.75 - 2.00	319	63.3%	£52.56	16.5%	1.85
	2.00 - 2.25	159	50.3%	£10.54	6.6%	2.12
	2.25 - 2.50	56	55.4%	£16.39	29.3%	2.34
	2.50 - 2.75	5	60.0%	£2.83	56.7%	2.62
	2.75 - 3.00	8	62.5%	£6.18	77.2%	2.82
	3.00 - 3.50	1	100.0%	£2.01	201.4%	3.01
Grand Total		1,760	64.8%	£76.49	4.3%	1.64

It is a well known problem within football betting that betting at short odds, whilst providing regular winners, is incredibly hard to drive a significant profit from in the long term. The sensible step here is to remove matches where, at average odds, we cannot obtain odds for a home win better than or equal to 1.75.

PlopPlop System
Rating (1-10)	Odds Band	# Bets	Strike Rate	Returns	Yield (%)	Avg Odds
1	1.75 - 2.00	319	63.3%	£52.56	16.5%	1.85
	2.00 - 2.25	159	50.3%	£10.54	6.6%	2.12
	2.25 - 2.50	56	55.4%	£16.39	29.3%	2.34
	2.50 - 2.75	5	60.0%	2.83	56.7%	2.62
	2.75 - 3.00	8	62.5%	£6.18	77.2%	2.82
	3.00 - 3.50	1	100.0%	£2.01	201.4%	3.01
Grand Total		548	58.8%	£90.52	16.5%	2.00

We have removed two thirds of non-profiting bets for this rating. As these were largely loss making bets we can see that overall, across 548 bets, we have a yield of 16.5% at average odds of evens (2.00).

Now we move on to away win bets; rating = 10:

PlopPlop System
Rating (1-10)	Odds Band	# Bets	Strike Rate	Returns	Yield (%)	Avg Odds
10	0.00 - 1.25	2	100.0%	£0.37	18.3%	1.18
	1.25 - 1.50	71	78.9%	£7.65	10.8%	1.40
	1.50 - 1.75	151	69.5%	£19.06	12.6%	1.62
	1.75 - 2.00	186	47.8%	-£21.22	-11.4%	1.86
	2.00 - 2.25	299	47.5%	£4.02	1.3%	2.14
	2.25 - 2.50	357	44.3%	£16.21	4.5%	2.37
	2.50 - 2.75	303	47.2%	£72.26	23.8%	2.62
	2.75 - 3.00	158	32.3%	-£12.36	-7.8%	2.84
	3.00 - 3.50	130	41.5%	£44.33	34.1%	3.23
	3.50 - 4.00	58	34.5%	£17.22	29.7%	3.72
	4.00 - 5.00	39	28.2%	£9.21	23.6%	4.43
	5.00 - 6.00	5	40.0%	£5.78	115.6%	5.35
	6.00 - 7.00	1	0.0%	-£1.00	-100.0%	6.36
	8.00 - 9.00	1	100.0%	£7.33	733.3%	8.33
Grand Total		1,761	47.4%	£168.87	9.6%	2.43

Again we must adhere to the old adage that profits are hard to achieve at short odds. Despite some impressive yield figures in the very short odds ranges, by reducing the number of matches in a similar fashion to the home wins, more profit per £1 staked can be realised. Going one step further than home wins; lets look at the effect of removing all games with odds of less than evens (2.00).

PlopPlop System
Rating (1-10)	Odds Band	# Bets	Strike Rate	Returns	Yield (%)	Avg Odds
10	2.00 - 2.25	299	47.5%	£4.02	1.3%	2.14
	2.25 - 2.50	357	44.3%	£16.21	4.5%	2.37
	2.50 - 2.75	303	47.2%	£72.26	23.8%	2.62
	2.75 - 3.00	158	32.3%	-£12.36	-7.8%	2.84
	3.00 - 3.50	130	41.5%	£44.33	34.1%	3.23
	3.50 - 4.00	58	34.5%	£17.22	29.7%	3.72
	4.00 - 5.00	39	28.2%	£9.21	23.6%	4.43
	5.00 - 6.00	5	40.0%	£5.78	115.6%	5.35
	6.00 - 7.00	1	0.0%	-£1.00	-100.0%	6.36
	8.00 - 9.00	1	100.0%	£7.33	733.3%	8.33
Grand Total		1,351	43.1%	£163.01	12.1%	2.65

Unlike with home wins, this has not reduced the number of matches in the sample by much, 23% in this case. We do however see an increase in yield to 12.1% at average odds of 2.65. The Odds Band '2.00 - 2.25' drives very little profit for some considerable effort, an additional 299 bets. The decision therefore is to remove this marginal group and concentrate our efforts where we clearly have an advantage over the bookmaker. It would be improper to ignore the '2.75 - 3.00'. There is however no sense to removing this group. Why would a bet with a rating of 10 be a bad bet just because the odds are longer? Longer odds bets with a rating of 10 return handsome profits. The feeling here is that this is a slight statistical anomaly that over seasons to come will iron out and we will observe a trend similar to the home win bets where profits increase in line with the odds. Betting at longer odds levels the playing field somewhat where bookmakers over-round is concerned.

The final away win table; rating of 10 and odds >= 2.25:

PlopPlop System
Rating (1-10)	Odds Band	# Bets	Strike Rate	Returns	Yield (%)	Avg Odds
10	2.25 - 2.50	357	44.3%	£16.21	4.5%	2.37
	2.50 - 2.75	303	47.2%	£72.26	23.8%	2.62
	2.75 - 3.00	158	32.3%	-£12.36	-7.8%	2.84
	3.00 - 3.50	130	41.5%	£44.33	34.1%	3.23
	3.50 - 4.00	58	34.5%	£17.22	29.7%	3.72
	4.00 - 5.00	39	28.2%	£9.21	23.6%	4.43
	5.00 - 6.00	5	40.0%	£5.78	115.6%	5.35
	6.00 - 7.00	1	0.0%	-£1.00	-100.0%	6.36
	8.00 - 9.00	1	100.0%	£7.33	733.3%	8.33
Grand Total		1,052	41.8%	£158.99	15.1%	2.65

We settle on a yield of 15.1% across 1,052 bets at average odds of 2.65.

The following tables are a further dump to show the consistency and growing profits over the past 9 seasons as more and more bookmakers move online and competition forces the book closer and closer to 100%. It is also worth noting how when the proportion of home wins and away wins moves around season to season, as we are betting on both sides of the coin, one type of bet compensates for a reduction in profits on the other side.

To reiterate the conditions under which a bet is placed:
Home Win Bet: Rating = 1 (+8.00 or higher) and HW Odds >= 1.75
Away Win Bet: Rating = 10 (-9.33 or lower) and AW Odds >= 2.25

Season By Season:

Season	# Bets	Strike Rate	Returns	Yield (%)	Avg Odds
2000/01	159	46.5%	£23.36	14.7%	2.48
2001/02	137	40.9%	£1.15	0.8%	2.51
2002/03	136	43.4%	£7.55	5.6%	2.48
2003/04	203	45.3%	£24.46	12.0%	2.54
2004/05	180	46.7%	£18.98	10.5%	2.51
2005/06	176	48.9%	£33.14	18.8%	2.51
2006/07	214	52.3%	£61.10	28.6%	2.57
2007/08	190	49.5%	£34.21	18.0%	2.53
2008/09	205	51.2%	£45.55	22.2%	2.53
Grand Total	1,600	47.6%	£249.51	15.6%	2.52

Season By Season, split by Rating:

PlopPlop System
Rating (1-10)	Season	# Bets	Strike Rate	Returns	Yield (%)	Avg Odds
1	2000/01	42	57.1%	£6.16	14.7%	2.01
	2001/02	35	45.7%	-£4.64	-13.3%	1.92
	2002/03	40	47.5%	-£3.60	-9.0%	1.93
	2003/04	72	50.0%	£0.07	0.1%	2.03
	2004/05	65	66.2%	£20.74	31.9%	2.00
	2005/06	64	62.5%	£13.75	21.5%	1.97
	2006/07	89	56.2%	£12.70	14.3%	2.04
	2007/08	64	59.4%	£9.92	15.5%	1.97
	2008/09	77	72.7%	£35.43	46.0%	2.04
1 Total		548	58.8%	£90.52	16.5%	2.00
10	2000/01	117	42.7%	£17.20	14.7%	2.65
	2001/02	102	39.2%	£5.79	5.7%	2.71
	2002/03	96	41.7%	£11.16	11.6%	2.71
	2003/04	131	42.7%	£24.39	18.6%	2.83
	2004/05	115	35.7%	-£1.77	-1.5%	2.79
	2005/06	112	41.1%	£19.39	17.3%	2.82
	2006/07	125	49.6%	£48.41	38.7%	2.96
	2007/08	126	44.4%	£24.29	19.3%	2.81
	2008/09	128	38.3%	£10.12	7.9%	2.83
10 Total		1,052	41.8%	£158.99	15.1%	2.79
Grand Total		1,600	47.6%	£249.51	15.6%	2.52