Sports Day Trading

[NotQuiteRandom]

Great info, thank you. Another question if I may, do you understand why your nets work? I don't mean in terms of the parameters that you use but in terms of the models that drive the functions. Can you give some additional insight into the technicalities and the maths?

I am not an expert in nets by any means but I have observed examples where such a model has been shown to work for a period of time but the operator was unable to explain fully why.

Thank you for your patient answering of my questions. This is very interesting information.

NQR

Hello NQR,

In answer to you question, yes, entirely and completely.

If ever I try to add in my own expertise(?) and knowledge(?) returns are far less reliable. The temptation for going against the neural predictions is never far away ... and I do admit to occasionally falling foul of my own golden rule which is;
"DM, the stats understand the situation far better than you do, so don't meddle!"
Tonight for instance, my nets tell me to oppose both Man Utd and Arsenal on the spreads currently quoted. My head tells me that it must surely be a dangerous ploy going against Man Utd especially in their current mood ... But hey! what do I know ... bets have now been duly placed.

A couple of good recent examples; Last month my football trades were going through a torrid time and a midweek-weekend-midweek patch saw me lose 600 points. I examined the "evidence" and decided that the tail-end of the season was not representative, and I should stop my football trading for the season. Maybe the cowards way is not to check-out "what would have been" after that decision, but why bury your head in the sand? The following weekend-midweek-weekend trading would have recouped all of the 600 points, and made profit on top, had I not invented reasons to curtail my trading in this area. Bugger! Football trading has been resumed

Example two: On Saturday (15 April) my NNets suggested a sell of both distances and SPs at Kelso. However, the racecard included a two-runner, three-mile steeplechase ... cue for my head to takes over the situation ...
Common sense (?) tells me to leave this well alone - too risky.
The outcome almost goes without saying? The distance spread would have returned a 296+ profit and the SP a 48+ profit. Double bugger!

I suppose my expertise & knowledge (this time without the question marks) is in the area of making best use of neural networks, which is by no means a straightforward business. But I've been using them in this area for over 10 years now, and you would have though by now I'd be able to stick to my own rules.

I'm still working on it. In a darkened room I'm repeating to myself over and over
"DM, the stats understand the situation far better than you do, so don't meddle!"

DM

 

[Data Miner]

19 April 2007
+802.04 b/fwd

Today's racing;
87] SOLD Cheltenham SPs @ 56 (SX) 3pts

===== Previous trading results (from 18 April 2007) =====
+784.99 b/fwd
82] SOLD Beverley Favs @ 74 (SX) 3pts MU=85, [-33.00]
83] SOLD Newmarket Favs @ 80 (SP) 3pts MU=70, [+30.00]
84] SOLD Cheltenham SPs @ 45 (All) 3pts MU=24.3, [+62.10]
85] SOLD Blackburn supremacy (v Watford) @ 0.9 (SP) 25pts MU=2, [-35.75]
86] SOLD Liverpool supremacy (v Middsbro) @ 1.4 (SP) 25pts MU=2, [-25.75]

Day's return +16.60
+802.04 to c/fwd
(slight adjustments to this figure after consolidating my records with spread statements received)

 

[Data Miner]

20 April 2007
+844.34 b/fwd

Today's racing;
88] SOLD Newbury Favs @ 72 (SP,SX) 3pts
89a] SOLD Thirsk SPs @ 52 (SF) 1pt
89b] SOLD Thirsk SPs @ 51 (SX) 2pts
90] BOUGHT Thirsk Favs @ 65 (IG,SX) 2pts
(There's also two early evening meetings today, so may be back later when the quotes are published)

===== Previous trading results (from 19 April 2007) =====
+802.04 b/fwd

Today's racing;
87] SOLD Cheltenham SPs @ 56 (SX) 3pts MU=41.9, [+42.30]

Day's return +42.30
+844.34 to c/fwd

 

[Data Miner]

20 April 2007 supplementary

91] SOLD Hereford distances @ 35 (SX) 5pts
92] SOLD Lingfield SPs @ 51 (SP, IG) 3pts
93] SOLD Lingfield Favs @ 50 (SF) 3pts
... football ...
94] SOLD Derby supremacy @ 1.3goals (All) 25pts


Happy spreading

 

[Data Miner]

21 April 2007
+1011.61 b/fwd

I won't be around to get the quotes for today's racing, so no trades on that front today. I have however already made my football plays for the afternoon, and wow! do I have plenty of 'em!

Supremacies (all 25pts stakes)
95] SOLD Man Utd @ 1.8 (IG)
96] SOLD West Ham @ 0.0 (SX,IG)
97] SOLD Barnsley @ 0.2 (All)
98] SOLD Sunderland @ 0.2 (All)
99] BOUGHT Plymouth @ 0.46 (SF)
100] BOUGHT Southampton @ 0.54 (SF)
101] SOLD Gillingham @ 0.4 (IG)
102] SOLD Yeovil @ 1.1 (SX)
103] SOLD Boston @ 0.3 (IG)
104] SOLD Peterborough @ 0.76 (SF)
105] BOUGHT Rochdale @ 0.5 (SP)
106] SOLD Swindon @ 1.1 (SP)
107] SOLD Wrexham @ 0.8 (SX)
108] SOLD Morecambe @ 0.1 (IG)
109] SOLD Burton @ 0.2 (SP)

Total Goals (All 25pts stakes)
110] SOLD Man Utd v Middbro @ 2.9 (SP)
111] SOLD Preston v Leicester @ 2.5 (All)
112] BOUGHT Boston v Chester @ 2.3 (SP)
113] BOUGHT Grimsby v Barnet @ 2.5 (SP)

Mini-performances (All 5pts stakes)
114] BOUGHT Sheff Utd @ 15 (SP)
115] SOLD Barnsley @ 23.5 (SF)
116] BOUGHT Crystal Pal @ 18 (SP)
117] BOUGHT Cardiff @ 18 (SF)
118] BOUGHT Port Vale @ 17.5 (SF)
119] BOUGHT Chester @ 17 (SF)
120] BOUGHT Barnet @ 16 (SP)



===== Previous trading results (from 20 April 2007) =====

88] SOLD Newbury Favs @ 72 (SP,SX) 3pts MU=85, [-39.00]
89a] SOLD Thirsk SPs @ 52 (SF) 1pt MU=52.7, [-0.70]
89b] SOLD Thirsk SPs @ 51 (SX) 2pts MU=52.7 [-3.40]
90] BOUGHT Thirsk Favs @ 65 (IG,SX) 2pts MU=80, [+30.00]
91] SOLD Hereford distances @ 35 (SX) 5pts MU=19.5, [+77.50]
92] SOLD Lingfield SPs @ 51 (SP, IG) 3pts MU=38.25, [+38.25]
93] SOLD Lingfield Favs @ 50 (SF) 3pts MU=30, [+57.12]
94] SOLD Derby supremacy @ 1.3goals (All) 25pts MU=1, [+7.50]

Day's return +167.27
+1011.61 to c/fwd



Happy spreading

 

[Data Miner]

22 April 2007
+646.30 b/fwd

Again not around to get today's racing quotes, and just a solitary football trade.

BOUGHT Norwich v Ipswich total goals @ 2.5 (SP) 25pts


===== Previous trading results (from 21 April 2007) =====
+1011.61 b/fwd

Supremacies
95] SOLD Man Utd @ 1.8 (IG) MU=0, [+45.00]
96] SOLD West Ham @ 0.0 (SX,IG) MU=1, [-25.00]
97] SOLD Barnsley @ 0.2 (All) MU=2, [-45.00]
98] SOLD Sunderland @ 0.2 (All) MU=-2, [+55.00]
99] BOUGHT Plymouth @ 0.46 (SF) MU=1, [+12.85]
100] BOUGHT Southampton @ 0.54 (SF) MU=1, [+10.95]
101] SOLD Gillingham @ 0.4 (IG) MU=1, [-15.00]
102] SOLD Yeovil @ 1.1 (SX) MU=1, [+2.50]
103] SOLD Boston @ 0.3 (IG) MU=1, [-17.50]
104] SOLD Peterborough @ 0.76 (SF) MU=3, [-56.00]
105] BOUGHT Rochdale @ 0.5 (SP) MU=5, [+112.50]
106] SOLD Swindon @ 1.1 (SP) MU=2, [-22.50]
107] SOLD Wrexham @ 0.8 (SX) MU=1, [-5.00]
108] SOLD Morecambe @ 0.1 (IG) MU=1, [-22.50]
109] SOLD Burton @ 0.2 (SP) MU=3, [-70.00]
Tot Gls
110] SOLD Man Utd v Middbro @ 2.9 (SP) MU=2, [+22.50]
111] SOLD Preston v Leicester @ 2.5 (All) MU=1, [+37.50]
112] BOUGHT Boston v Chester @ 2.3 (SP) MU=1, [-32.50]
113] BOUGHT Grimsby v Barnet @ 2.5 (SP) MU=5, [+62.50]
Mini-performances
114] BOUGHT Sheff Utd @ 15 (SP) MU=15, [break-even]
115] SOLD Barnsley @ 23.5 (SF) MU=40, [-82.50]
116] BOUGHT Crystal Pal @ 18 (SP) MU=0, [-90.00]
117] BOUGHT Cardiff @ 18 (SF) MU=0, [-90.00]
118] BOUGHT Port Vale @ 17.5 (SF) MU=20, [+11.90]
119] BOUGHT Chester @ 17 (SF) MU=0, [-85.00]
120] BOUGHT Barnet @ 16 (SP) MU=0, [-80.00]

Day's return -365.31
(No need to be disheartened, I'm fine with that. As with any trading regime, some days you take, others you give back.
So long as in the longer term the takes continue to outweigh the give-backs)

+646.30 to c/fwd

 

[Data Miner]

23 April 2007
+633.80 b/fwd

Today's racing;
122] SOLD Hexham SPs @ 52 (All) 3pts
123] SOLD Plumpton distances @ 41 (SF) 5pts
124] SOLD Plumpton SPs @ 29 (SP,IG) 3pts
125] SOLD Pontefract SPs @ 62 (SP,SX) 3pts

===== Previous trading results (from 22 April 2007) =====
+646.30 b/fwd

121] BOUGHT Norwich v Ipswich total goals @ 2.5 (SP) 25pts MU=2, [-12.50]

Day's return -12.50
+633.80 to c/fwd

 

[Data Miner]

Great info, thank you. Another question if I may, do you understand why your nets work? I don't mean in terms of the parameters that you use but in terms of the models that drive the functions. Can you give some additional insight into the technicalities and the maths?

I am not an expert in nets by any means but I have observed examples where such a model has been shown to work for a period of time but the operator was unable to explain fully why.

Thank you for your patient answering of my questions. This is very interesting information.

NQR

Sorry for the delayed response NQR, I've been busy elsewhere.

If I understand your question correctly the best answer I can offer is that neural nets are very good at smoothing noisy data into a useable, and pretty reliable predictions. Unlike traditional betting where the focus is more geared towards win/not win frequencies, spread betting success relies upon getting as close as possible to mean averages.

If for example a quote was 1.7-1.9 and the results from 7 identical situations turned out to be;

1, 1, 1, 1, 1, 1, 10

... as a traditional better (where the line would be "below 2 -or- equal to/greater than 2" - but essentially the same bet) I'd want to know that more often than not such a situation the result is lower (example here would return a "line" bettor 6 wins & a single loss. Betting "UNDERS" at even money, a return of +6, with -1 loss, 5pts profit from 7 bets.)
As a spread trader however it is more important for me to know the mean average. As such I see the exact same situation as 7 BUYING opportunities. My return is 6 losses @ 0.9 (=5.4) and a single win @ 8.1 for an overall return of +2.7.
(The above example of course is using extremes, to better illustrate the point. No conclusions should be inferred from comparing the profit levels)

So, if I were to evaluate past performances of Premier football teams using league positions. I could use x axis for home side league positions, y axis for away teams and plot mean average results on a 3rd dimensional z axis. I could mathematically smooth both gaps in the information and points that looked unusually high or low. But this is a pain in the butt for an amateur statistician - processing this info through a neural net and getting a useful output is a breeze.

But, such a simplistic model using just league positions is not likely to be as useful as one based upon a greater variety of pertinent indicators. So we add in, points gained, goals scored/conceded, recent form, injuries to key players, etc.

Whilst having the potential to give a more accurate picture, the above tactic also reduces the number of samples in each unique category, and many may have no previous examples in the training set. So the info that is there becomes very noisy, which is further complicated by our newly engineered gaps. The data also becomes mind-blowingly (well THIS mind anyhow!) multi-dimensional and complex.

Step forward my trusty artificial neural network. Feed the info in, and in double quick time its ability to minimise errors across the entire spectrum does all the work for me.

How does it do it? Or what does it do? I have no idea ... nor do I really care! I am a neural net USER, they do what I want them to do, far better than I ever could. My investments are based entirely on the neural net outputs, and 10 winning months from the last 12 confirms their effectiveness.

DM

 

[Data Miner]

23 April 2007, supplementary

Evening racing;
126] SOLD Sedgefield distances @ 31 (SF) 5pts
127] SOLD Windsor SPs @ 49 (IG) 3pts

Football;
128] SOLD West Brom superiority @ 0.2 (SX) 25pts

Happy spreading!

 

[NotQuiteRandom]

Hi DataMiner,

Thank you for the overview of your methodology.

I can't pretend that the lack of tractability and my lack of understanding of what takes place within the hidden layer of the net doesn't concern me but I obviously your results speak for themselves.

As I said, I know little more about nets than that which I have read here

http://en.wikipedia.org/wiki/Artificial_neural_network#Feedforward_neural_network

and some knowledge of genetic algorithms which helps.

What software do you use? Can you disclose the type of net you are running (forward feed, recurrent etc).

Are there sites online which publish useful sports data in a coherent format or do you gather your own?

My apologies for all the questions, like anything which is new it is very interesting.

Thanks, you write a great column.

NQR

Sorry for the delayed response NQR, I've been busy elsewhere.

If I understand your question correctly the best answer I can offer is that neural nets are very good at smoothing noisy data into a useable, and pretty reliable predictions. Unlike traditional betting where the focus is more geared towards win/not win frequencies, spread betting success relies upon getting as close as possible to mean averages.

If for example a quote was 1.7-1.9 and the results from 7 identical situations turned out to be;

1, 1, 1, 1, 1, 1, 10

... as a traditional better (where the line would be "below 2 -or- equal to/greater than 2" - but essentially the same bet) I'd want to know that more often than not such a situation the result is lower (example here would return a "line" bettor 6 wins & a single loss. Betting "UNDERS" at even money, a return of +6, with -1 loss, 5pts profit from 7 bets.)
As a spread trader however it is more important for me to know the mean average. As such I see the exact same situation as 7 BUYING opportunities. My return is 6 losses @ 0.9 (=5.4) and a single win @ 8.1 for an overall return of +2.7.
(The above example of course is using extremes, to better illustrate the point. No conclusions should be inferred from comparing the profit levels)

So, if I were to evaluate past performances of Premier football teams using league positions. I could use x axis for home side league positions, y axis for away teams and plot mean average results on a 3rd dimensional z axis. I could mathematically smooth both gaps in the information and points that looked unusually high or low. But this is a pain in the butt for an amateur statistician - processing this info through a neural net and getting a useful output is a breeze.

But, such a simplistic model using just league positions is not likely to be as useful as one based upon a greater variety of pertinent indicators. So we add in, points gained, goals scored/conceded, recent form, injuries to key players, etc.

Whilst having the potential to give a more accurate picture, the above tactic also reduces the number of samples in each unique category, and many may have no previous examples in the training set. So the info that is there becomes very noisy, which is further complicated by our newly engineered gaps. The data also becomes mind-blowingly (well THIS mind anyhow!) multi-dimensional and complex.

Step forward my trusty artificial neural network. Feed the info in, and in double quick time its ability to minimise errors across the entire spectrum does all the work for me.

How does it do it? Or what does it do? I have no idea ... nor do I really care! I am a neural net USER, they do what I want them to do, far better than I ever could. My investments are based entirely on the neural net outputs, and 10 winning months from the last 12 confirms their effectiveness.

DM

 

[Data Miner]

24 April 2007
+297.05 b/fwd

Today's racing;
129] SOLD Folkestone Favs @ 90 (SF) 3pts
130] SOLD Southwell distances @ 11.25 (IG) 15pts
131] SOLD Wolverhampton Favs @ 73 (SX) 3pts


===== Previous trading results (from 23 April 2007) =====
+633.80 b/fwd

122] SOLD Hexham SPs @ 52 (All) 3pts MU=41.6, [+31.20]
123] SOLD Plumpton distances @ 41 (SF) 5pts MU=60.7, [-98.50]
124] SOLD Plumpton SPs @ 29 (SP,IG) 3pts MU=23.5, [+16.50]
125] SOLD Pontefract SPs @ 62 (SP,SX) 3pts MU=98.1, [-108.30]
126] SOLD Sedgefield distances @ 31 (SF) 5pts MU=55.25, [-121.25]
127] SOLD Windsor SPs @ 49 (IG) 3pts MU=77.8, [-86.40]
128] SOLD West Brom superiority @ 0.2 (SX) 25pts MU=-1, [+30.00]

Day's return -336.75

+297.05 to c/fwd

 

[Data Miner]

24 April supplementary

++ Evening racing;
132] SOLD Towcester distances @ 40 (IG) 5pts
133] SOLD Bath distances @ 8.3 (SF) 15pts
134] SOLD Bath SPs @ 49 (SP,IG) 3pts
++ Football;
135] BOUGHT Macclesfield v BristolR Total goals @ 2.5 (IG) 25pts

Happy spreading!

 

[Data Miner]

25 April 2007
+343.91 b/fwd

Today's racing;
136] SOLD Catterick distances @ 10.25 (SP) 15pts
137] SOLD Epsom Favs @ 55 (IG) 3pts
138] SOLD Perth distances @ 39 (SF) 5pts
139] SOLD Perth SPs @ 60 (SF) 3pts


===== Previous trading results (from 24 April 2007) =====
+297.05 b/fwd

129] SOLD Folkestone Favs @ 90 (SF) 3pts MU=70, [+57.12]
130] SOLD Southwell distances @ 11.25 (IG) 15pts MU=16.1, [-72.75]
131] SOLD Wolverhampton Favs @ 73 (SX) 3pts MU=70, [+9.00]
132] SOLD Towcester distances @ 40 (IG) 5pts MU=37.25, [+13.75]
133] SOLD Bath distances @ 8.3 (SF) 15pts MU=4.95, [+47.83]
134] SOLD Bath SPs @ 49 (SP,IG) 3pts MU=39.2, [+29.40]
135] BOUGHT Macclesfield v BristolR Total goals @ 2.5 (IG) 25pts MU=1, [-37.50]

Day's return +46.86

+343.91 to c/fwd

 

[Data Miner]

25 April 2007 supplementary

140] SOLD Worcester SPs @ 57 (SP,SX) 3pts
141] SOLD Kempton SPs @ 57 (SX) 3pts
142] BOUGHT Northwich superiority @ -0.3 (SX) 25pts

Happy spreading!

 

[Data Miner]

Hi DataMiner,

Thank you for the overview of your methodology.

I can't pretend that the lack of tractability and my lack of understanding of what takes place within the hidden layer of the net doesn't concern me but I obviously your results speak for themselves.

As I said, I know little more about nets than that which I have read here

http://en.wikipedia.org/wiki/Artificial_neural_network#Feedforward_neural_network

and some knowledge of genetic algorithms which helps.

I attempt some basic explanations on my web site.
Also might be well worth taking a look at TRAJAN and downloading the demo. I can't vouch for the software .. and it is VERY expensive .. but the manual that comes with it is one of the best neural net overviews I've seen - and in understandable language.

What software do you use? Can you disclose the type of net you are running (forward feed, recurrent etc).

A bit like recommending a car, it is personal preference rather than "This is the the very best, etc". Currently I favour three products because they all have their particular strengths.
Predictor (Ward Systems) which is a hybrid ANN and unlike anything else I've tried. But it is a very strong predictor which stand up well to comparisons. Unfortunately no demo is available.
Tiberius (Neusolutions). Has the bonus of allowing the trained net to be output as code, so predictions (for me) can be performed within the spreadsheet I use rather than having to process everything through the black-box master module. A demo is available for download
Gene X Pro This is a great genetic programming toolkit (so NOT a neural net) with more options than you can shake a stick at. The very nature of this type of program is that it also generates code to use elsewhere (a great time saver). I've only had this one for a short time and have I feel only scratched the surface, but am very excited about its potential. Demo available from their website.

The feature I like about all such software is that the data fed in and the parameters you set are unique to each user - so the end product is that you are definitely NOT going with the crowd.

Are there sites online which publish useful sports data in a coherent format or do you gather your own?

Football-data is the source I use for my soccer betting. Inexpensive one-off payment brings several years league football results and current season results/fixtures updated almost daily.

Horse racing data is not so easily (or cheaply) obtained. I use Proform for daily up-to-date results & fixtures and a 9-year database of racing results. Subscriptions range from £10.25 to £11.50 per week. (Anyone interested don't blindly place an order, I can organise a £20 cash-back)

I don't do golf, but there's a free updated database of international results at bettingmad

Sports spread betting make-ups for the past six months are available on the sportingindex website, but historical spread quotes are virtually impossible to obtain (so far as I've been able to find). So much so I attempted to start a sports-spread-quote exchange on a web site - but work there is still in progress.

My apologies for all the questions, like anything which is new it is very interesting.
Thanks, you write a great column.

NQR

No problems NQR - I enjoy discussing it!

DM

 

[Data Miner]

Taking the trip from +1011 to -163 in just 5 days - such is life ...

26 April 2007
-163.95 b/fwd

Today's racing;
143] SOLD Perth distances @ 34 (IG) 5pts
144A] SOLD Fontwell SPs @ 41 (SF) 1pt
144B] SOLD Fontwell SPs @ 40 (IG) 2pts


===== Previous trading results (from 25 April 2007) =====
+343.91 b/fwd
136] SOLD Catterick distances @ 10.25 (SP) 15pts MU=21.25 [-165.00]
137] SOLD Epsom Favs @ 55 (IG) 3pts MU=120 [-195.00]
138] SOLD Perth distances @ 39 (SF) 5pts MU=49.25 [-51.25]
139] SOLD Perth SPs @ 60 (SF) 3pts MU=68.3 [-24.90]
140] SOLD Worcester SPs @ 57 (SP,SX) 3pts MU=91 [-102.00]
141] SOLD Kempton SPs @ 57 (SX) 3pts MU=49.4 [+22.80]
142] BOUGHT Northwich superiority @ -0.3 (SX) 25pts MU=0 [+7.50]

Day's outcome -507.85

-163.95 to c/fwd

 

[Data Miner]

26 April supplementary

145] SOLD Yarmouth distances @ 10 (SP) 15pts
146] SOLD Yarmouth Favs @ 77 (SP) 3pts
147] SOLD Southwell Favs @ 60 (SF) 3pts

Happy spreading!

 

[Data Miner]

27 April 2007
-33.67 b/fwd

Today's daytime racing;
148] SOLD Perth distances @ 38 (SP,IG) 5pts
149] SOLD Wolverhampton favs @ 88 (SF) 3pts


===== Previous trading results (from 26 April 2007) =====
-163.95 b/fwd
143] SOLD Perth distances @ 34 (IG) 5pts MU=43.75, [-48.75]
144A] SOLD Fontwell SPs @ 41 (SF) 1pt MU=36.1, [+4.66]
144B] SOLD Fontwell SPs @ 40 (IG) 2pts MU=36.1, [+7.80]
145] SOLD Yarmouth distances @ 10 (SP) 15pts MU=7.7, [+34.50]
146] SOLD Yarmouth Favs @ 77 (SP) 3pts MU=42.5, [103.50]
147] SOLD Southwell Favs @ 60 (SF) 3pts MU=50, [+28.56]

Day's outcome +130.27

-33.67 to c/fwd

 

[Data Miner]

27 April 2007 (evening session)

150] SOLD Sunderland supremacy @ 1.2 (SX,IG) 25pts

2 losses from 2 this afternoon. This one will either give me the loser's hat-trick ... or will it prove to be a saver?!?

Happy spreading!

 

[Data Miner]

T2W's day off (upgrading) yesterday means there's a one day gap in my reporting on here. Still, the new look is an improvement ... but best of all the page loads far quicker than it used to!

Anyway, on with the daily trading ...

29 April 2007
+34.95 b/fwd

Today's racing;
187] SOLD Ludlow distances @ 33 (All) 5pts
188] SOLD Ludlow SPs @ 37 (SP) 3pts
189] SOLD Wetherby SPs @ 49 (SX) 3pts
Football
190] SOLD Arsenal superiority @ 1.5 (SX) 25pts



===== Previous trading results (from 28 April 2007) =====
-158.55 b/fwd

151] SOLD Chelsea @ 1.6 (SX,IG) MU=0, [+40.00]
152] SOLD Man Utd @ 0.8 (SP,SX) MU=2, [-30.00]
153] SOLD Tottenham @ 0.0 (SP,SX) MU=1, [-25.00]
154] SOLD Liverpool @ 0.2 (All) MU=-1, [+30.00]
155] SOLD Sheff Utd @ 0.9 (SX) MU=1, [-2.50]
156] BOUGHT Wigan @ 0.26 (SF) MU=-3, [-81.50]
157] BOUGHT Cardiff @ 0.14 (SF) MU=-1, [-28.50]
158] SOLD West Brom @ 0.2 (SX) MU=1, [-20.00]
159] SOLD Leeds @ 0.6 (SX) MU=0, [+15.00]
160] SOLD Southend @ 0.8 (SX) MU=-2, [+70.00]
161] SOLD Wolves @ 0.9 (SX,IG) MU=2, [-27.50]
162] SOLD Oldham @ 0.2 (SX) MU=1, [-20.00]
163] SOLD Bristol C @ 0.3 (SX) MU=-1, [+32.50]
164] SOLD Cheltenham @ 0.08 (SF) MU=2, [-48.00]
165] SOLD Yeovil @ 0.7 (SP) MU=1, [-7.50]
166] BOUGHT Accrington @ 0.3 (SP) MU=1, [+17.50]
167] SOLD Boston @ 0.8 (SP,SX) MU=0, [+20.00]
168] SOLD Peterborough @ 0.3 (SP,SX) MU=0, [+7.50]
169] SOLD Shrewsbury @ 0.2 (SP,SX) MU=1, [-20.00]
170] BOUGHT Walsall @ 0.8 (SX) MU=1, [+5.00]
171] SOLD MKDons @ 0.0 (SX) MU=2, [-50.00]
172] SOLD Altringham @ 0.6 (All) MU=0, [+15.00]
173] SOLD Burton @ 1.0 (IG) MU=-1, [+50.00]
174] SOLD Cambridge @ 0.7 (SX) MU=1, [-7.50]
175] BOUGHT Dagenham&Red @ 0.3 (SX) MU=1, [+17.50]
176] SOLD Exeter @ 1.3 (IG) MU=1, [+7.50]
177] SOLD Stafford @ 0.6 (IG) MU=-1, [+40.00]
178] SOLD York @ 0.5 (SX) MU=1, [-12.50]
Total Goals
179] SOLD Crewe/Tranmere @ 2.9 (SP,IG) MU=2, [+22.50]
180] SOLD St Albans/Weymouth @ 2.9 (SP,SX) MU=1, [+47.50]
Racing
181] SOLD Leicester distances @ 10.75 (IG) 15pts MU=12.05, [-19.50]
182] SOLD Leicester Favs @ 87 (SP,IG) 3pts MU=75, [+36.00]
183] SOLD Market Rasen distances @ 34 (IG) 5pts MU=22, [+60.00]
184] SOLD Market Rasen SPs @ 46 (IG,SX) 3pts MU=57, [-33.00]
185] SOLD Market Rasen Favs @ 79 (SP,IG) 3pts MU=40, [+117.00]
186] SOLD Wolves Favs @ 57 (SP,SX) 3pts MU=65, [-24.00]

Day's outcome +193.50

+34.95 to c/fwd