Sports Day Trading

11 May 2007 (afternoon session)
-607.46

68] BOUGHT Chester Favourites @ 60 (SF) 3pts
69] SOLD Nottingham Favourites @



===== Previous trading results (from 09 May 2007) =====
-575.34 b/fwd

62] SOLD Chester Favs @ 76 (SF) 3pts MU=100, [-72.00]
63] SOLD Goodwood distances @ 12.3 (SF) 15pts MU=9.15, [+44.98]
64] SOLD Goodwood Favs @ 83 (SP) 3pts MU=110, [-81.00]
65] SOLD Newton Abbot SPs @ 49 (SP) 3pts MU=22.9, [+78.30]
66] BOUGHT Newton Abbot Favs @ 72 (SP) 3pts MU=140, [+204.00]
67] SOLD Ayr SPs @ 57(SP) 3pts MU=125.8, [-206.40]

Day's outcome -32.12

-607.46 to c/fwd :(
 
11 May 2007 (evening session)

70] SOLD Hamilton SPs @ 38 (SX) 3pts
71] BOUGHT Hamilton Favs @ 57 (IG) 3pts
72] SOLD Wincanton distances @ 34 (SP,SX) 5pts

Happy spreading!
 
12 May 2007 (afternoon session)
-473.90 b/fwd

73] SOLD Ascot Favs @ 56 (IG) 3pts
74] SOLD Lingfield distances @ 13.3 (SF) 15pts
75] SOLD Nottingham SPs @ 53 (SP.IG) 3pts
76] BOUGHT Hexham Favs @ 50 (IG) 3pts



===== Previous trading results (from 09 May 2007) =====
-607.46 b/fwd
68] BOUGHT Chester Favourites @ 60 (SF) 3pts MU=70, [+28.56]
69] SOLD Nottingham Favourites @ 66 (IG) 3pts MU=50, [+48.00]
70] SOLD Hamilton SPs @ 38 (SX) 3pts MU=36.3, [+5.10]
71] BOUGHT Hamilton Favs @ 57 (IG) 3pts MU=75, [+54.00]
72] SOLD Wincanton distances @ 34 (SP,SX) 5pts MU=36, [-10.00]

Day's outcome +125.66

-473.90 to c/fwd :(
(amount adjusted slightly in line with actual trading statements)
 
12 May 2007 (evening session)

77] SOLD Thirsk SPs @ 39 (SF) 3pts
78] BOUGHT Thirsk Favs @ 66.5 (SF) 3pts
79] BOUGHT Warwick Favs @ 60 (SX) 3pts

Happy spreading!
 
13 May 2007
-560.42 b/fwd

80] BOUGHT Plumpton Favs @ 74 (SP.IG) 3pts
81] SOLD Uttoxeter distances @ 45 (SX) 5pts
82] SOLD Worcester distances @ 38 (SX) 5pts
83] SOLD Worcester SPs @ 53 (SX) 3pts
84] BOUGHT Worcester Favourites @ 66 (SF) 3pts



===== Previous trading results (from 12 May 2007) =====
-473.90 b/fwd

73] SOLD Ascot Favs @ 56 (IG) 3pts MU=90, [-102.00]
74] SOLD Lingfield distances @ 13.3 (SF) 15pts MU=13.6, [-4.50]
75] SOLD Nottingham SPs @ 53 (SP.IG) 3pts MU=30.8, [+66.60]
76] BOUGHT Hexham Favs @ 50 (IG) 3pts MU=35, [-45.00]
77] SOLD Thirsk SPs @ 39 (SF) 3pts MU=16.3, [+64.83]
78] BOUGHT Thirsk Favs @ 66.5 (SF) 3pts MU=80, [+38.55]
79] BOUGHT Warwick Favs @ 60 (SX) 3pts MU=25, [-105.00]

Day's outcome -86.52

-560.42 to c/fwd :(
 
14 May 2007 (afternoon session)
-403.42 b/fwd

85] SOLD Southwell SPs @ 56 (SP,IG) 3pts
86] SOLD Wolverhampton SPs @ 46 (SX) 3pts


===== Previous trading results (from 13 May 2007) =====
-560.42 b/fwd
80] BOUGHT Plumpton Favs @ 74 (SP.IG) 3pts MU=70, [-12.00]
81] SOLD Uttoxeter distances @ 45 (SX) 5pts MU=27.8, [+86.00]
82] SOLD Worcester distances @ 38 (SX) 5pts MU=25, [+65.00]
83] SOLD Worcester SPs @ 53 (SX) 3pts MU=41, [+36.00]
84] BOUGHT Worcester Favourites @ 66 (SF) 3pts MU=60, [-18.00]

Day's outcome +157.00

-403.42 to c/fwd :(
 
14 May 2007 (evening session)

87] SOLD Windsor distances @ 11 (SP) 15pts
88] SOLD Towcester favourites @ 59 (IG) 3pts

Happy spreading!
 
15 May 2007 (afternoon session)
-256.87 b/fwd

89] SOLD Southwell SPs @ 53.5 (SF) 3pts


===== Previous trading results (from 14 May 2007) =====
-403.42 b/fwd

85] SOLD Southwell SPs @ 56 (SP,IG) 3pts MU=41.5, [+43.50]
86] SOLD Wolverhampton SPs @ 46 (SX) 3pts MU=81.4, [-106.20]
87] SOLD Windsor distances @ 11 (SP) 15pts MU=3.85, [+107.25]
88] SOLD Towcester favourites @ 59 (IG) 3pts MU=25, [+102.00]

Day's return +146.55

-256.87 to c/fwd :(
 
15 May 2007 (evening session)

90] SOLD Huntingdon distances @ 31 (All) 5pts


Happy spreading!
 
More than the sum of its parts?

The implementation of "ensembles", "committee" and/or "team" solutions are frequently discussed in machine learning circles. The terms used merely refer to the act of taking a mean average from several AI generated predictions, and using this as the favourable option. There is proof that the resulting output will not just give a "mushy" middle-ground prediction (as intuition might suggest), but will in fact frequently give a more accurate result than any of the individual sources used, and will never be worse than the least accurate.

Being an individual who is all for the easy life, I always shunned the idea as increasing my data processing workload. Do I need to make life more complicated for the possibility little extra accuracy?

But a recent interesting observation on another newsgroup highlighted the fact that in a recent data mining competition (where machine learning solutions go head-to-head for top prize) combining the predictions of the top performers gave a better result than any one of the individuals. In many ways then, the whole, combined to a single solution outperformed any of the individual parts.

The data mining competition was hosted by the Pacific-Asia Knowledge Discovery and Data Mining conference (PAKDD 2007) and was co-organized by the Singapore Institute of Statistics.

Even though some of the individual players employed "team" solutions themselves, a simple average compiled from the top 27 performers would have won the competition outright.
If the top 36 were averaged, this would have made runner-up spot.

Seems counter-intuitive somehow, that "average" can ever win the day, but in the realms of AI, two machine "heads" are obviously more often than not, better than one.

To cross-check, I compared two of my programs on some of my existing horse race data, then examined how an average of the two would fare.

The figures refer to cumulative absolute errors, so the lower the number the better.


NNet1 NNet2 "Team" combination
P1 5962 5956 5807
P2 6912 6957 6892
P3 11207 11558 11167
P4 1137 1124 1123
P5 5068 5110 5074
P6 7272 7499 7332


The rows, P1, P2 ... are the cases to which I'm seeking a solution
NNet1 & 2 are stand alone software programs, "Team" is the mean average of the outputs of the two programs, NNet1 & 2.

"Team" average scores top place in 4 of the 6 cases, and 2nd place in the remaining two cases (one of these a very close 2nd)

This I think is impressive. An area I have been ignoring for far too long perhaps?

DM
 
17 May 2007 (afternoon session)
-120.96 b/fwd

91] SOLD Salisbury SPs @ 48 (IG) 3pts
92] SOLD Salisbury Favourites @ 85 (SP,IG) 3pts

===== Previous trading results (from 15 May 2007) =====
-256.87 b/fwd

89] SOLD Southwell SPs @ 53.5 (SF) 3pts MU=34.8, [+53.41]
90] SOLD Huntingdon distances @ 31 (All) 5pts MU=14.5 [+82.50]

Day's outcome +135.91

-120.96 to c/fwd :(
 
17 May 2007 (evening session)

93] SOLD Folkestone distances @ 40 (SP,IG) 5pts
94] BOUGHT Folkestone Favourites @ 75 (SP,SX) 3pts
95] SOLD Ludlow SPs @ 43 (IG,SX) 3pts

Happy spreading!
 
18 May 2007 (afternoon session)
-92.86 b/fwd

96] SOLD Newbury Favourites @ 86 (SP) 3pts
97] SOLD Newmarket distances @ 10.5 (SP) 5pts
98] SOLD Newmarket SPs @ 53 (SP) 3pts
99] BOUGHT Newmarket Favourites @ 62 (SP,IG) 3pts


===== Previous trading results (from 17 May 2007) =====
-120.96 b/fwd

91] SOLD Salisbury SPs @ 48 (IG) 3pts MU=37.6, [+31.20]
92] SOLD Salisbury Favourites @ 85 (SP,IG) 3pts MU=65, [+60.00]
93] SOLD Folkestone distances @ 40 (SP,IG) 5pts MU=45, [-25.00]
94] BOUGHT Folkestone Favourites @ 75 (SP,SX) 3pts MU=45, [-90.00]
95] SOLD Ludlow SPs @ 43 (IG,SX) 3pts MU=25.7, [+51.90]

Day's outcome +28.10

-92.86 to c/fwd :(
 
18 May 2007 (evening session)

100] SOLD Aintree distances @ 35 (All) 5pts
101] SOLD Aintree favourites @ 70 (SP) 3pts

Happy spreading!
 
19 May 2007 (afternoon session)
-411.61 to b/fwd

102] SOLD Bangor Favourites @ 80 (SP,IG) 3pts
103] SOLD Newbury Favourites @ 88.5 (SF) 3pts
104] SOLD Thirsk SPs @ 60 (All) 3pts



===== Previous trading results (from 18 May 2007) =====
-92.86 b/fwd

96] SOLD Newbury Favourites @ 86 (SP) 3pts MU=95, [-27.00]
97] SOLD Newmarket distances @ 10.5 (SP) 5pts MU=10.05, [+2.25]
98] SOLD Newmarket SPs @ 53 (SP) 3pts MU=51.5, [+4.50]
99] BOUGHT Newmarket Favourites @ 62 (SP,IG) 3pts MU=30, [-81.00]
100] SOLD Aintree distances @ 35 (All) 5pts MU=63.5, [-142.50]
101] SOLD Aintree favourites @ 70 (SP) 3pts MU=90, [-60.00]

Day's outcome -318.75

-411.61 to c/fwd :(
 
19 May 2007 (evening session)

105] SOLD Uttoxeter distances @ 35 (SP) 5pts
106] SOLD Worcester distances @ 36 (SP,SX) 5pts

Happy spreading!
 
20 May 2007
-71.35 b/fwd

107] SOLD Market Rasen SPs @ 54 (SP) 3pts
108] BOUGHT Market Rasen favourites @ 67 (SF) 3pts
109] SOLD Ripon SPs @ 47 (SP,SX) 3pts
110] BOUGHT Ripon favourites @ 69 (SP) 3pts


===== Previous trading results (from 19 May 2007) =====
-411.61 b/fwd

102] SOLD Bangor Favourites @ 80 (SP,IG) 3pts MU=65, [+45.00]
103] SOLD Newbury Favourites @ 88.5 (SF) 3pts MU=15, [+209.92]
104] SOLD Thirsk SPs @ 60 (All) 3pts MU=72.8, [-38.40]
105] SOLD Uttoxeter distances @ 35 (SP) 5pts MU=27.75, [+36.25]
106] SOLD Worcester distances @ 36 (SP,SX) 5pts MU=18.5, [+87.50]

Day's return +340.27
-71.35 to c/fwd
 
21 May 2007 (afternoon session)
+53.86 b/fwd

111] SOLD Bath favourites @ 70 (SP) 3pts
112] BOUGHT Newton Abbot favourites @ 77 (SF) 3pts


===== Previous trading results (from 20 May 2007) =====
-71.35 b/fwd

107] SOLD Market Rasen SPs @ 54 (SP) 3pts MU=28.7, [+75.90]
108] BOUGHT Market Rasen favourites @ 67 (SF) 3pts MU=110, [+122.81]
109] SOLD Ripon SPs @ 47 (SP,SX) 3pts MU=57.5, [-31.50]
110] BOUGHT Ripon favourites @ 69 (SP) 3pts MU=55, [-42.00]

Day's outcome +125.21

+53.86 to c/fwd :)
 
Top