Martin,
I don't know if this is relevant to your approach but it may be of some interest.
It's alot of work to keep up manually (pen and paper) and I have done it in the past, but my school motto was labor omnia vincit

PACEWINNERS
(Quoted from the Raceform Handicap Book No. 331, 25 Nov 89)

"A consideration of the pace of a race seems to be overlooked when considering speed figures and race times. It could prove very rewarding, particularly at the beginning of a season in pinpointing fit horses.
The simplest method of calculating the pace of a race is explained in the back of any Superform annual. Using a calculator anyone should be able to calculate the pace of all races at a meeting in under 5 minutes. Using a computer is faster. My computer allows me to calculate and print out pace figures for a week's returns in about 30 minutes.
Your analysis of Mr Proudlock's system for pinpointing winners started me looking at early season pace figures.
It seems logical to me that a horse winning at a very fast pace at the start of the season gives an indication of fitness that could give it a decisive advantage.
At some meetings, the fastest pace race was well above the pace of the next fastest. I noted the winner of the race and the second if it was within 5 lengths of the winner. The results are quite pleasing.
A very high proportion of winners (and seconds) win their next race, although the odds are not all that marvellous, possibly due to small fields.
A couple of examples of pace calculations are shown below, if anyone wishes to check list in table 1 (I used Sporting Life standard times so there might be a slight difference if RHB figures are used).

(1) DEVON 16/8

Race (1) Dis (2) Dif Std Time(3) Dif / Dis x 100(4) Pace (5)
1 3yo H 17 2.6 15 +8
2 Hc Nov H 21 24.5 117 -94
3 Nov Ch 17 -2.7 -16 +3
4 CJSH 17 10.2 60 -3
5 Hc Ch 17 0.0 0 +23
6 Hc H 17 -6.5 -38 +61*

Add the six figures in column (4), total = 138. Divide this by the number of races, 6 = 23. This gives the average pace of the meeting. Column (5) is simply the difference between the average pace and the figure in column (4). As can be seen, races 2 and 4 are slow run races while 1,3, 5 and 6 are faster than average. Race 6 has been run at a much faster pace than the others and the winner (Midnight Strike and 2nd within 5 lengths, Imadyna) are noted.

HEREFORD 19/8

Race (1) Dis (2) Dis Std Time (3) Dif/Dis x 100 (4) Pace (5)
1 Nov H 19 1.6 8 +38
2 Nov H 16 9.0 56 -10
3 Hc Ch 16 1.6 10 +36
4 Am H 25 12.4 50 -4
5 Nov Ch 25 22.4 90 -44
6 Hc H 16 10.0 63 -17

Average pace of this meeting is 46. Since there is no race that is clearly faster no horses are listed from this meeting.

(There follows a Sample Table of winners and seconds of fast pace races that I have omitted -Jack)

One of your Forumites says a check of several seasons is needed. Last season I calculated and stored on computer over 8000 speed ratings for the 88/89 season. With the speed ratings are stored 12 other factors including, date, course, going, lengths won or lost and the pace of the race. The fastest pace of the meeting has a * by the pace figure. I cannot submit these figures since over 150 pages of printout is involved.
I then looked at chasers that had won or placed within 10 lengths and their record after recording the fast pace (I took 8 weeks as a reasonable time to keep form). From 150 qualifiers 99 horses won a total of 151 races in that period. This is giving about 5 winners a week and seems to be useful to follow up, however you still have to find the winners."

(There follows an analysis of the Cheltenham Festival which I have omitted - Jack)