Hey Mike (and Ken), whatcha think....

[Carl]

We talked briefly at the Convention about putting the last three races for a horse on one line.

This made sense to me, and If I can be of any help in pursuing it I will. A few preliminary thoughts.
1) We have 255 fields to work with, and we are only using maybe 100 including IMPACT.
2) IMPACT is not a problem, since it would I think be the same no matter which of the horses last three races were chosen.
3) Looks to me as if around 40% of the HTR fields are "fixed", which is to say that they wouldn't change regardless of paceline chosen. You know much more about this than I do.
4) We could call fields by races back (rEv1LR, rEV12B, rEV13B for "Last race", "two races back", "three races back"). Some amazing patterns to research.

To me, being able to work with data from the last three races rather than just the last one would be like moving from a two dimensional world to real world in full color. A horses first off layoff query would look like
1) nLayLR <=30
2) nLay2B >=90

And so on. Cool query to look for improving Ev, or improving/declining par numbers, and......THE SKY'S THE LIMIT!

later,
Carl

[MikeDee]

I don't have a problem with it, but it is really up to Ken and what he is able to do. Need to keep everything one a single line though

[ERNIE]

Been there,done that dept.=
For my two cents worth : to use last three lines on a spreadsheet, was looking at combo's of whole figures and rankings.
I used "htr4msa" to create a file useing pl-1.....imported it to the spreadsheet,saved it then went back to htr4msa created a file useing pl-2 only this time I selected only the second line of every horse..........then when I went to Lotus I imported into the same spreadsheet only starting at the last field ........did the same thing for the third paceline..........

only to find that this isn't what I wanted from a spread sheet..........access may be able to do something that I couldn't see , the spreadsheet being so large by then..........

however I have built spreadsheet using screen #5 the past performances useing each horses last three lines "only marking what line it was 1-2-3"........results were that what paceline it was didn't matter........... ranking all three lines the winner will be the best pace or per figure ranked "a or b" most of the time.........

because it is so time consumming to build this by hand I can't claim to have thousands of races but I think I have enough to suggest that an export program that exports the last 4-5 lines of every horse is what your looking for.......then you can compare the best of the best performances..........Now to the examples: look at cd 7/4 race (question should I example the longshot first or not ????) race #1 the highest horse per figure is #4's last race "90" paid $4.00 to win........top horse wins second race @$5.60 and race three you look and wait for horse #7 ranked second and third ,wins paying $23.00 On to bel 7/4 R8 paid $15.00 top ranked horse ....... results of the day 7/4/01 @ cd wins 7 of 11 races +$1.13 bel wins 6 of 10 $1.34 betting top two blind............However, as always you stay at the crap table long enough you find a losing streak..... go to hol same day win 4 of 10 return .62 cents on the $$$
Now that this has turned into a book........my thoughts and recomendation is that to spread more information out on one line only gives you more information but nothing to compare to but if you look at and compare a number of lines for each horse you will be able to see the best of the bunch..... hope this helps your food for thought for today.......I know seeing your post on this subject got me started again................. ;) :D

[Carl]

O.K., I want to get started on having more than one paceline available.

So..... can I
1) Download data twice, into different tables, once as pace line one (last race) and once as paceline three (best two out last three) and then
2) Link the tables or do a make table combining the two?

Yes no maybe? Am I going to get into trouble OR into uncharted territory OR has someone "been there done that"?

Inquiring minds want to know....but not to re-invent the wheel.

later,
Carl

[MikeDee]

I have thought about doing something like that as well but never did it becasue I doubted that the result would be worth the effort.

Yes it can be done. Here is what I would do. Import PL5 into one table and PL1 into another.
Join them in the normal maner with the 4 fields we use to key with.

Then put all the fields from one of the table (lets say PL5 into the design grid of a query. Then put only the fields that are pace line dependent from PL1 in the design grid.

Now the fun part, change the query to a make table query run it and see what happens. I'm not sure if access will let you make a table becasuse it will have duplicate field names unless it includes the table name that it came from as part of the field name of the new table.

If access makes a table you will need to go in and change the field names to something you can live with that gives each field a unigue name.

If access won't let you then you need to make a table with the fields from PL1 and then manually add the field names that you want from pl5 to the desgin of this table, the field names will have to be unique as you cannot have duplicates in field names.

When finshed with the new table go back to your query that has joined data and do a append query. One by one in the design grid tell access which field to append the data to in your new table.

now you have it and the append query will map the data from pl1 and pl5 into a single table.

see how easy this will be? If you decide to do it let me know be glad to answer any questions as you build it.

[hurrikane]

Carl,
I've been doing what you ask for a while. Was using it with PL1 and PL5 to produce Georges K1s each day.

Just a word of caution. Be very careful as the all_msa.txt doesn't say which table is which PL it is easy to get the wrong data into the wrong table. This corrupted my main database file as I got som PL1 data in with the PL5 data in there. Had to delete the PL1 data and redo those days. Just a word of caution. It is a lot of work..

[Carl]

Thanks guys.

I thought I would try it with pl 3 ( the description is "best two out of three last races") and Pl1.

I brought in 2013 race lines in each. Not so bad, out of 222 races, I have 97 that have different horses chosen for rAP, and 87 that have different ev1.

I think I can "remark" certain factors easily on a make table by changing column headings I dump the data into.

My question is, what is Pl 3? When it says "best out of last two" and only brings in one paceline, what is that paceline? A hybrid or a "real"? Anyone know?

later,
Carl

[Glen]

My guess is average of the two...?!??

[George]

Glen is right that PL3 is an average. PL3 is average of "best two" out of last three.

PL2 is a single line "best of last three".

PL1 is a single line last race unless last was dnf or dis.