RDSS Modeller Export

[Ted Craven]

Hi folks. I am testing the layout of output to an Excel file from RDSS. May I have commentary on how suitable the layout is? All factors are dumped to both an Excel and CSV file.

Today's race is identified, the output from all Analysis screens including results and Mutuels and tote info. From this file you would cut and paste into a master spreadsheet however you wanted to organize things, for example one spreadsheet file per track, one tab per surface and distance, or group distances, etc.

Your feedback is appreciated!

Some races and sample files follow (I'll just keep adding to this post for a while, as races run tonight).

WO 07/22 #3
WO20090722-03.xls
WO20090722-03.csv

WO 07/22 #4
WO20090722-04.xls

DMR 07/22 #8
DMR20090722-08.xls

EVD 07/22 #5
EVD20090722-05.xls

Ted

[Ted Craven]

WO 07/22 #6
WO20090722-06.xls

That's enough for now! Depending on feedback, I'll try to get this update to a few folks to test tomorrow or Friday, with a release late weekend or early next week.

Cheers,

Ted

[mackenziem]

This is definately going to be a huge time saver!

[gandalf380]

i have minor issues with some of the formatting. I think it is easier to read for instance if the payoffs are in currency form. I also would like to see some of the columns switched around but I guess that is something that the individual would set up?

[dogkatcher]

Just curious Ted, will one be able to export only the winners data and not the whole field? I guess we could delete/hide until only the winner remained and then export. These are just samples your showing and may not represent what one would or could do with the data. Thanks, as this looks like a terrific time saver.

[Ted Craven]

A bit of a preamble:

I consider this modelling export to be an interim solution (i.e. better than nothing) while I work on the next version of RDSS which will have a much more robust and elegant modeller, wager tracker, slice-it-and-dice-it-any-way data tracking solution. And thus, I was hoping to not spent too, too much time on this. Also, this output, while it does provide results and mutuels if you are connected to the RDSS toteboard module, is not really intended as a Wager Decision Form - just factor tracker. Though with the raw data, those who know how could do anything they wish, and the rest of us can simply cut and paste.

That said, I agree money columns should be formatted $000.00, rank columns should be centered, all velocites should have 1 decimal, etc. My 'quick and dirty' solution means I can't apply these formats on an automated basis, but when you cut and paste from the output file to your master spreadsheet, that target sheet could of course have all such formatting, colour, whatever.

Charlie, what re-arrangements of data columns would you suggest?

John, every line in the Analysis screens gets exported. If you have 5 lines (including Winner, Place and Show horses) they are all exported. Just copy the Winner's line into a master spreadsheet for your Win Model, by track (and by whatever fine-grained organization you further require: e.g. one tab per distance/surface, group distances together, etc).

If any Excel experts can take the layout as supplied (or as amended before the update is released) and add some formatting to a master template, which formatting is applied to rows of data copied from the RDSS output file, please feel free to post or send to me.

BTW, I am making the output both in Excel 2003 (XLS) and Comma Separated Value (CSV) format. If you don't use Excel, you can likely import CSV to any software. There's a CSV sample file for the first WO race above.

Cheers,

Ted

[SilentRun]

Ted,

I was able to read the CSV file..thx.

There is adequate data to pick and choose whatever suites your tracking
criteria I think this is great idea for record keeping.

Personally I would just like to see data for the win and place horses only.
My tracking criteria is limited to about 8 data elements such as E/L, LPR, DCL,
TPP, 2nd call... I remember in one issue of the Follow-Up Sartin had an
article that showed more data elevated ones confidence but lowered the results.
I guess you can say , "Less is More".

Ernie

[Profiler8]

Very good work Ted ! I think it is good to, that the all the contenders are presented.
Everybody can copy the things he or she needs. ;-)

Greetings
Tim

[Ted Craven]

That's great, re the CSV files!

I think there might possibly be a misunderstanding about the purpose of these export files. First, the export file (samples above) are not intended to be your modelling spreadsheet - merely the raw output from RDSS which you use to copy 1 or 2 lines INTO your permanent modelling spreadsheet (e.g. by track). If you want to copy the line for the Winner (i.e. marked "W" in the Result column), just copy that 1 line into your model sheet. Similarly, if you model Place, copy the line for the Place horse into your Place model spreadsheet. After you've copied the rows you're interested in, you can delete the export file.

If you normally have 5 lines in your RDSS Analysis screens, you'll see 5 lines in the export file. I will provide a master spreadsheet template with layout matching the export file for folks to use to keep their models in. Really, 4 keystrokes transfers ALL the data items into your permanent model spreadsheet.

I agree there is way too much data in the export file! In addition to identity, result and mutuel columns, I myself will likely normally look at the following factors: TE, BL/BL, VDC, E/ep, L/ep, TPP, E/L, TP+F3, maybe Hid, possibly 2nd call Position and Beaten Lengths. These factors, I suspect, apply (by track, distance group, surface) more-or-less regardless of a matchup within the race. Other factors, such as F1, F2 (either PoH or PoR) F3, DCL are likely matchup sensitive and might become rather meaningless when removed from the context of what the other runners looked like. But I'm open to whatever the info will show me!

Notice I have omitted above some popular (and useful) factors found on the Segments screen - only because (when shown as ranks) they are wholly redundant to other factors. For example: CBL1 = PoH F1, CBL-SC = PoH SC, CBL Final = PoH TS, MUG panels 1 & 2 = PoH F1, F2. Only the visual presentation of the rank info and differential from best is different. Spend some time browsing the ranks of the different columns in the spreadsheet to see which column ranks are always completely redundant.

I included (at the far right-side of the sheet) all the adjusted time and velocity numbers (because some folks asked me to), and because some like to calculate all the older factors from Phase III (i.e. Brohamers MPH), such as EP, SP, AP, %E, etc (which are no longer shown in the more modern software). I can recommend doing this only to eventually prove that reliance on those older factors alone produces lower average mutuels than using the more modern factors. But hands-on discovery is the essence of understanding.

Note: one column, the E/L Rank, is new and also does not appear on the RDSS Analysis screens, though you should be visually deducing this rank info. If I can make a calculation to compute this rank, I will add it to the final export (and maybe to the E/L Analysis screen), otherwise there's a column to record it yourself if you like. In this ranking, the most Early is rank +1, 2nd most Early is rank +2, etc. If all E/L values are >= 0 (RED), all ranks will be positive (e.g. for 5 horses, ranked +1 to +5). This might typically be the case for a 6f sprint. Similarly (e.g. typically for a route), all E/L values might be negative (Late): thus the most Late will be -1, 2nd most Late will be -2, etc. When there is a mix of Early and Late, say 3 Early and 2 Late, the 3 Earlies would show +1, +2, +3 and the 2 Lates would show -1, -2. By this means, when you look at your Spreadsheet model of Winners for, say 6f at PHA (just hypothetically), you may see an E/L range of positive values and also that the ranks are concentrated around +1 or +2, meaning the 2 most Early E/L values in a set. You may have to co-relate this to mutuels (and maybe race category, e.g. Maidens) to know whether this has any profitable predictive value, but hey, that's what a model is for...

Maybe there is a more useful way to show this E/L rank impact, and I'm open to others' thoughts (e.g. have 2 columns one for Early rank +1...+5, and the inverse for Late rank -1...-5 ??)

I would advise not to get hung up on modelling factors just because you can. I think there are only a few useful factors for any given track, distance (or distance group) and surface over any one short period of time, but knowing those in the short run may add the few hits you need per 20 races to put a 40-50% hit rate solidly in the black. I expect the universal factors (prominently: Total Energy, BLBL, V/DC, E/L, TPP) will be useful at all tracks, distances and surfaces, being overridden only by a race's matchup considerations.

Random musing: I would like to see the mix of Visual and Energy running styles (RS and ESP) in a summary for each race I model, so when scanning accumulated model info, I can be aware how many Earlies and how many Other than Early runners were in any given race (i.e. the matchup mix). Definitely for later...

I want to say that my future plans for RDSS modelling factors, mutuel impact values and tracking wagers are a bit more sophisticated than what this current simple data dump to an Excel file provides. But this export can be useful in the interim (since that interim will likely last a few more months yet...)

Looking forward to more comments!

Ted

[gandalf380]

upon further thought and reading your last posting, i retract my previous statement about rearranging the output. You are providing the output and it is up to us to use it in any way that we need to or want to. Therefore, the more info you provide in the export the better and the order does not matter.