FIG2 Patterns and Robot

[km]

I want to elicit some feedback for how to provide functionality into the Robot for testing and extracting various generic fig number patterns (from FIG2 screen)

Here's my plan, nothing has been worked on yet except on paper, so feel free to comment.

Rather than contrive various patterns that you may have, i want to provide the tools to let you work on them independently. The Robot could have three filter sections with each representing the horse's last three races (and flowing from the last four starts).

Each would have the following filters to select

T the horse ran a cyclical or lifetime top in this start
+ the horse improved its figure from the previous sstart
- the horse declined its fig from the previous start
= the horse ran the same, or "paired" it figure from the previous start (within 1.5 points)
@ the horses figure dropped way below its previous by 7 or more points (bounce)
N neutral, anything is ok or no race

I'm using 7 point drop as the "bounce" but maybe some of you think it should be 5 pts or 10 pts ?

Example: (+) -> (T) -> (@)
i set my (3-back) filter for "+" the horse improved its figure from the 4th back to the 3rd back race
i set my (2-back) filter for "T" (ran its lifetime or cyclical Top 2-back)
i set my (1-back) filter for "@" (bounced badly last outing )

I'm looking for "bounce back" horses here - or is that a myth? 3yr are most interesting to discover.

How do these horses perform today? The Robot will know in minutes.

Its generic - yes - but nuance can be added with other Robot filters, such as Age/Sex/Surface and Class type; the permutations are endless and the spot play potential is huge, especially with the Connections coming soon, you can test Trainers for these patterns. Are $$ a bad bet if the horse ran a Top or Bounced last time, how do K=110s do in those cases? etc. etc. Now you'll know.

Certain patterns will produce positive sustainable ROI because so few handicappers can recognize them without a lot of effort, scrutiny and expense.

Let me know what you think, don't over complicate it please.

note to Export and db users, you already have this in a sense with HX5 and the Fig numbers for all PP lines (Field 111-120). However, it would take you about 100 years to program and query all these pattern combinations that the Robot interface makes user friendly, not to mention combining them with 135 separate items in the Learn All report. Let the Robot do the grunt work on this one.

[Mike Brown]

Ken, nothing to add at this time, but I absolutely LOVE the concept ! Would your pattern symbols be added to the FIG2 report ?

Thanks
Mike

[del_dog]

Ken


What about a pattern that involves running (+ - ) to the race pace-per par using the same method you outlined below?

[DanG]

Off the top of my head and very random as they occur to my out too late head… :eek:

• The ‘amount of the top and the size of the figure is significant. The larger the top; the more dramatic the reaction depending on age etc. (in theory)

• Grouping of races was a concept taught by Friedman in sheet reading when I was a kid. If they have at least 6 starts group in 3’s to get a mock trend line. 9 start runners for example…groups of 3 each with ascending tops and or bottoms / descending or neutral. Averages can be very misleading when grouping figures.

• If the PER is used as the base it can be tricky (imo) with turf and some poly route runners. At times descending pace #’s are actually a positive in grass runners while learning to relax and distribute their EEN% correctly and the PER does have a pace variable.

• The same theory you’re describing Ken can be interesting with PAC #’s. If users limit their study to early types patterns in pace #’s it can be very revealing. Normally a healthy sign when a lay-off early type shows major early gas in their first start back.

As you said; putting these tools in the Robot2 will yield many angles we couldn’t possibly think of now. Great idea and look forward to seeing it in action; thanks! :cool:

[Donnie]

with feedback from others always accepted.....

I personally set that bounce line (last race a stinker) at 10 points. But I don't expect a bounce today. I do a toss-out. Seems they do not rebound as much as those with a lesser drop in their last race. Maybe the trainer used the last race as a conditioner, or they used the horse to fill a race with no intentions of even trying (with smaller drops of pre-race to last). But when a horse dropped 10 points off it's previous, something may be wrong with the horse.

I would bet when she comes back, Indian Blessing may show a 10 point drop from her previous race to yesterday's. The Dubai Bounce tho must be anticipated immediately after their return. If they give her time to rest, she may be one of those exceptions that come off that 10 point drop just fine. But in lower levels, not so much.

Love the idea of being able to test trainers with these patterns.

Gambler844 utilizes some of these same thought processes in his reports using flags to note some of the same patterns you mentioned above. I would wager he would love the ability to test them against ongoing data!

[Donnie]

One point, and this was stressed last year at the seminar:

The Figs2 has the PAC numbers built into them. So I cautioned people to make sure you keep that top of mind when you interpret the numbers. While similar to Ragizon/ThoroughGraph, they are not always compatible. The faster the pace of the horse, the more likely the horse earns a better (lower) Figs2. This is why I had asked Ken to put the FR3 to the right of the graph....easy to see who had more gas in the tank towards the end of the race. Also, easy to see who consistently has more gas at the end of their races.

[Huguenot]

Ken,
Great idea

IMHO this whole discussion constitutes a very slippery slope. Ask 20 people about the bounce and you'll get about 40 different answers. Even the Raggies probably argue amongst themselves as to how many points does it take to make a horse bounce.

then you have the 3 yr old "explosive" move whiich has worked well in the Belmont (at least with HTR figs) whereby a horse barely exceeds his prior top. Horses who get a little rest after that are known to run figures that are off the charts. Sometimes off a small setback, sometimes not.

A good example of an "explosive" move came at CD on Saturday in the 3rd race with the winner, the 5 horse, Kiddari. A 31 to a 20 to a 19 and 6 weeks rest with regular workouts.

I disagree with Donnie on using the Fig 2 figures with the PAC, etc I dont see any reason you can't incorporate the pace a horse ran into his chart. After all it's about the EFFORT, the energy used The faster the horse ran early the more effort he used and the more likely he will regress If that's reflected in a higher FINAL velocity figure that's even better in my opinion.

I'd also consult the best book I've ever read about form cycles and patterns. Cary Fotias' Blinkers Off.


Of course you cant directly translate his figures to HTR -- different scales for different distances, etc., but the theory is sound. And if there is any way to test his theories using HTR figs that's the road i would go down -- of course i realize there will be programming limitations (For Fotias at 4f, 1 point is 1 length. At 6f, 1 pt is 1.5 lengths and it progresses from there. Those are rough measurements. )

---------------------------

Finally, to me a chart is more than just a top or a bounce but the total "look." Is it smooth or jagged? has the horse thrown in unaccountable random races -- good or bad -- or are there patterns generally involved. Is a horse cycling back to earlier good efforts?

You want to see a pattern? Check out the 4 horse Charlie Thumper at CD 3rd race Saturday. 21 22 21 20 22 21 The 2B horse, Headache, the winner was also an explosive horse

And remember these are just probabilities. They are wrong a lot.

[Donnie]

I disagree with Donnie on using the Fig 2 figures with the PAC, etc I dont see any reason you can't incorporate the pace a horse ran into his chart.

Don't know if I was misunderstood.....

The PAC is already a part of the number you are looking at when you look at the chart in FIGs2. I had Ken add the PAC-PER figures to the left of the chart as well. I consult them on a regular basis also.

[km]

Fig2 sheet style numbers are final time only - simply translated from Cramer figure

PER is similar to A/P and includes the pace quantity

I'm getting the feeling the Andy (Hugo) is right - there is too much variety of opinion on every aspect of the "pattern" - and don't forget, i can only provide the last few races, not the complete picture. I'm rethinking this... it has become over-complicated already.

Why do speed figures users love to fight it out so much? It's like arguing over ice cream flavors, everyone can have their own preference without anyone being "wrong".

I think i may instead provide a few basic last race filters, well defined, such as Top and Bounce on the Robot and leave it at that. I'll run some tests on my own and find out where the sweet spot is for those and put them in. Down the road we may evolve it further.

thank you

[DanG]

I'm rethinking this... it has become over-complicated already.

Great…now look at the fine mess we’ve made…:D
http://3.bp.blogspot.com/_EmJorwXmPBY/SMi9bA1AitI/AAAAAAAAAF0/hRMRz0aewcw/s400/laurel_hardy.jpg

[Huguenot]

Fig2 sheet style numbers are final time only - simply translated from Cramer figure

PER is similar to A/P and includes the pace quantity

I'm getting the feeling the Andy (Hugo) is right - there is too much variety of opinion on every aspect of the "pattern" - and don't forget, i can only provide the last few races, not the complete picture. I'm rethinking this... it has become over-complicated already.

Why do speed figures users love to fight it out so much? It's like arguing over ice cream flavors, everyone can have their own preference without anyone being "wrong".

I think i may instead provide a few basic last race filters, well defined, such as Top and Bounce on the Robot and leave it at that. I'll run some tests on my own and find out where the sweet spot is for those and put them in. Down the road we may evolve it further.

thank you

You had a "top" on the Robot before -- for the PAC and PER i believe -- and removed them. I trust because users didn't find them helpful.

For me I'd like to know how a horse does if he runs a top from his last 10 races and then returns within three weeks. And i'd to be able to quantify that top. A one-point top makes a much different statement than a 10-point top. So lets say you have a filter in the Robot for the extent of the top (how much the best was his last out of the last 10 races.) If you could use that filter in accordance with a days away filter you might have something worth looking at

Secondly, there are those sudden improvements from a previous race Since I tend to follow Fotias' guidelines, I get concerned when a horse improves about six to eight PER points or more from his previous race. (Very rough guidelines here), especially if he also ran a new pace top in that race. And especially if he's coming back without much rest. So you'd need a filter for last race improvement over his prior race (assuming there was no layoff in between to interrupt the form cycle), again combined with a days off filter. And an option to denote if the last race was also a new pace top.

Horses win all the time even of these tops but how much do they affect the favorite's win rate/ROI?


And turf racing is a completely different animal, according to Fotias. While on dirt you want to see a pace figure which is a top with a non-top final number, on the grass you want a much LOWER pace figure with a final figure within a few points of the previous race

[km]

I'm going to replant the PAC/FIG2 TOPs for you Andy, proabably 2-pointers (ran a top 2points or more higher than any previous start)

But I also have another idea for a filter

= Projects to improve and/or run a new top today

This is one of those items that won't have a clear algorithm to share. But I have some strong feedback on how to formulate it from an experience sheet player in NY. So it will be an accumulation of various ideas, somewhat generic, but your ability to work with it depends on your creativity. 3yr Fillies 2TS might prove hugely profitable, but 5y males in Stakes may be terrible. It might be fantasitc with some trainers, and pitiful with others. It could be a winner at SAR and a bust at DMR. The concept is sound, but you have to do your homework.

[Judicious Player]

Ken,

I am having success with the following:

(1) First time the horses's 2nd figure has been higher than the first: e.g., 98 102. An especially good angle for a turf race, as the horse is learning to conserve its energy for the latter part of the race.

(1) New lifetime pace top: wins can result off such an effort in either the very next race, or the 2nd or 3rd race after the new lifetime pace top has been acheived -- often at a price.

Steve

[Judicious Player]

Ken,
Great idea

IMHO this whole discussion constitutes a very slippery slope. Ask 20 people about the bounce and you'll get about 40 different answers. Even the Raggies probably argue amongst themselves as to how many points does it take to make a horse bounce.


Ken,

I look at the "bounce" this way: if the horse is going at, say, 8/5 (i.e., an underlay), it is going to bounce off a new top or a much-improved top from its last race; if the horse is going off at 6-1 or higher (i.e., an overlay), it is not going to bounce. Let the odds dictate the possibility of a bounce -- or not.

Steve

[tomcat]

Part of the problems dealing with these numbers and the 2d race (not 3d) at CD last Sat is, we assume the horse raced hard the whole time. However, it is hard to use the time of any horse who was slowed at the end of the race for whatever reason and in the CD race the horse that won big (+3 or more). In that latter case Kidari won easily. But how fast would have run if pressed? Probably faster. Maybe a 17 or 18 instead of the 19. The racing sec'y made him the ML favorite but he seemed to ignore by everyone else and pd a nice $10. Just using numbers w/out understanding the conditions that created them can lead us in the wrong direction.

[tomcat]

Ken, since I tend to keep some track files is separate folders...like CD or Kee, I would like to keep
my spot plays when I change files(folders). But I am finding that spot plays are not the same when changing folders. Some of the spot plays are lost. I suppose I'm doing something wrong?

[Rick]

The spot plays are saved in the directory the races are in.

You can copy and paste them to your other directories.

[km]

Tom, as Rick said, your spot play files are saved to the current work folder. When you change folders, you can't access them unless you copy them over to the new folder.

The spot plays are easy to find because they have the SPOT--- as the file name. Copy them over to the other folders if you want to use.

I suppose i could save them in a common bin folder, but then you would not be able to have separate spot plays in different folders or hard drives. That would limit the flexibility.

Are you coming to the seminar? We can show you how to handle this quickly.

[Rick]

And the Robot II spot plays are named from 001.HSP to 999. HSP.

[tomcat]

I would love to come to the seminar since I started with you a gazillion yrs ago, but age
restrictions are making it difficult. I will be supporting you in spirit if I don't get there.

I made the spot play adjustments, thank you.