We are recruiting highly intelligent horseplayers! (a puzzle)

[Robert Fischer]

#Puzzle 33 = "EROCS_A_ROF_SDRAWKCAB_KROW"

... While this remains an unsolved puzzle, feel free to 'Reply' guesses below.

#CLUE='Inversion'

[Dave Schwartz]

Work backwards for a score.

(ChatGPT is my friend.)

Actual conversation:
________________
User
I was given the following as a test.

Do you understand it?
______
#Puzzle 33 = "EROCS_A_ROF_SDRAWKCAB_KROW"

... While this remains an unsolved puzzle, feel free to 'Reply' guesses below.

#CLUE='Inversion'
______
ChatGPT
Yes, it seems like a puzzle where you need to solve it using the clue provided. The clue "Inversion" typically means reversing or turning something upside down or inside out.

Given the puzzle string #Puzzle 33 = "EROCS_A_ROF_SDRAWKCAB_KROW", it seems like you might need to reverse the string to solve the puzzle.

Let's reverse the string:

Copy code
WORK_BACKWARDS_FOR_A_SCORE
So, the solved puzzle appears to be "WORK BACKWARDS FOR A SCORE."

[Jeff P]

Are you alluding to something from Bill Heller's Book "Overlay Overlay?"

"pu_mottob_eht_morf_secnamrofrep_tsap_daer"


-jp
.

[Robert Fischer]

Originally Posted by Dave Schwartz

Work backwards for a score.

(ChatGPT is my friend.)

Actual conversation:
________________
User
I was given the following as a test.

Do you understand it?
______
#Puzzle 33 = "EROCS_A_ROF_SDRAWKCAB_KROW"

... While this remains an unsolved puzzle, feel free to 'Reply' guesses below.

#CLUE='Inversion'
______
ChatGPT
Yes, it seems like a puzzle where you need to solve it using the clue provided. The clue "Inversion" typically means reversing or turning something upside down or inside out.

Given the puzzle string #Puzzle 33 = "EROCS_A_ROF_SDRAWKCAB_KROW", it seems like you might need to reverse the string to solve the puzzle.

Let's reverse the string:

Copy code
WORK_BACKWARDS_FOR_A_SCORE
So, the solved puzzle appears to be "WORK BACKWARDS FOR A SCORE."

Well done!
you will be receiving covert contact and further information

Seriously, impressive.
Also the use of a good tool.

Very important to use good tools. You seem to be good with ChatGPT.
Some players are impressed that they can handicap with "just the form" or less, but that can be a losing flaw.

ChatGPT, DRF Formulator, TimeformUS, video replay, etc.. etc...etc..

and back to your winning answer; Inversion. Work backwards for a score. Starting with good payout results and you work backwards to see if it fits. You start to get comfortable with Race Selection that provides possible scores, as well as some of the similar patterns that produce prices.

[Robert Fischer]

Originally Posted by Jeff P

Are you alluding to something from Bill Heller's Book "Overlay Overlay?"

"pu_mottob_eht_morf_secnamrofrep_tsap_daer"


-jp
.

that's a good one, also.

Good way to learn the history and sometimes the growth or regression of a horse.

[PaceAdvantage]

Cicada 3301?

YouTube

[Dave Schwartz]

Originally Posted by Jeff P

Are you alluding to something from Bill Heller's Book "Overlay Overlay?"

"pu_mottob_eht_morf_secnamrofrep_tsap_daer"


-jp
.

Child's play for ChatGPT.

_______
Here's some more words.
Might it be in a foreign language?

"pu_mottob_eht_morf_secnamrofrep_tsap_daer"
____________
ChatGPT
The string "pu_mottob_eht_morf_secnamrofrep_tsap_daer" seems to be written in reverse, similar to the previous puzzle. Reversing the string gives:

"read_past_performances_from_the_bottom_up"

It doesn't appear to be in a foreign language; it's in English, just written backwards. The phrase suggests that one should read past performances from the bottom upwards.

[Inner Dirt]

Originally Posted by Jeff P

Are you alluding to something from Bill Heller's Book "Overlay Overlay?"

"pu_mottob_eht_morf_secnamrofrep_tsap_daer"


-jp
.

I kid thee not, my dad did that. He wasn't the brightest bulb on the Christmas tree, and had went to Los Alamitos Harness with me like two times, previously. He had to go to Illinois for some training when he worked for Northrup. He cleaned up betting exactas, ridiculous amounts for what he was betting. He went to Quad City Downs, he showed me the programs with his bets. He made over $500 after expenses. My dad is not to be confused with my step dad who was a bright man who took me to Los Al QH when I was 14.

[Jeff P]

Originally Posted by Dave Schwartz

Child's play for ChatGPT.

_______
Here's some more words.
Might it be in a foreign language?

"pu_mottob_eht_morf_secnamrofrep_tsap_daer"
____________
ChatGPT
The string "pu_mottob_eht_morf_secnamrofrep_tsap_daer" seems to be written in reverse, similar to the previous puzzle. Reversing the string gives:

"read_past_performances_from_the_bottom_up"

It doesn't appear to be in a foreign language; it's in English, just written backwards. The phrase suggests that one should read past performances from the bottom upwards.

About 6 seconds after seeing the original string I knew (without outside help) it said Work Backwards For A Score.

It took me an extra second or two because of the underscore characters.

This thread implies a fascinating question.

Is there a place for machine learning?

Of course there is. (Obviously.)

But is there also a place for intuition?

I would argue both questions share the same answer.


-jp
.

[Dave Schwartz]

Originally Posted by Jeff P

About 6 seconds after seeing the original string I knew (without outside help) it said Work Backwards For A Score.

It took me an extra second or two because of the underscore characters.

This thread implies a fascinating question.

Is there a place for machine learning?

Of course there is. (Obviously.)

But is there also a place for intuition?

I would argue both questions share the same answer.


-jp
.

This is a very excellent question.

In the past 10 months or so I've done a lot of work learning how ChatGPT works. As everyone says, "It's all about the prompts."

That is, of course, true.

However, CG has a lot of preconceived ideas about horse racing - many (most?) of which are not statistically sound.

This relates to your question in the sense that we have to ask ourselves, "What drives our intuition?"


I did give GPT some horse racing data to look at. Some Handicapping Objects from HSH. The columns were labeled with things like "Odds, Jockey, Trainer, Speed Ratings, Pace Ratings, etc."

It made selections based mostly upon Odds, Jockey, and Pace.

Started a new session and removed the Odds. It switched its selections to the jockey column.

When I started still another new session, I switched the column names around without changing the data.

It changed its selections to match "jockey."

The point being that it was not handicapping the actual data, but the underlying knowledge it believes to be true: jockey is a very important factor.

MY POINT:
I think that the drive behind intuitional ideas is our core beliefs, and combinations thereof. We will also defend those beliefs to the point of resisting looking at real data that might fracture our belief system.


Your thoughts?

[Bustin Stones]

In any test of data, the key is what question is posed and how is it posed.
So, in using any GPT, the burden continues to be on the user. The best way to develop one's skill set is to answer the original question on one's own without using GPT. GPT is only a tool that will be best used by people who are interested in acquiring the right skills to use it better.

[Robert Fischer]

Originally Posted by PaceAdvantage

Cicada 3301?

https://www.youtube.com/watch?v=3ZfKfkuYwyM

I'm very interested in watching this today. I've noticed you have a good eye for some quality, and hard-hitting video content.

You caught on, that, at least some of the inspiration of my post was credited to the Cicada 3301 type puzzles. (I half hope your video provides a solution or description of a puzzle or two, - and half hope your video claims to be unable to solve some or all of the puzzles. The difficulty and mystery was some of the charm.)

The puzzle I presented here was meant to be more in fun, and in a fresh introductory fashion, than it was actually meant to be some perplexing problem, unsolvable w/out sufficient genius or invested time.

It also genuinely was meant to present 'inversion' as one method of handicapping. Useful, as well as sometimes an alternative or possibly positive diversion from what can get repetitive and even stale among the handicappers who are churning out their work and/or have lost some interest and imagination. No panacea, but a useful model in the toolbox.

Genuinely enjoying the path this thread has taken.

[Robert Fischer]

Originally Posted by PaceAdvantage

Cicada 3301?

https://www.youtube.com/watch?v=3ZfKfkuYwyM

good video. A channel I may browse sometime. thx

[thaskalos]

I read that ChatGPT continually gets better informed, and its future responses get better and better. This "self-training" of the chatbox is done with the use of human trainers who upvote or downvote the responses by clicking on thumbs-up or thumbs-down icons. When we are talking about a specialized topic like horse handicapping...how can we be sure that these "human trainers" are better informed about the handicapping process than the rest of us are?

[Dave Schwartz]

Originally Posted by thaskalos

I read that ChatGPT continually gets better informed, and its future responses get better and better. This "self-training" of the chatbox is done with the use of human trainers who upvote or downvote the responses by clicking on thumbs-up or thumbs-down icons. When we are talking about a specialized topic like horse handicapping...how can we be sure that these "human trainers" are better informed about the handicapping process than the rest of us are?

Those thumbs up & down are done by the USERS, although I've never used them myself.

Instead, verbal compliments and criticisms are typically used.

My take on its learning process is that it learns from what it reads EVERYWHERE. It is, after all, a language model.

Thus, it believes that whatever is said most often must be true.

These become "facts," to be regurgitated whenever a question is asked.

Thus, its horse racing answers are derived from the very most common source of horse players: people who show up on the big race days.

A private session with the top 200 posters in the handicapping section of PA would improve the answers tremendously.