Horse Racing Bettors Companion

Horse Racing Bettors Companion — Technical details

A closer description of how the research is put together. In Detail remains the shorter guide.

18

AnnieNEW

Annie

1. Purpose

Annie is a second reading of one race. Frank's RFMS rank stays on the card. Annie says where the stored evidence points, in five words, and whether Frank's top 3 should change.

The model is Grok 4.7, set to high effort. The questions put to it are fixed. From the pack below, it is asked for a rating for each declared runner, the reason for that rating, whether Frank's top 3 should change, an overview of the race, and an overview of each runner.

Three uses follow from that. The five ratings are a way to read Frank's selections again, so the profit taken from those selections can be improved by keeping or setting aside a horse on the strength of the evidence. The top-3 note is a check on whether those three should change. The same five ratings can be used without Frank: Bet finder can keep Strong, Good, Average, Poor or Weak and leave his rank out, and Today's winners can show the profit or loss of that set at a level stake, with 2% commission taken off winnings. That figure is the past result of the filter.

She does not place a bet, and she does not say a horse will win. The decision remains with the user.

2. The race pack

One pack is built for the race. Annie uses that pack and nothing else. Today's market price is not supplied. A BSP from a race already run is.

The race records the date and off time, the course, the race name, the distance in furlongs, the going, the surface, the race code, the class, the declared field size, whether it is a handicap, a novice or a maiden, and the prize where that is stored. Race code F is Flat, A is all-weather, H is hurdle, C is chase, and N is a bumper. The surface repeats that label.

Frank's RFMS rank and score are copied onto each runner. They are not rebuilt, and they are not rearranged so that Annie's rating agrees with them. Agreement is useful. Disagreement is allowed. A missing rank or score stays missing. It is not stored as zero.

3. Each runner

A runner records the horse, age, sex, weight, claim, official rating, headgear, and wind surgery. A non-runner stays in the pack and is marked as such.

The horse's own starts are included newest first, up to eight. A start records the date, course, race name, distance, going, surface, race code, class, field size, handicap or novice, weight, claim, official rating, headgear, jockey, trainer, finish, the share of the other runners beaten, the beaten distance, the BSP and the BSP rank. The stored form figure and speed figure for that start are included when they exist. A missing figure stays missing.

The comment, the early pace and the finishing style are the account of what the horse did in that start. A finish that was not completed is kept, and it is not treated as a completed run.

The jockey and the trainer are the names on today's card. The pack records whether today's trainer trained the supplied starts, and whether today's jockey rode in them, including wins and finishes of second or third with this horse. A stored record for the jockey, the trainer, and the partnership is included when one exists. Those records support the rating. They do not replace the horse's own starts.

The sire, the dam and the dam's sire are named. Sibling runs are added only when this horse has fewer than eight starts of its own, and only enough to make eight in total. A horse with eight starts has none. They are the most recent starts of horses with the same dam, newest first. The same sire as well is a full sibling. The same dam only is a half sibling. A sibling run is not this horse's run. It is a hint of inherited ability, not proof, and it is omitted when no sibling start is stored.

4. The rating

Each declared runner is given one band. Strong is the front of the field. Good sits between the front and the middle. Average is the middle. Poor sits between the middle and the back. Weak is the back. In a field of six, Strong is the first two, Average is the middle two, and Weak is the last two. In any other field, Strong stays the first two, Weak stays the last two, and the runners between them are Good, Average or Poor. Several runners may share a band.

The band is the part of the field the evidence points to more than any other part. It is not a claim that the horse will finish there. A horse that has never raced is not Weak for that reason. If the evidence does not point to one band more than the others, the band is Average and the commentary says so.

Confidence is separate: high, moderate or low. It describes how complete, relevant and consistent the evidence is. Strong with low confidence is allowed. Weak with high confidence is allowed. On the racecard the band is the Strength column.

5. Frank's top 3

For each of Frank's top three, the latest start is read against the starts before it. The result is that the latest start supports the horse, that the uncertainty is ordinary, that the latest start materially misrepresents the horse for today, or that the evidence is insufficient. A non-runner is insufficient.

A material contradiction is a latest start that seriously misrepresents the horse for today: the code, going or distance is not comparable, or the closer starts point somewhere else. A one-step difference in conditions is a concern. It is material when that latest start is the main reason the horse is near the top and the closer starts point to a different part of the field.

Each of the three is then left in, left in with a concern, or marked for a possible replacement. A replacement is another declared runner whose evidence points to the front, and it is named. A runner already in the middle of the field is not a replacement on its own. Every runner outside the three is marked as outside the three. The suggestion does not change Frank's rank.

6. The commentaries

The race commentary is one overview. It names who the evidence places at the front, who it places at the back, which runners outside the top three are interesting, where Frank and Annie agree, where they disagree, and whether a substitution is worth a further look. Where more than one runner is placed in the first two, it says they cannot all finish there. The same applies to the last two. It does not tell the reader what to bet.

The runner commentary is that horse only. It covers the recent runs, the opposition, the historical market in races already run, today's conditions, the trainer, the jockey, and the breeding. Strengths, concerns and unknowns are listed when the pack supports them. An empty list means there are none. The historical market compares a completed finish with the market position in that past race. It is not today's price, and beating that past market is not a betting claim.

On the racecard, Annie's Critique is the race note, and the information button beside a Strength word opens the runner note. Settings can show the words as dialogue or as bullets. The two say the same thing.

7. Packages

Bronze does not include Annie.

Silver opens whether she would change the top 3, and nothing else. No rating, no race commentary, and no runner commentary.

Gold opens the change to the top 3, naming who would be looked at instead, and a rating for every declared runner. Race commentary and runner commentary stay closed.

Diamond adds the race commentary and the runner commentary.

Bet finder can filter by the five ratings from Gold. Silver does not open those filters. Using the rating without Frank's rank is how a set can be counted on its own in Today's winners.

8. What is not claimed

The rating does not choose a bet and does not say that a horse will win. A missing figure stays missing. Frank's rank is not moved.

A profit or loss on Today's winners is the past result of the horses that matched. It is not a forecast.

The same account, in the guide, is the In Detail chapter.

An unhandled error has occurred. Reload 🗙