{ "id": 355204, "title": "Demo of One Race at Saratoga.", "url": "/competitions/big-data-derby-2022/discussion/355204", "author_name": "", "post_date": "2022-09-25T23:06:00.253581100Z", "votes": 4, "comment_count": 4, "views": 0, "content": "

Hello All,

\n

Hope you all will like my work.

\n

Basically, I have analyzed one Race that happened at Saratoga on July 27, 2019. The actual video of the race can be found here.

\n

Question: Why did Javier Castellano lose to Joel Rosario when he had a very good chance of winning?
\nAnswer: I would first emphasize that the following is just a sample visualization. Although, I have already made a more sophisticated animation and will release it before the competition's deadline.

\n

\"\"

\n

In the visualization above, I am comparing and contrasting the race-tracking data of Javier Castellano and Joel Rosario in order to understand what led Joel Rosario to win and what led Javier Castellano to lose. Here, I have already plotted Joel Rosario's race-tracking data just for clarity and comparing purposes.

\n

\"\"

\n
Little In-Depth Insight
\n

Typical Time Speed Curve

\n

\"\"

\n

1 Joel Rosario Speed Time Curve

\n

\"\"

\n

2 Javier Castellano Speed Time Curve

\n

\"\"

\n\n

Also. I've used Stats and Physics to assess the Optimal Path for each race type and its category, defined each Jockey's Performance metric, created a Machine Learning Model for Drafting Strategy, and much more. Moreover, I would really appreciate comments and feedback on this. Thank you.

\n

Visit this Notebook to see my work.

", "messages": [ { "id": "1955312", "postDate": "09/25/2022 23:06:00", "content": "

Hello All,

\n

Hope you all will like my work.

\n

Basically, I have analyzed one Race that happened at Saratoga on July 27, 2019. The actual video of the race can be found here.

\n

Question: Why did Javier Castellano lose to Joel Rosario when he had a very good chance of winning?
\nAnswer: I would first emphasize that the following is just a sample visualization. Although, I have already made a more sophisticated animation and will release it before the competition's deadline.

\n

\"\"

\n

In the visualization above, I am comparing and contrasting the race-tracking data of Javier Castellano and Joel Rosario in order to understand what led Joel Rosario to win and what led Javier Castellano to lose. Here, I have already plotted Joel Rosario's race-tracking data just for clarity and comparing purposes.

\n

\"\"

\n
Little In-Depth Insight
\n

Typical Time Speed Curve

\n

\"\"

\n

1 Joel Rosario Speed Time Curve

\n

\"\"

\n

2 Javier Castellano Speed Time Curve

\n

\"\"

\n\n

Also. I've used Stats and Physics to assess the Optimal Path for each race type and its category, defined each Jockey's Performance metric, created a Machine Learning Model for Drafting Strategy, and much more. Moreover, I would really appreciate comments and feedback on this. Thank you.

\n

Visit this Notebook to see my work.

", "rawMarkdown": "Hello All,\n\nHope you all will like my work.\n\nBasically, I have analyzed one Race that happened at Saratoga on July 27, 2019. The actual video of the race can be found [here](https://youtu.be/-OId_SqGt_I?t=2141). \n\n**Question:** Why did **Javier Castellano lose** to **Joel Rosario** when he had a very good chance of winning?\n**Answer:** I would first emphasize that the following is just a sample visualization. Although, I have already made a more sophisticated animation and will release it before the competition's deadline. \n\n![](https://i.postimg.cc/nVv03rmH/jvj.gif)\n\nIn the visualization above, I am **comparing** and **contrasting** the **race-tracking data** of Javier Castellano and Joel Rosario in order to understand what led Joel Rosario to win and what led Javier Castellano to lose. Here, I have already plotted Joel Rosario's race-tracking data just for clarity and comparing purposes.\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F8412510%2Fbf413bd197b67134b1e6166e154b24a0%2FSummary.png?generation=1664143904781760&alt=media)\n\n##### Little In-Depth Insight\n\n**Typical Time Speed Curve**\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F8412510%2F8aba1981983c84423c7579bcb340233e%2FSpeed-time-curve.jpeg?generation=1665796320262949&alt=media)\n\n1 **Joel Rosario Speed Time Curve**\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F8412510%2F0fcc90339d1cdc856a7eeed6ee149bcc%2Fjoel_rosario_sar.png?generation=1666450429793585&alt=media)\n\n2 **Javier Castellano Speed Time Curve**\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F8412510%2F423e9a1bde8e2f6a4d584349f9e959a2%2Fjavier_castellano_sar.png?generation=1666450416978191&alt=media)\n\n- If we compare the **above visualizations** with the **typical speed time curve**, we can see both **Joel Rosario** and **Javier Castellano** were running **head to head** at given time intervals. **However,** Javier Castellano had the **lead at 10 -15 sec intervals** **(** could be seen in **2nd visualization** **)** **but** it was a **smart move** by **Joel Rosario** by **dropping little speed just before** the **braking point** and then by took a **wider turn** and accelerated **(** it could be seen with a **blue dot** in **1st visualization** at **20-25 sec interval** **)** by charging the horse and **overtakes** Javier Castellano **between 25-35 sec intervals** where he **apex late** and **blocked** the **path for Javier Castellano** **(** it could be seen in **2nd visualization** at **30-35 sec intervals** where there was a **sudden drop in speed** **)**. Then both of them forced the horse to run till the finish line.\n\nAlso. I've used **Stats** and **Physics** to assess the **Optimal Path** for each race type and its category, defined each **Jockey's Performance** metric, created a **Machine Learning Model for Drafting Strategy**, and **much more**. Moreover, I would really appreciate comments and feedback on this. Thank you.\n\n#### Visit this [Notebook](https://www.kaggle.com/code/kushtrivedi14728/advance-eda-with-ml-big-data-derby) to see my work.", "votes": null }, { "id": "1956663", "postDate": "09/26/2022 14:47:46", "content": "

Hello Kush:

\n

Here's what would be the conventional interpretation of what happened in that race--

\n

1) Castellano and Rosario are dueling on the lead early in the race--horses can expend too much energy to early competing to be the lead horse early.

\n

2) The conventional wisdom in racing is that it is more advantangeous to pass from the outside of the leading horse than the inside even though the inside would obviously be a shorter distance. The horse on the inside is up against the rail and risks being subtly, or not so subtly, \"cut-off\" by the leading horse.

\n

3) So Castellano's logic is I can't beat Rosario's horse in a pure speed duel, I'm going to drop back a little then try and pass him on the outside in the final portion of the race. It didn't work out, but the strategy was sound given how the race \"set up\".

\n

--Brian

", "rawMarkdown": "Hello Kush:\n\nHere's what would be the conventional interpretation of what happened in that race--\n\n1) Castellano and Rosario are dueling on the lead early in the race--horses can expend too much energy to early competing to be the lead horse early.\n\n2) The conventional wisdom in racing is that it is more advantangeous to pass from the *outside* of the leading horse than the inside even though the inside would obviously be a shorter distance. The horse on the inside is up against the rail and risks being subtly, or not so subtly, \"cut-off\" by the leading horse.\n\n3) So Castellano's logic is I can't beat Rosario's horse in a pure speed duel, I'm going to drop back a little then try and pass him on the outside in the final portion of the race. It didn't work out, but the strategy was sound given how the race \"set up\".\n\n--Brian", "votes": null }, { "id": "1956861", "postDate": "09/26/2022 16:21:24", "content": "

Hello @btimoney,

\n

Firstly I really appreciate your response.

\n

Indeed you are true about Castellano's logic. I have also found that logic and he had actually tried that same way by dropping back a little and trying to pass him afterward. However, this is just a basic sample of my work I have already made a sophisticated version for that and will release it before the competition deadline.

\n

Also, Castellano was fortunate to finish second since Manuel Franco would have won \"Place\" if his horse hadn't been exhausted as he was charging the horse too much before the braking point and Castellano was clever and saved his horse's energy before catching up to Franco then by taking a slightly wider turn at the exit point to get himself little room and then charged the horse. Manuel also charged his horse but the horse was already been exhausted which could be seen here.

\n

Side Note: Franco also took a big wider turn and Apexing late than Rosario. However, he had made 2 main mistakes. First not having a good head start in comparison to Rosario and Castellano, and second, charging a horse before or at a breaking point to get lead. Also, an interesting thing to note is that Tyler stumbled at the start of the race, but he still managed to finish in fourth place by taking the shorter route and charging the horse after the exit point.

\n

In short words:

\n
    \n

  1. Rosario: Had good head start and charged the horse at the correct time and angle following the standard racing convention. (Wide turn and Late Apex)

    2. \n

  2. Castellano: Also had a good head start, however, he made a mistake by coming towards the rails at the braking point as it was not necessary he was already on the right of Rosario and then he thought to step back a little as Franco was also neck to neck with him and was lucky that Franco's horse was exhausted.

    3. \n
\n

Key Findings for this particular race: (Should not Consider this as Main Finding)

\n

If you stumbled upon the start always try to aim for a shorter distance instead of a fast route and if you get off to a good start, stick to the traditional racing line rather than a geometric one.

\n

Lastly, thank you for your response I really appreciate it.

", "rawMarkdown": "Hello @btimoney,\n\nFirstly I really appreciate your response.\n\n**Indeed you are true** about **Castellano's logic**. I have also found that logic and he had actually tried that same way by dropping back a little and trying to pass him afterward. However, this is just a basic sample of my work I have already made a sophisticated version for that and will release it before the competition deadline.\n\nAlso, **Castellano** was fortunate to finish second since **Manuel Franco** would have won \"**Place**\" if his horse hadn't been exhausted as he was charging the horse too much before the braking point and **Castellano** was clever and saved his horse's energy before catching up to **Franco** then by taking a slightly wider turn at the exit point to get himself little room and then charged the horse. Manuel also charged his horse but the horse was already been exhausted which could be seen [here](https://youtu.be/-OId_SqGt_I?t=2200).\n\n**Side Note:** **Franco** also took a big wider turn and Apexing late than **Rosario**. However, he had made **2 main mistakes**. **First** not having a good head start in comparison to Rosario and Castellano, and **second**, charging a horse before or at a breaking point to get lead. Also, an interesting thing to note is that **Tyler** stumbled at the start of the race, but he still managed to finish in fourth place by taking the shorter route and charging the horse after the exit point.\n\n#### **In short words**:\n\n1. **Rosario:** Had good head start and charged the horse at the correct time and angle following the standard racing convention. (Wide turn and Late Apex)\n\n2. **Castellano:** Also had a good head start, however, he made a mistake by coming towards the rails at the braking point as it was not necessary he was already on the right of Rosario and then he thought to step back a little as Franco was also neck to neck with him and was lucky that Franco's horse was exhausted.\n\n#### **Key Findings for this particular race:** (Should not Consider this as Main Finding)\n\nIf you stumbled upon the start always try to aim for a shorter distance instead of a fast route and if you get off to a good start, stick to the traditional racing line rather than a geometric one.\n\nLastly, thank you for your response I really appreciate it.", "votes": null }, { "id": "1957289", "postDate": "09/26/2022 21:01:47", "content": "

Kush:

\n

Great stuff. Focusing on velocity/acceleration metrics going in and coming out of turns is a very good use of this dataset.

", "rawMarkdown": "Kush:\n\nGreat stuff. Focusing on velocity/acceleration metrics going in and coming out of turns is a very good use of this dataset.", "votes": null }, { "id": "1957353", "postDate": "09/26/2022 22:56:06", "content": "

Correct! also we have to find the minimum curvature probabilistic inference as that could help to improve lap time as well as increment in speed.

", "rawMarkdown": "Correct! also we have to find the minimum curvature probabilistic inference as that could help to improve lap time as well as increment in speed.", "votes": null } ], "comments": [ { "id": 1956663, "author_name": "btimoney", "author_url": "", "post_date": "09/26/2022 14:47:46", "content": "

Hello Kush:

\n

Here's what would be the conventional interpretation of what happened in that race--

\n

1) Castellano and Rosario are dueling on the lead early in the race--horses can expend too much energy to early competing to be the lead horse early.

\n

2) The conventional wisdom in racing is that it is more advantangeous to pass from the outside of the leading horse than the inside even though the inside would obviously be a shorter distance. The horse on the inside is up against the rail and risks being subtly, or not so subtly, \"cut-off\" by the leading horse.

\n

3) So Castellano's logic is I can't beat Rosario's horse in a pure speed duel, I'm going to drop back a little then try and pass him on the outside in the final portion of the race. It didn't work out, but the strategy was sound given how the race \"set up\".

\n

--Brian

", "votes": null, "replies": [ { "id": 1956861, "author_name": "kushtrivedi14728", "author_url": "", "post_date": "09/26/2022 16:21:24", "content": "

Hello @btimoney,

\n

Firstly I really appreciate your response.

\n

Indeed you are true about Castellano's logic. I have also found that logic and he had actually tried that same way by dropping back a little and trying to pass him afterward. However, this is just a basic sample of my work I have already made a sophisticated version for that and will release it before the competition deadline.

\n

Also, Castellano was fortunate to finish second since Manuel Franco would have won \"Place\" if his horse hadn't been exhausted as he was charging the horse too much before the braking point and Castellano was clever and saved his horse's energy before catching up to Franco then by taking a slightly wider turn at the exit point to get himself little room and then charged the horse. Manuel also charged his horse but the horse was already been exhausted which could be seen here.

\n

Side Note: Franco also took a big wider turn and Apexing late than Rosario. However, he had made 2 main mistakes. First not having a good head start in comparison to Rosario and Castellano, and second, charging a horse before or at a breaking point to get lead. Also, an interesting thing to note is that Tyler stumbled at the start of the race, but he still managed to finish in fourth place by taking the shorter route and charging the horse after the exit point.

\n

In short words:

\n
    \n

  1. Rosario: Had good head start and charged the horse at the correct time and angle following the standard racing convention. (Wide turn and Late Apex)

    2. \n

  2. Castellano: Also had a good head start, however, he made a mistake by coming towards the rails at the braking point as it was not necessary he was already on the right of Rosario and then he thought to step back a little as Franco was also neck to neck with him and was lucky that Franco's horse was exhausted.

    3. \n
\n

Key Findings for this particular race: (Should not Consider this as Main Finding)

\n

If you stumbled upon the start always try to aim for a shorter distance instead of a fast route and if you get off to a good start, stick to the traditional racing line rather than a geometric one.

\n

Lastly, thank you for your response I really appreciate it.

", "votes": null, "replies": [] } ] }, { "id": 1957289, "author_name": "btimoney", "author_url": "", "post_date": "09/26/2022 21:01:47", "content": "

Kush:

\n

Great stuff. Focusing on velocity/acceleration metrics going in and coming out of turns is a very good use of this dataset.

", "votes": null, "replies": [ { "id": 1957353, "author_name": "kushtrivedi14728", "author_url": "", "post_date": "09/26/2022 22:56:06", "content": "

Correct! also we have to find the minimum curvature probabilistic inference as that could help to improve lap time as well as increment in speed.

", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "1955312": "Hello All,\n\nHope you all will like my work.\n\nBasically, I have analyzed one Race that happened at Saratoga on July 27, 2019. The actual video of the race can be found [here](https://youtu.be/-OId_SqGt_I?t=2141). \n\n**Question:** Why did **Javier Castellano lose** to **Joel Rosario** when he had a very good chance of winning?\n**Answer:** I would first emphasize that the following is just a sample visualization. Although, I have already made a more sophisticated animation and will release it before the competition's deadline. \n\n![](https://i.postimg.cc/nVv03rmH/jvj.gif)\n\nIn the visualization above, I am **comparing** and **contrasting** the **race-tracking data** of Javier Castellano and Joel Rosario in order to understand what led Joel Rosario to win and what led Javier Castellano to lose. Here, I have already plotted Joel Rosario's race-tracking data just for clarity and comparing purposes.\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F8412510%2Fbf413bd197b67134b1e6166e154b24a0%2FSummary.png?generation=1664143904781760&alt=media)\n\n##### Little In-Depth Insight\n\n**Typical Time Speed Curve**\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F8412510%2F8aba1981983c84423c7579bcb340233e%2FSpeed-time-curve.jpeg?generation=1665796320262949&alt=media)\n\n1 **Joel Rosario Speed Time Curve**\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F8412510%2F0fcc90339d1cdc856a7eeed6ee149bcc%2Fjoel_rosario_sar.png?generation=1666450429793585&alt=media)\n\n2 **Javier Castellano Speed Time Curve**\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F8412510%2F423e9a1bde8e2f6a4d584349f9e959a2%2Fjavier_castellano_sar.png?generation=1666450416978191&alt=media)\n\n- If we compare the **above visualizations** with the **typical speed time curve**, we can see both **Joel Rosario** and **Javier Castellano** were running **head to head** at given time intervals. **However,** Javier Castellano had the **lead at 10 -15 sec intervals** **(** could be seen in **2nd visualization** **)** **but** it was a **smart move** by **Joel Rosario** by **dropping little speed just before** the **braking point** and then by took a **wider turn** and accelerated **(** it could be seen with a **blue dot** in **1st visualization** at **20-25 sec interval** **)** by charging the horse and **overtakes** Javier Castellano **between 25-35 sec intervals** where he **apex late** and **blocked** the **path for Javier Castellano** **(** it could be seen in **2nd visualization** at **30-35 sec intervals** where there was a **sudden drop in speed** **)**. Then both of them forced the horse to run till the finish line.\n\nAlso. I've used **Stats** and **Physics** to assess the **Optimal Path** for each race type and its category, defined each **Jockey's Performance** metric, created a **Machine Learning Model for Drafting Strategy**, and **much more**. Moreover, I would really appreciate comments and feedback on this. Thank you.\n\n#### Visit this [Notebook](https://www.kaggle.com/code/kushtrivedi14728/advance-eda-with-ml-big-data-derby) to see my work.", "1956663": "Hello Kush:\n\nHere's what would be the conventional interpretation of what happened in that race--\n\n1) Castellano and Rosario are dueling on the lead early in the race--horses can expend too much energy to early competing to be the lead horse early.\n\n2) The conventional wisdom in racing is that it is more advantangeous to pass from the *outside* of the leading horse than the inside even though the inside would obviously be a shorter distance. The horse on the inside is up against the rail and risks being subtly, or not so subtly, \"cut-off\" by the leading horse.\n\n3) So Castellano's logic is I can't beat Rosario's horse in a pure speed duel, I'm going to drop back a little then try and pass him on the outside in the final portion of the race. It didn't work out, but the strategy was sound given how the race \"set up\".\n\n--Brian", "1956861": "Hello @btimoney,\n\nFirstly I really appreciate your response.\n\n**Indeed you are true** about **Castellano's logic**. I have also found that logic and he had actually tried that same way by dropping back a little and trying to pass him afterward. However, this is just a basic sample of my work I have already made a sophisticated version for that and will release it before the competition deadline.\n\nAlso, **Castellano** was fortunate to finish second since **Manuel Franco** would have won \"**Place**\" if his horse hadn't been exhausted as he was charging the horse too much before the braking point and **Castellano** was clever and saved his horse's energy before catching up to **Franco** then by taking a slightly wider turn at the exit point to get himself little room and then charged the horse. Manuel also charged his horse but the horse was already been exhausted which could be seen [here](https://youtu.be/-OId_SqGt_I?t=2200).\n\n**Side Note:** **Franco** also took a big wider turn and Apexing late than **Rosario**. However, he had made **2 main mistakes**. **First** not having a good head start in comparison to Rosario and Castellano, and **second**, charging a horse before or at a breaking point to get lead. Also, an interesting thing to note is that **Tyler** stumbled at the start of the race, but he still managed to finish in fourth place by taking the shorter route and charging the horse after the exit point.\n\n#### **In short words**:\n\n1. **Rosario:** Had good head start and charged the horse at the correct time and angle following the standard racing convention. (Wide turn and Late Apex)\n\n2. **Castellano:** Also had a good head start, however, he made a mistake by coming towards the rails at the braking point as it was not necessary he was already on the right of Rosario and then he thought to step back a little as Franco was also neck to neck with him and was lucky that Franco's horse was exhausted.\n\n#### **Key Findings for this particular race:** (Should not Consider this as Main Finding)\n\nIf you stumbled upon the start always try to aim for a shorter distance instead of a fast route and if you get off to a good start, stick to the traditional racing line rather than a geometric one.\n\nLastly, thank you for your response I really appreciate it.", "1957289": "Kush:\n\nGreat stuff. Focusing on velocity/acceleration metrics going in and coming out of turns is a very good use of this dataset.", "1957353": "Correct! also we have to find the minimum curvature probabilistic inference as that could help to improve lap time as well as increment in speed." }, "source": "meta" }