I must say I am looking forward to seeing how the top people on the leaderboard got such impressive scores for this challenge. It feels like I've tried everything and got nowhere near. I'm sure to learn a lot of things by reading the solutions.
\nI'll run out of time to exhaust all avenues of exploration I had but I feel like I gave it a pretty good shot without getting anywhere
", "messages": [ { "id": "3093834", "postDate": "01/11/2025 12:01:08", "content": "I must say I am looking forward to seeing how the top people on the leaderboard got such impressive scores for this challenge. It feels like I've tried everything and got nowhere near. I'm sure to learn a lot of things by reading the solutions.
\nI'll run out of time to exhaust all avenues of exploration I had but I feel like I gave it a pretty good shot without getting anywhere
", "rawMarkdown": "I must say I am looking forward to seeing how the top people on the leaderboard got such impressive scores for this challenge. It feels like I've tried everything and got nowhere near. I'm sure to learn a lot of things by reading the solutions. \n\nI'll run out of time to exhaust all avenues of exploration I had but I feel like I gave it a pretty good shot without getting anywhere", "votes": null }, { "id": "3094010", "postDate": "01/11/2025 14:29:51", "content": "In optiver only two people in the top10 published solutions, and only one of them published code. Let's see if it will be better this time 🤣
", "rawMarkdown": "In optiver only two people in the top10 published solutions, and only one of them published code. Let's see if it will be better this time 🤣", "votes": null }, { "id": "3094014", "postDate": "01/11/2025 14:37:51", "content": "I just wonder what secret the top10 teams have got so that they can get a score higher than 0.01 or even 0.012 🤔
", "rawMarkdown": "I just wonder what secret the top10 teams have got so that they can get a score higher than 0.01 or even 0.012 🤔", "votes": null }, { "id": "3094015", "postDate": "01/11/2025 14:39:02", "content": "In quantitive trading competitions like this one, it's very different from other competitions. Even if one team decides to share their solutions, they will choose to do it after the finishing of private test (which will be 7 months later in this comptetition), since the leak of one team's solution might be used by someone or some organizations, which will have side effect on the team's final result in the privte ranking.
", "rawMarkdown": "In quantitive trading competitions like this one, it's very different from other competitions. Even if one team decides to share their solutions, they will choose to do it after the finishing of private test (which will be 7 months later in this comptetition), since the leak of one team's solution might be used by someone or some organizations, which will have side effect on the team's final result in the privte ranking.", "votes": null }, { "id": "3094016", "postDate": "01/11/2025 14:40:12", "content": "Oh damn that sucks. I thought it was required to publish code to be eligible for prize. But this is the only kaggle competition I've ever looked at other than ARC so that's obviously a misunderstanding on my part :(
", "rawMarkdown": "Oh damn that sucks. I thought it was required to publish code to be eligible for prize. But this is the only kaggle competition I've ever looked at other than ARC so that's obviously a misunderstanding on my part :(", "votes": null }, { "id": "3094017", "postDate": "01/11/2025 14:40:44", "content": "I'm going to kick myself if its something simple or even a path I went down and gave up on
", "rawMarkdown": "I'm going to kick myself if its something simple or even a path I went down and gave up on", "votes": null }, { "id": "3094030", "postDate": "01/11/2025 15:04:08", "content": "We are required to publish code in competitions with winner solutions open source conditions. Those are usually competitions that are more research/academic in their nature. This competition and other past time series competitions, on the other hand, are more 'commercial' in nature, sponsored by companies and the winner solutions conditions are licensing the solutions to the hosts, not open sourcing it. Jane Street may even prefer that the solution is not open sourced, considering the nature and secrecy of quantitative trading.
", "rawMarkdown": "We are required to publish code in competitions with winner solutions open source conditions. Those are usually competitions that are more research/academic in their nature. This competition and other past time series competitions, on the other hand, are more 'commercial' in nature, sponsored by companies and the winner solutions conditions are licensing the solutions to the hosts, not open sourcing it. Jane Street may even prefer that the solution is not open sourced, considering the nature and secrecy of quantitative trading.", "votes": null }, { "id": "3094033", "postDate": "01/11/2025 15:07:00", "content": "Oh damn, I had not even thought about it!
", "rawMarkdown": "Oh damn, I had not even thought about it!", "votes": null }, { "id": "3094034", "postDate": "01/11/2025 15:09:31", "content": "@i2nfinit3y I would love to know how to get even a score like yours! TS is hard ~sigh 🤣
", "rawMarkdown": "i2nfinit3y I would love to know how to get even a score like yours! TS is hard ~sigh 🤣", "votes": null }, { "id": "3094110", "postDate": "01/11/2025 17:06:56", "content": "We have not yet found any precious secret that other people don't know 😂 But online learning indeed play an important role.
", "rawMarkdown": "We have not yet found any precious secret that other people don't know 😂 But online learning indeed play an important role.", "votes": null }, { "id": "3094232", "postDate": "01/11/2025 19:04:14", "content": "While it would be ideal if the code were shared, I’d still appreciate it if they could share their ideas, like insights into data preprocessing, model architecture, or other strategies. Even without ‘magic,’ we can learn to improve our model training. And if it turns out to be a unique trick tailored for this competition, I would try to push harder and be more creative next time.
", "rawMarkdown": "While it would be ideal if the code were shared, I’d still appreciate it if they could share their ideas, like insights into data preprocessing, model architecture, or other strategies. Even without ‘magic,’ we can learn to improve our model training. And if it turns out to be a unique trick tailored for this competition, I would try to push harder and be more creative next time.", "votes": null }, { "id": "3094392", "postDate": "01/12/2025 03:25:05", "content": "Top teams must have used an unknown feature engineering?
", "rawMarkdown": "Top teams must have used an unknown feature engineering?", "votes": null }, { "id": "3094471", "postDate": "01/12/2025 06:36:16", "content": "same same! hopefully they will share some general ideas of their solutions.
", "rawMarkdown": "same same! hopefully they will share some general ideas of their solutions.", "votes": null } ], "comments": [ { "id": 3094010, "author_name": "shlomoron", "author_url": "", "post_date": "01/11/2025 14:29:51", "content": "In optiver only two people in the top10 published solutions, and only one of them published code. Let's see if it will be better this time 🤣
", "votes": null, "replies": [ { "id": 3094016, "author_name": "michaeltimbs", "author_url": "", "post_date": "01/11/2025 14:40:12", "content": "Oh damn that sucks. I thought it was required to publish code to be eligible for prize. But this is the only kaggle competition I've ever looked at other than ARC so that's obviously a misunderstanding on my part :(
", "votes": null, "replies": [ { "id": 3094030, "author_name": "shlomoron", "author_url": "", "post_date": "01/11/2025 15:04:08", "content": "We are required to publish code in competitions with winner solutions open source conditions. Those are usually competitions that are more research/academic in their nature. This competition and other past time series competitions, on the other hand, are more 'commercial' in nature, sponsored by companies and the winner solutions conditions are licensing the solutions to the hosts, not open sourcing it. Jane Street may even prefer that the solution is not open sourced, considering the nature and secrecy of quantitative trading.
", "votes": null, "replies": [] } ] }, { "id": 3094232, "author_name": "maaax", "author_url": "", "post_date": "01/11/2025 19:04:14", "content": "While it would be ideal if the code were shared, I’d still appreciate it if they could share their ideas, like insights into data preprocessing, model architecture, or other strategies. Even without ‘magic,’ we can learn to improve our model training. And if it turns out to be a unique trick tailored for this competition, I would try to push harder and be more creative next time.
", "votes": null, "replies": [] } ] }, { "id": 3094014, "author_name": "i2nfinit3y", "author_url": "", "post_date": "01/11/2025 14:37:51", "content": "I just wonder what secret the top10 teams have got so that they can get a score higher than 0.01 or even 0.012 🤔
", "votes": null, "replies": [ { "id": 3094017, "author_name": "michaeltimbs", "author_url": "", "post_date": "01/11/2025 14:40:44", "content": "I'm going to kick myself if its something simple or even a path I went down and gave up on
", "votes": null, "replies": [] }, { "id": 3094034, "author_name": "shlomoron", "author_url": "", "post_date": "01/11/2025 15:09:31", "content": "@i2nfinit3y I would love to know how to get even a score like yours! TS is hard ~sigh 🤣
", "votes": null, "replies": [ { "id": 3094110, "author_name": "i2nfinit3y", "author_url": "", "post_date": "01/11/2025 17:06:56", "content": "We have not yet found any precious secret that other people don't know 😂 But online learning indeed play an important role.
", "votes": null, "replies": [] } ] } ] }, { "id": 3094015, "author_name": "lihaorocky", "author_url": "", "post_date": "01/11/2025 14:39:02", "content": "In quantitive trading competitions like this one, it's very different from other competitions. Even if one team decides to share their solutions, they will choose to do it after the finishing of private test (which will be 7 months later in this comptetition), since the leak of one team's solution might be used by someone or some organizations, which will have side effect on the team's final result in the privte ranking.
", "votes": null, "replies": [ { "id": 3094033, "author_name": "shlomoron", "author_url": "", "post_date": "01/11/2025 15:07:00", "content": "Oh damn, I had not even thought about it!
", "votes": null, "replies": [] } ] }, { "id": 3094392, "author_name": "lechengyan", "author_url": "", "post_date": "01/12/2025 03:25:05", "content": "Top teams must have used an unknown feature engineering?
", "votes": null, "replies": [] }, { "id": 3094471, "author_name": "sritichaimae", "author_url": "", "post_date": "01/12/2025 06:36:16", "content": "same same! hopefully they will share some general ideas of their solutions.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "3093834": "I must say I am looking forward to seeing how the top people on the leaderboard got such impressive scores for this challenge. It feels like I've tried everything and got nowhere near. I'm sure to learn a lot of things by reading the solutions. \n\nI'll run out of time to exhaust all avenues of exploration I had but I feel like I gave it a pretty good shot without getting anywhere", "3094010": "In optiver only two people in the top10 published solutions, and only one of them published code. Let's see if it will be better this time 🤣", "3094014": "I just wonder what secret the top10 teams have got so that they can get a score higher than 0.01 or even 0.012 🤔", "3094015": "In quantitive trading competitions like this one, it's very different from other competitions. Even if one team decides to share their solutions, they will choose to do it after the finishing of private test (which will be 7 months later in this comptetition), since the leak of one team's solution might be used by someone or some organizations, which will have side effect on the team's final result in the privte ranking.", "3094016": "Oh damn that sucks. I thought it was required to publish code to be eligible for prize. But this is the only kaggle competition I've ever looked at other than ARC so that's obviously a misunderstanding on my part :(", "3094017": "I'm going to kick myself if its something simple or even a path I went down and gave up on", "3094030": "We are required to publish code in competitions with winner solutions open source conditions. Those are usually competitions that are more research/academic in their nature. This competition and other past time series competitions, on the other hand, are more 'commercial' in nature, sponsored by companies and the winner solutions conditions are licensing the solutions to the hosts, not open sourcing it. Jane Street may even prefer that the solution is not open sourced, considering the nature and secrecy of quantitative trading.", "3094033": "Oh damn, I had not even thought about it!", "3094034": "i2nfinit3y I would love to know how to get even a score like yours! TS is hard ~sigh 🤣", "3094110": "We have not yet found any precious secret that other people don't know 😂 But online learning indeed play an important role.", "3094232": "While it would be ideal if the code were shared, I’d still appreciate it if they could share their ideas, like insights into data preprocessing, model architecture, or other strategies. Even without ‘magic,’ we can learn to improve our model training. And if it turns out to be a unique trick tailored for this competition, I would try to push harder and be more creative next time.", "3094392": "Top teams must have used an unknown feature engineering?", "3094471": "same same! hopefully they will share some general ideas of their solutions." }, "source": "meta" }