Dear Kagglers,
\nWelcome to the G-Research Crypto Challenge!
\nWe are excited to launch our first Kaggle competition and to see how the community tackles a highly challenging prediction problem: forecasting cryptocurrency returns.
\nWorking with a massive, real-world dataset, you will need to find meaningful signal amongst huge amounts of noise. After building, testing and fine-tuning your model, it will be tested against live market data during a three-month evaluation phase.
\nG-Research is Europeโs leading quantitative finance research firm. We have long explored the extent of market prediction possibilities, making use of machine learning, big data, and some of the most advanced technology available. Through this competition, you will gain real insight into the type of challenges we face in our everyday work. We are also hiring and if you enjoy tackling the Crypto Challenge, we would encourage you to check out our vacancies: https://www.gresearch.co.uk/join-us/
\nTo get started, we would encourage you to review the tutorial materials available on the competition page and to ask any questions on the forum.
\nThanks and very best of luck!
\nG-Research
", "messages": [ { "id": 1568272, "postDate": "2021-11-02T16:31:05.673Z", "content": "Dear Kagglers,
\nWelcome to the G-Research Crypto Challenge!
\nWe are excited to launch our first Kaggle competition and to see how the community tackles a highly challenging prediction problem: forecasting cryptocurrency returns.
\nWorking with a massive, real-world dataset, you will need to find meaningful signal amongst huge amounts of noise. After building, testing and fine-tuning your model, it will be tested against live market data during a three-month evaluation phase.
\nG-Research is Europeโs leading quantitative finance research firm. We have long explored the extent of market prediction possibilities, making use of machine learning, big data, and some of the most advanced technology available. Through this competition, you will gain real insight into the type of challenges we face in our everyday work. We are also hiring and if you enjoy tackling the Crypto Challenge, we would encourage you to check out our vacancies: https://www.gresearch.co.uk/join-us/
\nTo get started, we would encourage you to review the tutorial materials available on the competition page and to ask any questions on the forum.
\nThanks and very best of luck!
\nG-Research
", "rawMarkdown": "Dear Kagglers,\n\nWelcome to the G-Research Crypto Challenge!\n\nWe are excited to launch our first Kaggle competition and to see how the community tackles a highly challenging prediction problem: forecasting cryptocurrency returns.\n\nWorking with a massive, real-world dataset, you will need to find meaningful signal amongst huge amounts of noise. After building, testing and fine-tuning your model, it will be tested against live market data during a three-month evaluation phase.\n\nG-Research is Europeโs leading quantitative finance research firm. We have long explored the extent of market prediction possibilities, making use of machine learning, big data, and some of the most advanced technology available. Through this competition, you will gain real insight into the type of challenges we face in our everyday work. We are also hiring and if you enjoy tackling the Crypto Challenge, we would encourage you to check out our vacancies: [https://www.gresearch.co.uk/join-us/](https://www.gresearch.co.uk/join-us/)\n\nTo get started, we would encourage you to review the tutorial materials available on the competition page and to ask any questions on the forum.\n\nThanks and very best of luck!\n\nG-Research", "votes": 65 }, { "id": 1575278, "postDate": "2021-11-08T09:49:23.920Z", "content": "Thanks to G-Research and Kaggle for hosting an interesting and unique competition !!
\nAll the best to every participant ๐
Thank you everyone for your interest and excitement for this forecasting competition!
\nIt is fantastic to see all the engagement and sharing of suggested resources, techniques and approaches - please do keep them coming. Kudos to @yamqwe @julian3833 @andy1010 @crained @rdizzl3 @romulomadu @lucasmorin @yus002 @carlmcbrideellis @odins0n for your excellent and insightful posts, support and questions ๐
\nOur team will be following up this week to help provide guidance and clarification on the key points and questions raised over the past few days to help you on your ways in the coming months.
\nIn particular, on the public leaderboard, we will provide guidance to help you make best use of this to test your models ready for the evaluation phase in 3 months time which will use future market data.
\nIt is worth reiterating that - โThis public leaderboard uses publicly available data. Ranks here are not meaningful. Furthermore, the public leaderboard targets are publicly available and are provided as part of the competition dataset. Expect to see many people submitting perfect submissions for fun.โ
\nNevertheless, you can still use the leaderboard to test out your models on the existing market data provided. More to follow on this soon.
\nThanks everyone and good luck!
", "rawMarkdown": "Thank you everyone for your interest and excitement for this forecasting competition! \n\nIt is fantastic to see all the engagement and sharing of suggested resources, techniques and approaches - please do keep them coming. Kudos to @yamqwe @julian3833 @andy1010 @crained @rdizzl3 @romulomadu @lucasmorin @yus002 @carlmcbrideellis @odins0n for your excellent and insightful posts, support and questions ๐\n\nOur team will be following up this week to help provide guidance and clarification on the key points and questions raised over the past few days to help you on your ways in the coming months. \n\nIn particular, on the public leaderboard, we will provide guidance to help you make best use of this to test your models ready for the evaluation phase in 3 months time which will use future market data. \n\nIt is worth reiterating that - โThis public leaderboard uses publicly available data. Ranks here are not meaningful. Furthermore, the public leaderboard targets are publicly available and are provided as part of the competition dataset. Expect to see many people submitting perfect submissions for fun.โ \n\nNevertheless, you can still use the leaderboard to test out your models on the existing market data provided. More to follow on this soon. \n\nThanks everyone and good luck!\n", "votes": 8, "replies": [ { "id": 1579210, "postDate": "2021-11-11T16:33:50.947Z", "content": "Questions asked but not being answeredโฆ.
", "rawMarkdown": "Questions asked but not being answered...." } ] }, { "id": 1568768, "postDate": "2021-11-03T03:09:05.180Z", "content": "Thanks to G-Research and Kaggle for this competition.
\nExcited to learn and apply the art and science of short term crypto forecasting.
\nAll the best to all of us!
", "rawMarkdown": "Thanks to G-Research and Kaggle for this competition.\n\nExcited to learn and apply the art and science of short term crypto forecasting. \n\nAll the best to all of us!", "votes": 5 }, { "id": 1568695, "postDate": "2021-11-03T01:36:16.937Z", "content": "Thanks for hosting an interesting competition!
\nBTW, the following explanation seems to be wrong in https://www.kaggle.com/c/g-research-crypto-forecasting/data.
\nThanks to G-Research. ๐
\nGood luck to each participant. ๐
Thanks to G-Research and Kaggle for this competition.
", "rawMarkdown": "Thanks to G-Research and Kaggle for this competition.", "votes": 3 }, { "id": 1569646, "postDate": "2021-11-03T16:09:44.357Z", "content": "Thanks for hosting this competition.
\nI'm excited to learn and apply my knowlodge that I have been learning the past year in Data Analytics and ML.
\nRegards.
", "rawMarkdown": "Thanks for hosting this competition. \n\nI'm excited to learn and apply my knowlodge that I have been learning the past year in Data Analytics and ML. \n\nRegards. ", "votes": 3 }, { "id": 1569516, "postDate": "2021-11-03T14:03:31.633Z", "content": "Thanks a lot for setting up this competition. After the Jane Street and Optiver competitions, I was desperately waiting for a new \"machine learning in finance\" competition.The next three months will surely be exciting. :)
", "rawMarkdown": "Thanks a lot for setting up this competition. After the Jane Street and Optiver competitions, I was desperately waiting for a new \"machine learning in finance\" competition.The next three months will surely be exciting. :) ", "votes": 3 }, { "id": 1583280, "postDate": "2021-11-15T17:23:07.957Z", "content": "Thanks to G-Research and Kaggle for hosting an interesting competition !
", "rawMarkdown": "Thanks to G-Research and Kaggle for hosting an interesting competition !", "votes": 1 }, { "id": 1574359, "postDate": "2021-11-07T13:25:59.917Z", "content": "Thanks to G-Research and Kaggle for this competition.
", "rawMarkdown": "Thanks to G-Research and Kaggle for this competition.", "votes": 1 }, { "id": 1572880, "postDate": "2021-11-06T03:45:10.313Z", "content": "Thanks G-research and Kaggle for posting this competition. Its a Hot Topic nowadays for Analytics in Crypto world.
", "rawMarkdown": "Thanks G-research and Kaggle for posting this competition. Its a Hot Topic nowadays for Analytics in Crypto world.", "votes": 1 }, { "id": 1572737, "postDate": "2021-11-05T22:38:41.443Z", "content": "Can I use R ?
", "rawMarkdown": "Can I use R ?", "votes": 1 }, { "id": 1572736, "postDate": "2021-11-05T22:37:06.683Z", "content": "Can I use R instead python?
", "rawMarkdown": "Can I use R instead python?", "votes": 1, "replies": [ { "id": 1580383, "postDate": "2021-11-12T17:49:17.177Z", "content": "You can, by using the reticulate package to interact with the API.
", "rawMarkdown": "You can, by using [the reticulate package](https://rstudio.github.io/reticulate/) to interact with the API.", "votes": 1 } ] }, { "id": 1568784, "postDate": "2021-11-03T03:24:15.273Z", "content": "Good luck to everyone. Thanks to G-Research.
", "rawMarkdown": "Good luck to everyone. Thanks to G-Research.", "votes": 1 }, { "id": 1575141, "postDate": "2021-11-08T07:31:50.837Z", "content": "Thanks to G-Research ad Kaggle. All the best to all of us !
", "rawMarkdown": "Thanks to G-Research ad Kaggle. All the best to all of us !" }, { "id": 1590789, "postDate": "2021-11-21T16:43:05.013Z", "content": "you have developed a valuation method that has been used with bad results by those who have opened a competition in finance before you. The results have always been terrible when algorithms are tested on new values, behaving randomly. Aren't you worried about having the exact same result as your predecessors?
\nIn this link my post in which I explain why these methods do not work
\nhttps://www.kaggle.com/c/optiver-realized-volatility-prediction/discussion/265291
Andrea, the shared link is not opening.
", "rawMarkdown": "Andrea, the shared link is not opening." }, { "id": 1594184, "postDate": "2021-11-24T15:44:04.837Z", "content": "Thanks for reporting I have corrected it.
", "rawMarkdown": "Thanks for reporting I have corrected it." } ] }, { "id": 1734021, "postDate": "2022-03-24T22:09:42.657Z", "content": "Is there any friend who can send me the main data of this match, I really need it if they can send it to me on one of the social networks or other methods
", "rawMarkdown": "Is there any friend who can send me the main data of this match, I really need it if they can send it to me on one of the social networks or other methods" }, { "id": 1715139, "postDate": "2022-03-07T16:58:57.773Z", "content": "@andrewscherer @kierangarvey May I ask if the Rules Acceptance Deadline can be extended? I am new to Kaggle and I found this competition to be very interesting. I found that most competitions set Rules Acceptance Deadline one week before the submission deadline. I would be really appreciated it if you can extend the Rules Acceptance Deadline. Thanks!
", "rawMarkdown": "@andrewscherer @kierangarvey May I ask if the Rules Acceptance Deadline can be extended? I am new to Kaggle and I found this competition to be very interesting. I found that most competitions set Rules Acceptance Deadline one week before the submission deadline. I would be really appreciated it if you can extend the Rules Acceptance Deadline. Thanks! " }, { "id": 1672451, "postDate": "2022-02-02T05:15:20.453Z", "content": "Thanks to G-Research. The testing period starts. Good luck to everyone
", "rawMarkdown": "Thanks to G-Research. The testing period starts. Good luck to everyone" }, { "id": 1631318, "postDate": "2021-12-28T09:29:13.643Z", "content": "what makes this competition differently from other machine learning crypto market prediction models @kierangarvey @andrewscherer
", "rawMarkdown": "what makes this competition differently from other machine learning crypto market prediction models @kierangarvey @andrewscherer " }, { "id": 1629296, "postDate": "2021-12-26T01:07:20.273Z", "content": "is publicly available data useful ?
", "rawMarkdown": "is publicly available data useful ?" }, { "id": 1626503, "postDate": "2021-12-22T22:37:58.610Z", "content": "Out of all competitions, the crypto forecasting is most interesting topic for me personally, because it is quit practical. So, thanks for launching this competition.
", "rawMarkdown": "Out of all competitions, the crypto forecasting is most interesting topic for me personally, because it is quit practical. So, thanks for launching this competition. " }, { "id": 1577655, "postDate": "2021-11-10T09:25:45.720Z", "content": "simulated account competition is much more better than these trading forecast.
", "rawMarkdown": " simulated account competition is much more better than these trading forecast. " }, { "id": 1577379, "postDate": "2021-11-10T03:32:02.670Z", "content": "Thanks to G-Research and Kaggle for this competition.
", "rawMarkdown": "Thanks to G-Research and Kaggle for this competition." }, { "id": 1577021, "postDate": "2021-11-09T17:31:04.360Z", "content": "@kierangarvey
\n@andrewscherer
@kierangarvey ; @andrewscherer
\nAny comments on the above?
", "rawMarkdown": "@kierangarvey ; @andrewscherer \n\nAny comments on the above? " }, { "id": 1579941, "postDate": "2021-11-12T10:43:59.293Z", "content": "@kierangarvey ; @andrewscherer
\nAny comments on the above?
", "rawMarkdown": "@kierangarvey ; @andrewscherer\n\nAny comments on the above?" }, { "id": 1579969, "postDate": "2021-11-12T11:19:40.090Z", "content": "Hi @ambank1 hopefully your question on the eval metric is covered in the recent guidance.
\nFor the API - this is a necessary feature of a forecasting competition to prevent accidental look forward
@kierangarvey ; @andrewscherer
\nWhat about the range of data - is test data strictly continuous from train data?
\nAlso, as it's quoted:
\n\"Expect to see roughly three months worth of data in the test set. Until the forecasting phase of the competition, the API will just deliver a slice of the training data.\"
I really have no idea what this mean - can you clarify more on this?
", "rawMarkdown": "@kierangarvey ; @andrewscherer \n\nWhat about the range of data - is test data strictly continuous from train data? \n\nAlso, as it's quoted: \n\"Expect to see roughly three months worth of data in the test set. Until the forecasting phase of the competition, the API will just deliver a slice of the training data.\"\n\nI really have no idea what this mean - can you clarify more on this? " }, { "id": 1582230, "postDate": "2021-11-14T16:58:42.417Z", "content": "@kierangarvey ; @andrewscherer
\nAny queries on the date range? and perhaps some clarification on the above.
", "rawMarkdown": "@kierangarvey ; @andrewscherer\n\nAny queries on the date range? and perhaps some clarification on the above." }, { "id": 1583077, "postDate": "2021-11-15T14:14:32.667Z", "content": "Hi @ambank1 the date range of the test period is approximately the 3 last months of the training data
", "rawMarkdown": "Hi @ambank1 the date range of the test period is approximately the 3 last months of the training data " } ] }, { "id": 1576748, "postDate": "2021-11-09T12:39:45.290Z", "content": "good luck everyone
", "rawMarkdown": "good luck everyone\n" }, { "id": 1576728, "postDate": "2021-11-09T12:19:50.497Z", "content": "Thanks to G-Research. ๐
\nGood luck to each participant. ๐
Thank you G-Research and Kaggle for this competition :)
\nI am new to Kaggle competitions and was wondering if there is a thread where I can post questions if i have problems using libraries required for this competions?
\n( I cant run the import gresearch_crypto in my Kaggle notebook, nor pip install it)
Thanks, Best of luck all, from Gisle
\n :)
What will be the prediction resolution?
", "rawMarkdown": "What will be the prediction resolution?" }, { "id": 1570763, "postDate": "2021-11-04T11:01:39.200Z", "content": "Thanks for this.
\nI've been waiting patiently to test my knowledge for the past one year. This is a great opportunity for me.
I am new here. Anyone help me how it works?
", "rawMarkdown": "I am new here. Anyone help me how it works?" }, { "id": 1779712, "postDate": "2022-05-06T18:35:49.643Z", "rawMarkdown": "", "isDeleted": true }, { "id": 1759925, "postDate": "2022-04-19T01:57:43.840Z", "rawMarkdown": "", "votes": 1, "isDeleted": true }, { "id": 1568546, "postDate": "2021-11-02T20:45:36.123Z", "rawMarkdown": "", "votes": 5, "isDeleted": true, "replies": [ { "id": 1568581, "postDate": "2021-11-02T21:42:45.940Z", "rawMarkdown": "", "isDeleted": true }, { "id": 1568595, "postDate": "2021-11-02T22:13:26.950Z", "rawMarkdown": "", "votes": 1, "isDeleted": true }, { "id": 1569045, "postDate": "2021-11-03T08:01:05.660Z", "content": "Even after you won the Janstreet competition? That tells me something about the hope of getting a full-time job there even you get a good rank in this competition. Thanks.
", "rawMarkdown": "Even after you won the Janstreet competition? That tells me something about the hope of getting a full-time job there even you get a good rank in this competition. Thanks." }, { "id": 1569396, "postDate": "2021-11-03T12:07:15.323Z", "rawMarkdown": "", "isDeleted": true }, { "id": 1569998, "postDate": "2021-11-03T22:17:19.867Z", "rawMarkdown": "", "isDeleted": true }, { "id": 1570002, "postDate": "2021-11-03T22:26:29.733Z", "content": "it's a notorious company for
\n1) sending former employee into prison three times and using private investigator to stalk former employee
\nhttps://www.bloomberg.com/news/features/2018-11-19/the-triple-jeopardy-of-ke-xu-a-chinese-hedge-fund-quant
\n2) forcing employee to work from office during tier 4 lock down
\nhttps://www.reddit.com/r/cscareerquestionsEU/comments/kihdac/london_tier_4_return_to_office_enforced/
\n3) boss is somehow like yingguo zhao jia ren;
Thanks to G-Research and Kaggle for hosting an interesting and unique competition !!
\nAll the best to every participant ๐
Thank you everyone for your interest and excitement for this forecasting competition!
\nIt is fantastic to see all the engagement and sharing of suggested resources, techniques and approaches - please do keep them coming. Kudos to @yamqwe @julian3833 @andy1010 @crained @rdizzl3 @romulomadu @lucasmorin @yus002 @carlmcbrideellis @odins0n for your excellent and insightful posts, support and questions ๐
\nOur team will be following up this week to help provide guidance and clarification on the key points and questions raised over the past few days to help you on your ways in the coming months.
\nIn particular, on the public leaderboard, we will provide guidance to help you make best use of this to test your models ready for the evaluation phase in 3 months time which will use future market data.
\nIt is worth reiterating that - โThis public leaderboard uses publicly available data. Ranks here are not meaningful. Furthermore, the public leaderboard targets are publicly available and are provided as part of the competition dataset. Expect to see many people submitting perfect submissions for fun.โ
\nNevertheless, you can still use the leaderboard to test out your models on the existing market data provided. More to follow on this soon.
\nThanks everyone and good luck!
", "votes": 8, "replies": [ { "id": 1579210, "author_name": "ambank1", "author_url": "", "post_date": "2021-11-11T16:33:50.947000", "content": "Questions asked but not being answeredโฆ.
", "votes": 0, "replies": [] } ] }, { "id": 1568768, "author_name": "SRK", "author_url": "", "post_date": "2021-11-03T03:09:05.180000", "content": "Thanks to G-Research and Kaggle for this competition.
\nExcited to learn and apply the art and science of short term crypto forecasting.
\nAll the best to all of us!
", "votes": 5, "replies": [] }, { "id": 1568695, "author_name": "katsu1110", "author_url": "", "post_date": "2021-11-03T01:36:16.937000", "content": "Thanks for hosting an interesting competition!
\nBTW, the following explanation seems to be wrong in https://www.kaggle.com/c/g-research-crypto-forecasting/data.
\nThanks to G-Research. ๐
\nGood luck to each participant. ๐
Thanks to G-Research and Kaggle for this competition.
", "votes": 3, "replies": [] }, { "id": 1569646, "author_name": "Moha Arfaoui", "author_url": "", "post_date": "2021-11-03T16:09:44.357000", "content": "Thanks for hosting this competition.
\nI'm excited to learn and apply my knowlodge that I have been learning the past year in Data Analytics and ML.
\nRegards.
", "votes": 3, "replies": [] }, { "id": 1569516, "author_name": "Tarlan Nazarov", "author_url": "", "post_date": "2021-11-03T14:03:31.633000", "content": "Thanks a lot for setting up this competition. After the Jane Street and Optiver competitions, I was desperately waiting for a new \"machine learning in finance\" competition.The next three months will surely be exciting. :)
", "votes": 3, "replies": [] }, { "id": 1583280, "author_name": "Pan Zihan", "author_url": "", "post_date": "2021-11-15T17:23:07.957000", "content": "Thanks to G-Research and Kaggle for hosting an interesting competition !
", "votes": 1, "replies": [] }, { "id": 1574359, "author_name": "Billy Feng", "author_url": "", "post_date": "2021-11-07T13:25:59.917000", "content": "Thanks to G-Research and Kaggle for this competition.
", "votes": 1, "replies": [] }, { "id": 1572880, "author_name": "Brijesh Yadav", "author_url": "", "post_date": "2021-11-06T03:45:10.313000", "content": "Thanks G-research and Kaggle for posting this competition. Its a Hot Topic nowadays for Analytics in Crypto world.
", "votes": 1, "replies": [] }, { "id": 1572737, "author_name": "Lyes Bensaid", "author_url": "", "post_date": "2021-11-05T22:38:41.443000", "content": "Can I use R ?
", "votes": 1, "replies": [] }, { "id": 1572736, "author_name": "Lyes Bensaid", "author_url": "", "post_date": "2021-11-05T22:37:06.683000", "content": "Can I use R instead python?
", "votes": 1, "replies": [ { "id": 1580383, "author_name": "Sohier Dane", "author_url": "", "post_date": "2021-11-12T17:49:17.177000", "content": "You can, by using the reticulate package to interact with the API.
", "votes": 1, "replies": [] } ] }, { "id": 1568784, "author_name": "neoistheone", "author_url": "", "post_date": "2021-11-03T03:24:15.273000", "content": "Good luck to everyone. Thanks to G-Research.
", "votes": 1, "replies": [] }, { "id": 1575141, "author_name": "Anant Kumar", "author_url": "", "post_date": "2021-11-08T07:31:50.837000", "content": "Thanks to G-Research ad Kaggle. All the best to all of us !
", "votes": 0, "replies": [] }, { "id": 1590789, "author_name": "Andrea Berdondini", "author_url": "", "post_date": "2021-11-21T16:43:05.013000", "content": "you have developed a valuation method that has been used with bad results by those who have opened a competition in finance before you. The results have always been terrible when algorithms are tested on new values, behaving randomly. Aren't you worried about having the exact same result as your predecessors?
\nIn this link my post in which I explain why these methods do not work
\nhttps://www.kaggle.com/c/optiver-realized-volatility-prediction/discussion/265291
Andrea, the shared link is not opening.
", "votes": 0, "replies": [] }, { "id": 1594184, "author_name": "Andrea Berdondini", "author_url": "", "post_date": "2021-11-24T15:44:04.837000", "content": "Thanks for reporting I have corrected it.
", "votes": 0, "replies": [] } ] }, { "id": 1734021, "author_name": "Mahdi_asdzd", "author_url": "", "post_date": "2022-03-24T22:09:42.657000", "content": "Is there any friend who can send me the main data of this match, I really need it if they can send it to me on one of the social networks or other methods
", "votes": 0, "replies": [] }, { "id": 1715139, "author_name": "Mlin", "author_url": "", "post_date": "2022-03-07T16:58:57.773000", "content": "@andrewscherer @kierangarvey May I ask if the Rules Acceptance Deadline can be extended? I am new to Kaggle and I found this competition to be very interesting. I found that most competitions set Rules Acceptance Deadline one week before the submission deadline. I would be really appreciated it if you can extend the Rules Acceptance Deadline. Thanks!
", "votes": 0, "replies": [] }, { "id": 1672451, "author_name": "Faust", "author_url": "", "post_date": "2022-02-02T05:15:20.453000", "content": "Thanks to G-Research. The testing period starts. Good luck to everyone
", "votes": 0, "replies": [] }, { "id": 1631318, "author_name": "Muhammad Hafil T", "author_url": "", "post_date": "2021-12-28T09:29:13.643000", "content": "what makes this competition differently from other machine learning crypto market prediction models @kierangarvey @andrewscherer
", "votes": 0, "replies": [] }, { "id": 1629296, "author_name": "Hitesh kumar", "author_url": "", "post_date": "2021-12-26T01:07:20.273000", "content": "is publicly available data useful ?
", "votes": 0, "replies": [] }, { "id": 1626503, "author_name": "Selina Chen", "author_url": "", "post_date": "2021-12-22T22:37:58.610000", "content": "Out of all competitions, the crypto forecasting is most interesting topic for me personally, because it is quit practical. So, thanks for launching this competition.
", "votes": 0, "replies": [] }, { "id": 1577655, "author_name": "dragon zhang", "author_url": "", "post_date": "2021-11-10T09:25:45.720000", "content": "simulated account competition is much more better than these trading forecast.
", "votes": 0, "replies": [] }, { "id": 1577379, "author_name": "hllu1109", "author_url": "", "post_date": "2021-11-10T03:32:02.670000", "content": "Thanks to G-Research and Kaggle for this competition.
", "votes": 0, "replies": [] }, { "id": 1577021, "author_name": "ambank1", "author_url": "", "post_date": "2021-11-09T17:31:04.360000", "content": "@kierangarvey
\n@andrewscherer
@kierangarvey ; @andrewscherer
\nAny comments on the above?
", "votes": 0, "replies": [] }, { "id": 1579941, "author_name": "ambank1", "author_url": "", "post_date": "2021-11-12T10:43:59.293000", "content": "@kierangarvey ; @andrewscherer
\nAny comments on the above?
", "votes": 0, "replies": [] }, { "id": 1579969, "author_name": "Kieran Garvey", "author_url": "", "post_date": "2021-11-12T11:19:40.090000", "content": "Hi @ambank1 hopefully your question on the eval metric is covered in the recent guidance.
\nFor the API - this is a necessary feature of a forecasting competition to prevent accidental look forward
@kierangarvey ; @andrewscherer
\nWhat about the range of data - is test data strictly continuous from train data?
\nAlso, as it's quoted:
\n\"Expect to see roughly three months worth of data in the test set. Until the forecasting phase of the competition, the API will just deliver a slice of the training data.\"
I really have no idea what this mean - can you clarify more on this?
", "votes": 0, "replies": [] }, { "id": 1582230, "author_name": "ambank1", "author_url": "", "post_date": "2021-11-14T16:58:42.417000", "content": "@kierangarvey ; @andrewscherer
\nAny queries on the date range? and perhaps some clarification on the above.
", "votes": 0, "replies": [] }, { "id": 1583077, "author_name": "Kieran Garvey", "author_url": "", "post_date": "2021-11-15T14:14:32.667000", "content": "Hi @ambank1 the date range of the test period is approximately the 3 last months of the training data
", "votes": 0, "replies": [] } ] }, { "id": 1576748, "author_name": "Prajwal panta", "author_url": "", "post_date": "2021-11-09T12:39:45.290000", "content": "good luck everyone
", "votes": 0, "replies": [] }, { "id": 1576728, "author_name": "Jillani SofTech", "author_url": "", "post_date": "2021-11-09T12:19:50.497000", "content": "Thanks to G-Research. ๐
\nGood luck to each participant. ๐
Thank you G-Research and Kaggle for this competition :)
\nI am new to Kaggle competitions and was wondering if there is a thread where I can post questions if i have problems using libraries required for this competions?
\n( I cant run the import gresearch_crypto in my Kaggle notebook, nor pip install it)
Thanks, Best of luck all, from Gisle
\n :)
What will be the prediction resolution?
", "votes": 0, "replies": [] }, { "id": 1570763, "author_name": "", "author_url": "", "post_date": "2021-11-04T11:01:39.200000", "content": "", "votes": 0, "replies": [] }, { "id": 1569038, "author_name": "", "author_url": "", "post_date": "2021-11-03T07:58:29.337000", "content": "", "votes": 0, "replies": [] }, { "id": 1779712, "author_name": "", "author_url": "", "post_date": "2022-05-06T18:35:49.643000", "content": "", "votes": 0, "replies": [] }, { "id": 1759925, "author_name": "", "author_url": "", "post_date": "2022-04-19T01:57:43.840000", "content": "", "votes": 1, "replies": [] }, { "id": 1568546, "author_name": "", "author_url": "", "post_date": "2021-11-02T20:45:36.123000", "content": "", "votes": 5, "replies": [ { "id": 1568581, "author_name": "", "author_url": "", "post_date": "2021-11-02T21:42:45.940000", "content": "", "votes": 0, "replies": [] }, { "id": 1568595, "author_name": "", "author_url": "", "post_date": "2021-11-02T22:13:26.950000", "content": "", "votes": 1, "replies": [] }, { "id": 1569045, "author_name": "", "author_url": "", "post_date": "2021-11-03T08:01:05.660000", "content": "", "votes": 0, "replies": [] }, { "id": 1569396, "author_name": "", "author_url": "", "post_date": "2021-11-03T12:07:15.323000", "content": "", "votes": 0, "replies": [] }, { "id": 1569998, "author_name": "", "author_url": "", "post_date": "2021-11-03T22:17:19.867000", "content": "", "votes": 0, "replies": [] }, { "id": 1570002, "author_name": "", "author_url": "", "post_date": "2021-11-03T22:26:29.733000", "content": "", "votes": 2, "replies": [] }, { "id": 1570023, "author_name": "", "author_url": "", "post_date": "2021-11-03T23:01:21.913000", "content": "", "votes": 0, "replies": [] } ] } ], "raw_markdown_by_id": { "1568272": "Dear Kagglers,\n\nWelcome to the G-Research Crypto Challenge!\n\nWe are excited to launch our first Kaggle competition and to see how the community tackles a highly challenging prediction problem: forecasting cryptocurrency returns.\n\nWorking with a massive, real-world dataset, you will need to find meaningful signal amongst huge amounts of noise. After building, testing and fine-tuning your model, it will be tested against live market data during a three-month evaluation phase.\n\nG-Research is Europeโs leading quantitative finance research firm. We have long explored the extent of market prediction possibilities, making use of machine learning, big data, and some of the most advanced technology available. Through this competition, you will gain real insight into the type of challenges we face in our everyday work. We are also hiring and if you enjoy tackling the Crypto Challenge, we would encourage you to check out our vacancies: [https://www.gresearch.co.uk/join-us/](https://www.gresearch.co.uk/join-us/)\n\nTo get started, we would encourage you to review the tutorial materials available on the competition page and to ask any questions on the forum.\n\nThanks and very best of luck!\n\nG-Research", "1575278": "Thanks to G-Research and Kaggle for hosting an interesting and unique competition !!\nAll the best to every participant ๐", "1575146": "Thank you everyone for your interest and excitement for this forecasting competition! \n\nIt is fantastic to see all the engagement and sharing of suggested resources, techniques and approaches - please do keep them coming. Kudos to @yamqwe @julian3833 @andy1010 @crained @rdizzl3 @romulomadu @lucasmorin @yus002 @carlmcbrideellis @odins0n for your excellent and insightful posts, support and questions ๐\n\nOur team will be following up this week to help provide guidance and clarification on the key points and questions raised over the past few days to help you on your ways in the coming months. \n\nIn particular, on the public leaderboard, we will provide guidance to help you make best use of this to test your models ready for the evaluation phase in 3 months time which will use future market data. \n\nIt is worth reiterating that - โThis public leaderboard uses publicly available data. Ranks here are not meaningful. Furthermore, the public leaderboard targets are publicly available and are provided as part of the competition dataset. Expect to see many people submitting perfect submissions for fun.โ \n\nNevertheless, you can still use the leaderboard to test out your models on the existing market data provided. More to follow on this soon. \n\nThanks everyone and good luck!\n", "1568768": "Thanks to G-Research and Kaggle for this competition.\n\nExcited to learn and apply the art and science of short term crypto forecasting. \n\nAll the best to all of us!", "1568695": "Thanks for hosting an interesting competition!\n\nBTW, the following explanation seems to be wrong in https://www.kaggle.com/c/g-research-crypto-forecasting/data.\n\n- example_sample_submission.csv - An example of the data that will be delivered by the time series API. The data is just copied from train.csv.", "1568635": "Thanks to G-Research. ๐\nGood luck to each participant. ๐", "1570654": "Thanks to G-Research and Kaggle for this competition.", "1569646": "Thanks for hosting this competition. \n\nI'm excited to learn and apply my knowlodge that I have been learning the past year in Data Analytics and ML. \n\nRegards. ", "1569516": "Thanks a lot for setting up this competition. After the Jane Street and Optiver competitions, I was desperately waiting for a new \"machine learning in finance\" competition.The next three months will surely be exciting. :) ", "1583280": "Thanks to G-Research and Kaggle for hosting an interesting competition !", "1574359": "Thanks to G-Research and Kaggle for this competition.", "1572880": "Thanks G-research and Kaggle for posting this competition. Its a Hot Topic nowadays for Analytics in Crypto world.", "1572737": "Can I use R ?", "1572736": "Can I use R instead python?", "1568784": "Good luck to everyone. Thanks to G-Research.", "1575141": "Thanks to G-Research ad Kaggle. All the best to all of us !", "1590789": "you have developed a valuation method that has been used with bad results by those who have opened a competition in finance before you. The results have always been terrible when algorithms are tested on new values, behaving randomly. Aren't you worried about having the exact same result as your predecessors?\nIn this link my post in which I explain why these methods do not work\nhttps://www.kaggle.com/c/optiver-realized-volatility-prediction/discussion/265291", "1734021": "Is there any friend who can send me the main data of this match, I really need it if they can send it to me on one of the social networks or other methods", "1715139": "@andrewscherer @kierangarvey May I ask if the Rules Acceptance Deadline can be extended? I am new to Kaggle and I found this competition to be very interesting. I found that most competitions set Rules Acceptance Deadline one week before the submission deadline. I would be really appreciated it if you can extend the Rules Acceptance Deadline. Thanks! ", "1672451": "Thanks to G-Research. The testing period starts. Good luck to everyone", "1631318": "what makes this competition differently from other machine learning crypto market prediction models @kierangarvey @andrewscherer ", "1629296": "is publicly available data useful ?", "1626503": "Out of all competitions, the crypto forecasting is most interesting topic for me personally, because it is quit practical. So, thanks for launching this competition. ", "1577655": " simulated account competition is much more better than these trading forecast. ", "1577379": "Thanks to G-Research and Kaggle for this competition.", "1577021": "@kierangarvey \n@andrewscherer \n\n1. What is the range of test period; \n2. Is there reason why use time series API? Features requiring backward looking are tedious +to build when doing the prediction; \n3. Lastly, any reason for using this metric as evaluation? Would love to hear about this from host.", "1576748": "good luck everyone\n", "1576728": "Thanks to G-Research. ๐\nGood luck to each participant. ๐", "1575289": "Thank you G-Research and Kaggle for this competition :) \nI am new to Kaggle competitions and was wondering if there is a thread where I can post questions if i have problems using libraries required for this competions? \n( I cant run the import gresearch_crypto in my Kaggle notebook, nor pip install it)\n\nThanks, Best of luck all, from Gisle \n :) \n\n", "1570813": "What will be the prediction resolution?", "1570763": "Thanks for this.\nI've been waiting patiently to test my knowledge for the past one year. This is a great opportunity for me.", "1569038": "I am new here. Anyone help me how it works?", "1779712": "", "1759925": "", "1568546": "" } }