The number of teams has jumped in the last few days from about 2600 to close to 3000 .\nAnd they all have somewhat decent scores. I was around 50% on the leaderboard and now I am at about 33%. As a rather new person to Kaggle, where are these new people coming from ? Were they simply working offsite and decided just now to test their model on a kernel? Are they people who just jump on during the last few days and grab any public kernel to get a mildly good score to just look good without understanding it ? \nWhy the sudden influx of new people? when it takes quite a while to commit each model.
", "messages": [ { "id": 618048, "postDate": "2019-09-04T19:01:43.493Z", "content": "Welcome to Kaggle. Most of the reason you said are correct. Its hard to imagine people working on the side for two months and during last 3 days doing submission... I think Majority just want to take best kernels and blend them together in hopes to get in the medal area... Also some accounts are cheaters...
\n\nEdit: You can check this thread where @valanm and @octpath0302 did great job on finding potential cheaters .. https://www.kaggle.com/c/aptos2019-blindness-detection/discussion/105511#latest-617646
", "rawMarkdown": "Welcome to Kaggle. Most of the reason you said are correct. Its hard to imagine people working on the side for two months and during last 3 days doing submission... I think Majority just want to take best kernels and blend them together in hopes to get in the medal area... Also some accounts are cheaters... \n\nEdit: You can check this thread where @valanm and @octpath0302 did great job on finding potential cheaters .. https://www.kaggle.com/c/aptos2019-blindness-detection/discussion/105511#latest-617646", "votes": 7, "replies": [ { "id": 620091, "postDate": "2019-09-07T02:49:55.707Z", "content": "A fake medal really has no value... If they think simply a score can get them a job, they're wrong. The interviewer isn't a fool, they will ask questions about the candidate's project in detail and those who just fork will quickly fail
", "rawMarkdown": "A fake medal really has no value... If they think simply a score can get them a job, they're wrong. The interviewer isn't a fool, they will ask questions about the candidate's project in detail and those who just fork will quickly fail ", "votes": 2 } ] }, { "id": 618041, "postDate": "2019-09-04T18:54:10.280Z", "content": "The number of teams has jumped in the last few days from about 2600 to close to 3000 .\nAnd they all have somewhat decent scores. I was around 50% on the leaderboard and now I am at about 33%. As a rather new person to Kaggle, where are these new people coming from ? Were they simply working offsite and decided just now to test their model on a kernel? Are they people who just jump on during the last few days and grab any public kernel to get a mildly good score to just look good without understanding it ? \nWhy the sudden influx of new people? when it takes quite a while to commit each model.
", "rawMarkdown": "The number of teams has jumped in the last few days from about 2600 to close to 3000 .\nAnd they all have somewhat decent scores. I was around 50% on the leaderboard and now I am at about 33%. As a rather new person to Kaggle, where are these new people coming from ? Were they simply working offsite and decided just now to test their model on a kernel? Are they people who just jump on during the last few days and grab any public kernel to get a mildly good score to just look good without understanding it ? \nWhy the sudden influx of new people? when it takes quite a while to commit each model. ", "votes": 6 }, { "id": 620099, "postDate": "2019-09-07T03:10:34.303Z", "content": "You're only allowed 5 submissions per day. I can see people creating new accounts just to have more chances to test their models.
", "rawMarkdown": "You're only allowed 5 submissions per day. I can see people creating new accounts just to have more chances to test their models.", "votes": 1 }, { "id": 620087, "postDate": "2019-09-07T02:28:50.700Z", "content": "👀
", "rawMarkdown": "👀", "votes": 1, "replies": [ { "id": 729592, "postDate": "2020-01-26T12:26:33.230Z", "rawMarkdown": "" } ] }, { "id": 618061, "postDate": "2019-09-04T19:18:26.133Z", "rawMarkdown": "", "isDeleted": true } ], "comments": [ { "id": 618048, "author_name": "DrHB", "author_url": "", "post_date": "2019-09-04T19:01:43.493000", "content": "Welcome to Kaggle. Most of the reason you said are correct. Its hard to imagine people working on the side for two months and during last 3 days doing submission... I think Majority just want to take best kernels and blend them together in hopes to get in the medal area... Also some accounts are cheaters...
\n\nEdit: You can check this thread where @valanm and @octpath0302 did great job on finding potential cheaters .. https://www.kaggle.com/c/aptos2019-blindness-detection/discussion/105511#latest-617646
", "votes": 7, "replies": [ { "id": 620091, "author_name": "Homoalways", "author_url": "", "post_date": "2019-09-07T02:49:55.707000", "content": "A fake medal really has no value... If they think simply a score can get them a job, they're wrong. The interviewer isn't a fool, they will ask questions about the candidate's project in detail and those who just fork will quickly fail
", "votes": 2, "replies": [] } ] }, { "id": 620099, "author_name": "Sterls", "author_url": "", "post_date": "2019-09-07T03:10:34.303000", "content": "You're only allowed 5 submissions per day. I can see people creating new accounts just to have more chances to test their models.
", "votes": 1, "replies": [] }, { "id": 620087, "author_name": "kumapo", "author_url": "", "post_date": "2019-09-07T02:28:50.700000", "content": "👀
", "votes": 1, "replies": [ { "id": 729592, "author_name": "", "author_url": "", "post_date": "2020-01-26T12:26:33.230000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 618061, "author_name": "", "author_url": "", "post_date": "2019-09-04T19:18:26.133000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "618048": "Welcome to Kaggle. Most of the reason you said are correct. Its hard to imagine people working on the side for two months and during last 3 days doing submission... I think Majority just want to take best kernels and blend them together in hopes to get in the medal area... Also some accounts are cheaters... \n\nEdit: You can check this thread where @valanm and @octpath0302 did great job on finding potential cheaters .. https://www.kaggle.com/c/aptos2019-blindness-detection/discussion/105511#latest-617646", "618041": "The number of teams has jumped in the last few days from about 2600 to close to 3000 .\nAnd they all have somewhat decent scores. I was around 50% on the leaderboard and now I am at about 33%. As a rather new person to Kaggle, where are these new people coming from ? Were they simply working offsite and decided just now to test their model on a kernel? Are they people who just jump on during the last few days and grab any public kernel to get a mildly good score to just look good without understanding it ? \nWhy the sudden influx of new people? when it takes quite a while to commit each model. ", "620099": "You're only allowed 5 submissions per day. I can see people creating new accounts just to have more chances to test their models.", "620087": "👀", "618061": "" } }