{
"id": 627574,
"title": "What does \"Submission scoring error\" mean?",
"url": "/competitions/vesuvius-challenge-surface-detection/discussion/627574",
"author_name": "",
"post_date": "2025-11-16T22:52:16.753918500Z",
"votes": 2,
"comment_count": 1,
"views": 0,
"content": "![\"screenshot](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F30613376%2Fbe81d16f7479b0a3ca549c7c1bd07029%2FScreenshot%202025-11-16%20144310.png?generation=1763333167688658&alt=media\")
\nI'm new to Kaggle and trying to sort out that I am submitting things correctly. My first submission got scored, but the second didn't. The first one was just thresholding a slightly smoothed version of each image. To make sure that the content of what I was submitting made a difference, my second submission was just the logical inversion of the first one for each image. I figured that if the content was getting through, the second score should be dramatically worse than the original score (since I was basically trying to submit an approximate background mask instead of an approximate sheet mask). But intstead of getting a worse score, I got an error. Does that just mean that my guesses were so bad that the scoring code broke? That's my guess, but I would appreciate it if someone with more experience could confirm that. Thanks!
",
"messages": [
{
"id": "3332772",
"postDate": "11/16/2025 22:52:16",
"content": "\nI'm new to Kaggle and trying to sort out that I am submitting things correctly. My first submission got scored, but the second didn't. The first one was just thresholding a slightly smoothed version of each image. To make sure that the content of what I was submitting made a difference, my second submission was just the logical inversion of the first one for each image. I figured that if the content was getting through, the second score should be dramatically worse than the original score (since I was basically trying to submit an approximate background mask instead of an approximate sheet mask). But intstead of getting a worse score, I got an error. Does that just mean that my guesses were so bad that the scoring code broke? That's my guess, but I would appreciate it if someone with more experience could confirm that. Thanks!
",
"rawMarkdown": "\nI'm new to Kaggle and trying to sort out that I am submitting things correctly. My first submission got scored, but the second didn't. The first one was just thresholding a slightly smoothed version of each image. To make sure that the content of what I was submitting made a difference, my second submission was just the logical inversion of the first one for each image. I figured that if the content was getting through, the second score should be dramatically worse than the original score (since I was basically trying to submit an approximate background mask instead of an approximate sheet mask). But intstead of getting a worse score, I got an error. Does that just mean that my guesses were so bad that the scoring code broke? That's my guess, but I would appreciate it if someone with more experience could confirm that. Thanks!",
"votes": null
},
{
"id": "3337107",
"postDate": "11/18/2025 20:19:45",
"content": "Hi @mikebindschadler -- welcome to Kaggle! Here is a bit of documentation that we published with some common errors and how to troubleshoot them. Good luck!
",
"rawMarkdown": "Hi @mikebindschadler -- welcome to Kaggle! Here is a bit of [documentation](https://www.kaggle.com/code-competition-debugging) that we published with some common errors and how to troubleshoot them. Good luck!",
"votes": null
}
],
"comments": [
{
"id": 3337107,
"author_name": "macruzbar",
"author_url": "",
"post_date": "11/18/2025 20:19:45",
"content": "Hi @mikebindschadler -- welcome to Kaggle! Here is a bit of documentation that we published with some common errors and how to troubleshoot them. Good luck!
",
"votes": null,
"replies": []
}
],
"raw_markdown_by_id": {
"3332772": "\nI'm new to Kaggle and trying to sort out that I am submitting things correctly. My first submission got scored, but the second didn't. The first one was just thresholding a slightly smoothed version of each image. To make sure that the content of what I was submitting made a difference, my second submission was just the logical inversion of the first one for each image. I figured that if the content was getting through, the second score should be dramatically worse than the original score (since I was basically trying to submit an approximate background mask instead of an approximate sheet mask). But intstead of getting a worse score, I got an error. Does that just mean that my guesses were so bad that the scoring code broke? That's my guess, but I would appreciate it if someone with more experience could confirm that. Thanks!",
"3337107": "Hi @mikebindschadler -- welcome to Kaggle! Here is a bit of [documentation](https://www.kaggle.com/code-competition-debugging) that we published with some common errors and how to troubleshoot them. Good luck!"
},
"source": "meta"
}