There are several columns in the data that are present only in Train:
\nmentioning here for completeness:
\nand more related to possible Target:
\nplayer_scoring_next (train only): Which player scores at the end of the current event, in [0, 6), or -1 if the event does not end in a goal.team_scoring_next (train only): Which team scores at the end of the current event (A or B), or NaN if the event does not end in a goal.team_[A|B]_scoring_within_10sec (train only): [Target columns] Value of 1 if team_scoring_next == [A|B] and time_before_event is in [-10, 0], otherwise 0.Is it correct to say that:
\nevent_time is given only as a sanity check and for easier understanding of the columns team_[A|B]_scoring_within_10sec but since it is not present in test, can safely be ignored?team_[A|B]_scoring_within_10sec columns have originally been probabilities but have been converted to 1 or 0 by the competition hosts? (this is just to help me understand the problem better)player_scoring_next == 2 and team_scoring_next == A, the column team_[A|B]_scoring_within_10sec can still be equal to 0 because e.g. player A could have scored, but later than 10 seconds?team_[A|B]_scoring_within_10sec but before those would be converted to either 1 or 0?There are several columns in the data that are present only in Train:
\nmentioning here for completeness:
\nand more related to possible Target:
\nplayer_scoring_next (train only): Which player scores at the end of the current event, in [0, 6), or -1 if the event does not end in a goal.team_scoring_next (train only): Which team scores at the end of the current event (A or B), or NaN if the event does not end in a goal.team_[A|B]_scoring_within_10sec (train only): [Target columns] Value of 1 if team_scoring_next == [A|B] and time_before_event is in [-10, 0], otherwise 0.Is it correct to say that:
\nevent_time is given only as a sanity check and for easier understanding of the columns team_[A|B]_scoring_within_10sec but since it is not present in test, can safely be ignored?team_[A|B]_scoring_within_10sec columns have originally been probabilities but have been converted to 1 or 0 by the competition hosts? (this is just to help me understand the problem better)player_scoring_next == 2 and team_scoring_next == A, the column team_[A|B]_scoring_within_10sec can still be equal to 0 because e.g. player A could have scored, but later than 10 seconds?team_[A|B]_scoring_within_10sec but before those would be converted to either 1 or 0?\n", "rawMarkdown": "> 2. `team_[A|B]_scoring_within_10sec` columns have originally been probabilities but have been converted to `1` or `0` by the competition hosts? (this is just to help me understand the problem better)\n\n**No, there was a game and the Team either scored within 10 seconds (value = 1), or did not (value = 0).**\n\n> 4. do we need to essentially predict the probabilities for the column `team_[A|B]_scoring_within_10sec` but before those would be converted to either `1` or `0`?\n\n**Yes, because the log loss scoring scheme is clever and rewards accurate estimates of probabilities over the large test set. Although the 'ground truth' answer will be either 1 or 0, predicting these extreme values is not a recommended strategy because of the very substantial penalty for being confident but wrong.**", "votes": null } ], "comments": [ { "id": 1983260, "author_name": "jbomitchell", "author_url": "", "post_date": "10/11/2022 22:42:32", "content": "\n
\n- \n
team_[A|B]_scoring_within_10seccolumns have originally been probabilities but have been converted to1or0by the competition hosts? (this is just to help me understand the problem better)
\n No, there was a game and the Team either scored within 10 seconds (value = 1), or did not (value = 0).- do we need to essentially predict the probabilities for the column
\nteam_[A|B]_scoring_within_10secbut before those would be converted to either1or0?
\n Yes, because the log loss scoring scheme is clever and rewards accurate estimates of probabilities over the large test set. Although the 'ground truth' answer will be either 1 or 0, predicting these extreme values is not a recommended strategy because of the very substantial penalty for being confident but wrong.
\n", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "1983200": "There are several columns in the data that are present only in Train:\n\n*mentioning here for completeness:*\n- game_num (train only): Unique identifier for the game from which the event was taken.\n- event_id (train only): Unique identifier for the sequence of consecutive frames.\n- event_time (train only): Time in seconds before the event ended, either by a goal being scored or simply when we decided to truncate the timeseries if a goal was not scored.\n\n*and more related to possible Target:*\n\n- `player_scoring_next` (train only): Which player scores at the end of the current event, in [0, 6), or -1 if the event does not end in a goal.\n-`team_scoring_next` (train only): Which team scores at the end of the current event (A or B), or NaN if the event does not end in a goal.\n- `team_[A|B]_scoring_within_10sec` (train only): [Target columns] Value of 1 if team_scoring_next == [A|B] and time_before_event is in [-10, 0], otherwise 0.\n\nIs it correct to say that:\n1. `event_time` is given only as a sanity check and for easier understanding of the columns `team_[A|B]_scoring_within_10sec` but since it is not present in test, can safely be ignored?\n2. `team_[A|B]_scoring_within_10sec` columns have originally been probabilities but have been converted to `1` or `0` by the competition hosts? (this is just to help me understand the problem better)\n3. even if e.g. `player_scoring_next` == `2` and `team_scoring_next` == `A`, the column `team_[A|B]_scoring_within_10sec` can still be equal to `0` because e.g. player `A` could have scored, but later than `10` seconds?\n4. do we need to essentially predict the probabilities for the column `team_[A|B]_scoring_within_10sec` but before those would be converted to either `1` or `0`?", "1983260": "> 2. `team_[A|B]_scoring_within_10sec` columns have originally been probabilities but have been converted to `1` or `0` by the competition hosts? (this is just to help me understand the problem better)\n\n**No, there was a game and the Team either scored within 10 seconds (value = 1), or did not (value = 0).**\n\n> 4. do we need to essentially predict the probabilities for the column `team_[A|B]_scoring_within_10sec` but before those would be converted to either `1` or `0`?\n\n**Yes, because the log loss scoring scheme is clever and rewards accurate estimates of probabilities over the large test set. Although the 'ground truth' answer will be either 1 or 0, predicting these extreme values is not a recommended strategy because of the very substantial penalty for being confident but wrong.**" }, "source": "meta" }\n
\n- \n
team_[A|B]_scoring_within_10seccolumns have originally been probabilities but have been converted to1or0by the competition hosts? (this is just to help me understand the problem better)
\n No, there was a game and the Team either scored within 10 seconds (value = 1), or did not (value = 0).- do we need to essentially predict the probabilities for the column
\nteam_[A|B]_scoring_within_10secbut before those would be converted to either1or0?
\n Yes, because the log loss scoring scheme is clever and rewards accurate estimates of probabilities over the large test set. Although the 'ground truth' answer will be either 1 or 0, predicting these extreme values is not a recommended strategy because of the very substantial penalty for being confident but wrong.