## Task

Task is to learn to predict realizations of geological layer depth sequence. 

We ask to predict **10 realizations** (with **300 positions** each) with best possible accuracy. 
**10 realizations** are required to account for uncertainty of predictions. We encourage submissions to have capability of changing the number of realizations to any number.
Diverse predictions would likely increase the score.

Note: if your model uses randomness to predict uncertainty, please **fix the seed** for reproducibility.  

## Columns
* Position number (in column names) is "**X**" (vertical section) position in feet relative to the current location of the drill bit.
* Values in cells are **Z**  (oriented up, in feet) coordinates of a layer boundary near a drilled well 

### Column **geology_id**
Hash-based id of a specific row - unique ID string. 

### Columns **-299**..**0**
Inferred (known) geology **Z** (oriented up) coordinate of a layer up to the current drilling location along **X** (vertical section).

The data is normalized so that **Z(0) = 0**. 

Note: not all columns **-299**..**0** contain a value. However, we guarantee that the 
values for columns **-49**..**0** are present with a value in all the rows. 
The values will not have gaps starting from the column **-k** (first with data) until **0**. 

### Columns **1**..**300**
Geology to be predicted (realization 0 out of 10, zero-indexed). 

In the `train.csv` these columns contains the true geology of layer segments.

### Columns **r_1_pos_1**..**r_1_pos_300** ... **r_9_pos_1**..**r_9_pos_300**
Geology to be predicted (realizations 1 to 9 out of 10, zero-indexed). 

Note: The order of the realizations **0 to 10** does not change the score. 
This includes realization 0 described above.  

In the `train.csv` these columns contains the true geology of layer segments same as Columns **1**..**300**.