# How can I use the tournament data as a time series

**URL:** <https://forum.numer.ai/t/how-can-i-use-the-tournament-data-as-a-time-series/2176>\
**Category:** Tournament\
**Created:** [March 6, 2021, 11:26am UTC](https://forum.numer.ai/t/how-can-i-use-the-tournament-data-as-a-time-series/2176 "2021-03-06T11:26:42Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![aldente](https://avatars.discourse-cdn.com/v4/letter/a/eada6e/32.png) [@aldente](https://forum.numer.ai/u/aldente)\
**Post date:** [March 6, 2021, 11:26am UTC](https://forum.numer.ai/t/how-can-i-use-the-tournament-data-as-a-time-series/2176/1 "2021-03-06T11:26:42Z")

</div>

I wanna know if the row entries are actual ticks or not, and whether the targets are the values for this particular time, or is it the resulting target after a tick with the given features?

Assuming the features are a function of time and the target is also a function of time, is the data related as follows?  
Features(t) → target(t)  
OR  
Features(t) → target(t+1)

---

<div class="post-metadata">

**Author:** ![wigglemuse](https://yyz1.discourse-cdn.com/flex009/user_avatar/forum.numer.ai/wigglemuse/32/3582_2.png) [@wigglemuse](https://forum.numer.ai/u/wigglemuse)\
**Post date:** [March 6, 2021, 4:47pm UTC](https://forum.numer.ai/t/how-can-i-use-the-tournament-data-as-a-time-series/2176/2 "2021-03-06T16:47:13Z")

</div>

You can’t really do time-series with this data. You can’t match up rows from different periods. The target is 4 weeks (20 trading days) in the future relative to the features – that’s what you are predicting.

---

<div class="post-metadata">

**Author:** ![aldente](https://avatars.discourse-cdn.com/v4/letter/a/eada6e/32.png) [@aldente](https://forum.numer.ai/u/aldente)\
**Post date:** [March 7, 2021, 7:59am UTC](https://forum.numer.ai/t/how-can-i-use-the-tournament-data-as-a-time-series/2176/3 "2021-03-07T07:59:05Z")

</div>

But are the rows chronologically ordered?

---

<div class="post-metadata">

**Author:** ![wigglemuse](https://yyz1.discourse-cdn.com/flex009/user_avatar/forum.numer.ai/wigglemuse/32/3582_2.png) [@wigglemuse](https://forum.numer.ai/u/wigglemuse)\
**Post date:** [March 7, 2021, 2:50pm UTC](https://forum.numer.ai/t/how-can-i-use-the-tournament-data-as-a-time-series/2176/4 "2021-03-07T14:50:03Z")

</div>

The eras are chronologically ordered. The rows within each era are of the same time period. Each row is a different stock, so there is no time ordering within an era – an era is a snapshot of time of X number of stocks (X = number of rows in era). The row order within an era (i.e. the order of the stocks) should be considered random. And as I said, there is no way to link up a specific stock row in one era to the same stock in a different era (and it might not even be there anyway – the exact selection of stocks isn’t the same in each era either.)

---

<div class="post-metadata">

**Author:** ![aldente](https://avatars.discourse-cdn.com/v4/letter/a/eada6e/32.png) [@aldente](https://forum.numer.ai/u/aldente)\
**Post date:** [March 10, 2021, 2:44am UTC](https://forum.numer.ai/t/how-can-i-use-the-tournament-data-as-a-time-series/2176/5 "2021-03-10T02:44:45Z")

</div>

Alright, thank you for the explanation.