# Thread: Which Tests to applicate for Trends?

1. ## Which Tests to applicate for Trends?

Hi everyone, it's christmas time again, which for me means, some free time
During my free time I looked to acquire a bit more knowledge in some areas where my knowledge is somewhat not existing, so I was looking for a little project to do and what I came up with is Trend Analysis.

I tried to think of some "simple" use case where I can apply it and thought of my performance at sportsbetting, which is basically just my profit (or loss) from sportsbetting over the amount of bets I made.

Here is a simple graph of my performance, overall these are like ~10,000 bets, so quite a bit to go with and it looks roughly like that:

So at the beginning i tried around a bit and after some time I found myself a "good" strategy and since then I think I am kinda on an upward trend overall, with some minor setbacks inbetween, at least that's what i think when i just look at it.
( as said this is just an easy example of which i could think in order to apply the statistics somewhere )

So now I would like to statistically test, whether there actually is a trend in there or not, so I read up on a few trend tests, but found myself with quite a lot of overwhelming information.

First of all as a small question:
this graph, there is no monotic trend in there right? monotic means, that it's always increasing or breaking even and not decreasing over small amount of time, right?

so i can't test that for monotic trend, can i?

what do I then test for?

I would be happy if you could just drop a few names of some trend tests, that could be applicated here, cause I really need some pointers of what to look at, cause I can't seem to find the right tests.
I looked at the Mann-Kendall test as an example, but this is only for monotic

I am happy about every little reply, doesn't need to be long, just point me in the right direction, thanks!

2. ## Re: Which Tests to applicate for Trends?

Hi, usually you can decompose time-dependent data into 1.) an overall trend, 2.) a periodic component (such as a seasonal pattern), and 3.) noise. Some methods to extract trends from data are repeated LOESS-smooting, MAFA or dynamic factor analysis. But probably these methods are primarily for multiple time series, i.e. for multivariate data... I hope this helps.

3. ## Re: Which Tests to applicate for Trends?

Correct definition for monotonic. Also, monotonic is used in regards to missing data with repeated measures, and data can be organized to show latter data points are less likely to be completed.

 Tweet

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts