Session 4: advanced analytics

layout: true

<div class="my-footer">
  <div class="my-footer-box"><a href="https://openvolley.org/"><img style="display:inline;" src="extra/ovoutline-w.png"/>openvolley.org</a></div>
  <div class="my-footer-box"><a href="https://https://volleyball.ca/"><img src="extra/vc-w-wide.png"/></a></div>
  <div class="my-footer-box"><a href="https://untan.gl/"><img src="extra/su_title-w.png"/></a></div>
</div>

---

class: inverse, logo, center

## Session 4: Advanced analytics

### Adrien Ickowicz, Ben Raymond

##### with valuable contributions from many others...

---

## What are we going to talk about

- Statistical modelling
    - descriptive statistics vs inference
    - Is causality reachable ?
    - Inference via simulation vs Inference via statistical modelling
- Simulating matches
    - the <span class="pkg">volleysim</span> package
    - the <span class="pkg">bandit</span> package
- Statistical (inference-based) models
    - Use the example of the effect of player height on passing performance
    - the effect of serve speed on outcome? Serve error vs breakpoint ?

---

## Statistical modelling
#### descriptive statistics vs inference

From wiki:

A descriptive statistic (in the count noun sense) is a summary statistic that quantitatively describes or summarizes features from a collection of information, while descriptive statistics (in the mass noun sense) is the process of using and analysing those statistics.

Example of descriptive statistics:
* Number of serve errors per set
* Break-point rate per set

Example of inference:
* How do these two quantities relate to one another?

---

## Statistical modelling
#### Causality

credit: https://xkcd.com/

---

## Statistical modelling
#### Causality

credit: https://xkcd.com/

---

## Statistical modelling
#### Causality and correlation

https://www.tylervigen.com/spurious-correlations

So what is the problem?

- An increase in the number of people at the beach causes a higher number of ice cream sales
- An increase in the number of people at the beach causes a higher number of shark attacks
- Therefore, people need ice cream for comfort after a shark attack; or is it that sharks attack people after they ate ice cream because they taste better?

---

## Statistical modelling
#### Inference via simulation vs Inference via statistical modelling

When do we use either:
- When the world can be simplified without loss of (too much) information, we can simulate;
    - Example: Can we simplify a volleyball game if we only want to know who is going to win? Anyone has an opinion?
- When the world is too complex, and we have no clue or the process is convoluted, we may want to try something simpler at first;
    - Example: Are tall players better passers than short players?

In a nutshell, when you cannot narrow down the factors that influence the uncertainty to a handy few, forget about simulation.
    
    
---

## Simulating matches
#### the <span class="pkg">volleysim</span> package:

- vs_estimate_rates
- vs_example_file
- vs_match_win_probability
- vs_simulate_match
- vs_simulate_match_beach
- vs_simulate_match_mc
- vs_simulate_match_theor
- vs_simulate_set
- vs_simulate_set_mc
- vs_simulate_set_theor
- vs_theoretical_sideout_rates

---

## Simulating matches
#### the <span class="pkg">volleysim</span> package:

A quick overview: https://untan.gl/r-package-simulating-volleyball.html

Our turn. Open `s4-example-volleysim`.

---

## Simulating matches

The bandit functions in the <span class="pkg">ovlytics</span> package

- ov_simulate_setter_distribution
- ov_plot_ssd
- ov_plot_sequence_distribution
- ov_plot_distribution
- ov_create_history_table

What is this? See https://untan.gl/setter-choice.html

Our turn. Open `s4-example-bandit`.

And if you dont want to code... https://apps.untan.gl/bandit/

---

## Do you know...