1920:predictions_pl

Unterschiede

Hier werden die Unterschiede zwischen zwei Versionen angezeigt.

Link zu dieser Vergleichsansicht

Beide Seiten der vorigen Revision Vorhergehende Überarbeitung
Nächste Überarbeitung
Vorhergehende Überarbeitung
1920:predictions_pl [2019/08/10 06:41]
admin [Sites / Institutes]
1920:predictions_pl [2019/12/28 08:09] (aktuell)
admin [Interim results after the first round]
Zeile 98: Zeile 98:
 {{1920:​pred_pl.png?​640}} ​ {{1920:​pred_pl.png?​640}} ​
  
- + 
 +==== Interim results after the first round ==== 
 +=== Deviation from prediction === 
 +{{1920:​pred_mw-st_pl.png?​640}}  
 +=== Prediction score === 
 +{{1920:​mae_pl.png?​640}}
 /* /*
 ==== Comparison of average team predictions ==== ==== Comparison of average team predictions ====
 The following chart shows the final Premier League table together with the average predictions as $+$. The red circles mark the final standings. As well, the three best predictions are included (see discussion below), based on  the //best// mean-absolut-error prediction. ​ The following chart shows the final Premier League table together with the average predictions as $+$. The red circles mark the final standings. As well, the three best predictions are included (see discussion below), based on  the //best// mean-absolut-error prediction. ​
  
-{{1819:​pred_pl_st.png?​640}} ​+{{1920:​pred_pl_st.png?​640}} ​
  
 For a better comparison we use the difference: $d(m):​=\mathsf{rg}_{p}(m)-\mathsf{rg}_{38}(m)$ with $\mathsf{rg}_{p}(m)$ as the average prediction rank of team $m$ and $\mathsf{rg}_{38}(m)$ the rank of team $m$ after match day 38. This deviation from prediction, is a kind of measure, //how well they have done!// The larger $d$ the better they have performed, negativ $d$ means they have under performed, compared to the predictions. Sorting all teams accordingly from high $d$ to small $d$ is shown in the chart below. ​ For a better comparison we use the difference: $d(m):​=\mathsf{rg}_{p}(m)-\mathsf{rg}_{38}(m)$ with $\mathsf{rg}_{p}(m)$ as the average prediction rank of team $m$ and $\mathsf{rg}_{38}(m)$ the rank of team $m$ after match day 38. This deviation from prediction, is a kind of measure, //how well they have done!// The larger $d$ the better they have performed, negativ $d$ means they have under performed, compared to the predictions. Sorting all teams accordingly from high $d$ to small $d$ is shown in the chart below. ​
  
-{{1819:​pred_mw-st_pl.png?​640}}+{{1920:​pred_mw-st_pl.png?​640}}
  
 On top are those teams which have //over performed// and on the bottom those which have //under performed//​. ​ To judge whether it is really over- or under performed so far, those errors must be taken into account. All teams which are placed within this error bars, i.e. from 3.-17. rank, are roughly in the rang one could expect. On top are those teams which have //over performed// and on the bottom those which have //under performed//​. ​ To judge whether it is really over- or under performed so far, those errors must be taken into account. All teams which are placed within this error bars, i.e. from 3.-17. rank, are roughly in the rang one could expect.
  • Zuletzt geändert: 2019/08/10 06:41
  • von admin