Visualising a crude measure of personal skill using historical data:
Plotting each hero on a chart (X-axis games played; Y-axis Win Rate), a personal skill level can be visualised.
Theoretically, when you start playing a hero, it will start as a point on the left side of the chart. It moves towards the right as more games are played. Assuming each person has his own personal long-term win rate (at a given point in time), a simple linear regression line should be able to capture a person's Win Rate Progression. This is because as a person plays more games (regardless of heroes), it contributes to his overall win rate, which becomes more fine-tuned the more games he plays.
A decent player will exhibit an increasing win rate as more games are played. This will cause his linear regression line (in gold) to be sloped upwards. Using this concept, the line is a simple (albeit crude) way to instantly assess a player's base skill (Y-intercept), and his improvement rate (gradient of the line).
An experienced player (having played Dota 1) might exhibit an almost flat gradient because his 'new heroes' aren't really that new to him, therefore as he plays more of that hero, his win rate for that hero might not necessarily increase.
Visualising a crude measure of personal skill using historical data:
Plotting each hero on a chart (X-axis games played; Y-axis Win Rate), a personal skill level can be visualised.
Theoretically, when you start playing a hero, it will start as a point on the left side of the chart. It moves towards the right as more games are played. Assuming each person has his own personal long-term win rate (at a given point in time), a simple linear regression line should be able to capture a person's Win Rate Progression. This is because as a person plays more games (regardless of heroes), it contributes to his overall win rate, which becomes more fine-tuned the more games he plays.
<a href="http://img.photobucket.com/albums/v650/Brimspark/01_zps772c4585.png" target="_blank"><img src="http://img.photobucket.com/albums/v650/Brimspark/01_zps772c4585.png" height="800" width="665"></a>
A decent player will exhibit an increasing win rate as more games are played. This will cause his linear regression line (in gold) to be sloped upwards. Using this concept, the line is a simple (albeit crude) way to instantly assess a player's base skill (Y-intercept), and his improvement rate (gradient of the line).
An experienced player (having played Dota 1) might exhibit an almost flat gradient because his 'new heroes' aren't really that new to him, therefore as he plays more of that hero, his win rate for that hero might not necessarily increase.
<a href="http://img.photobucket.com/albums/v650/Brimspark/02_zpsed49703a.png" target="_blank"><img src="http://img.photobucket.com/albums/v650/Brimspark/02_zpsed49703a.png" height="800" width="665"></a>
Another way is to plot KDA on the Y-axis if you feel that win rate is not representative. Or you could consider both charts together.
<a href="http://img.photobucket.com/albums/v650/Brimspark/03_zpsf31e3959.png" target="_blank"><img src="http://img.photobucket.com/albums/v650/Brimspark/03_zpsf31e3959.png" height="800" width="665"></a>
Also, a time series of these charts might provide some info on the evolution of a player.
<a href="http://img.photobucket.com/albums/v650/Brimspark/04_zps71b8b4a4.png" target="_blank"><img src="http://img.photobucket.com/albums/v650/Brimspark/04_zps71b8b4a4.png" height="800" width="665"></a>