Good morning guys.
Just a heads up, I'll be pushing a release later on today to alleviate some of the main issues with the engine. It's not out there now, but I'll post again when the new version is available. I put on hold Injuries and Penalties. They will be coming soon, but I spent this last week focused on creating a more realistic distribution of game results.
I've been crunching away the last few days creating tools that help me determine the overall scoring and engine problems. One of the biggest issues I have been seeing is how quickly the scores increase as the average difference in players increases.
Here is a link to the full version
The Y axis is the number of points scored by the home team, and the X axis is the average DIFFERENCE in the player ratings per team. This is showing that if a team has any advantage the average scores increase quickly making the matches lopsided and the stats incredibly high.
I've made some tweaks to the passing, rushing, and overall matchup between the two teams to come up with a more realistic distributions of scoring per game.
As you can see, with these latest tweaks you may still get blowout games in an individual contest but in the aggregate things will follow a more realistic distribution. Point totals will come down, which will allow for a probabilistic chance that the better teams will win, but it will also reintroduce a chance for the opponent to win based on game planning, strategy, and of course luck.
Ok, a couple of more. Here are the before and after of Win Percentage of the Home team based on the same Rating difference scale.
Here we can see that the current distribution of win percentage will depend entirely on the difference in ratings. Right now, if you are better,even by a little, there is a great chance you are going to win, regardless of game plan and strategy. With the latest tweaks we can see that win percentage is still heavily weighted on talent, but it allows for the ability to create a better team with coaching. I feel that these are the general changes for which people are asking.
CAVEAT: I'm dealing entirely in the aggregate with these charts. Some specifics can be deduced, but these do not describe individual games. The jumps in the charts are coming from a relatively small sample size. The before charts were done over a 15 game average, and the afters over a 50 game average.
Also, overall score outputs will vary based on defense ratings vs offense ratings. I chose to create teams that have roughly the same ratings on offense and defense.