In regular casual play, you can set the AI to auto-adjust and it'll start to favor plays that worked earlier in the game and avoid ones that failed. This hardly ever seems to work and on defense, using it is suicide.
Now, before I go into my idea in specifics, let me explain what casual means to me. It has nothing at all to do with a casual desire to win games and championships. It has to do with the casual nature in the blocks of time that I can commit to coming up with game plans. I can commit to 20-30 minutes of scouting per game required to do a good plan in casual, but the hours upon hours needed to create your first plan in regular, plus the extra time needed to adjust it to the teams you play is something I can't do.
So, to me, any suggestions for improving casual league play need to come from a perspective where it adds competitiveness and versatility to the game without adding a lot of time to the process.
The Idea
In addition to the current "Automatically Adjust" settings, also have a "Historical Weighted Play Selection". This is a list of all the games you have already played (basically, it's the same as the offensive/defensive scouting page, but without the results printout - just the game selection part). From there I can check any or none of the games to figure into my game plan.
Now, chances are that I haven't played this specific team before, but I probably have played a team that uses the same types of playcall strategies or that is built the same way (e.g. another team with a rushing QB will be a good one to check off on my defensive plan if this game has a rushing QB as well).
Currently, in casual, I can pick what playbooks to call from and how often, but the plays that get picked are generally random - and then I can use the auto-adjust to affect the chances of those plays being called again (or not) in the future.
With this idea, though, it could run scouting report type data from the past games that I've checked off as relevant and have the success percentage and average gain data as a sort of "seed" to the play selection process even before the game starts. As the season goes on and you can check more and more games with relevant matchups, the playcalling gets better because there are certain plays that "never" work - and some that work really well.
A system like this doesn't add any real complexity to the creation of gameplans, but in the long term over the course of each season (especially in playoffs) it would step up the competitiveness of the games because it's starting with some sort of a "seed" rather than starting with nothing.
When a situation comes up, rather than running the numbers to get a set of possible plays and then randomly picking one, it'll have a set of seed values that will roll the dice and favor the ones that worked before over the ones that failed before - but not just for this game. It'll do it based upon the games that I ticked off.
The basic system for this is most likely already in place with the auto-adjust. This just sort of carries it over a longer period of time (multiple games) thus creating more data, and thus creating more efficient play calling choices.
A bonus from this system would be a the possibility cool new tool for OCs and DCs in regular leagues. As they create new packages and AIs there could be an option to tick off the boxes in a similar way and import automatically (or just see effectiveness ratios on the same page - or both) to give yourself a starting package that is populated with the plays that worked and skipping the ones that didn't. Then you can manually go in and adjust and add/remove. Instead of starting from zero, you can get something that is "roughly" a good idea and edit from there.
Summary of Suggestion
=Tickboxes of all past games and scrimmages from the season to date in the casual gameplan. Coordinators can check off which past games to include in the calculations
= At Gametime, when the sim decides what play to use it generates it's pool of plays the same as it always has, but rather than having an equal chance of each play being called, it weights them based upon the scouting data (success rates and yards gained) collected at the beginning of the game from the checked off past games. (Basically it just expands the dataset used to adjust the gameplan and will favor successful plays and unsucessful ones from a larger dataset.
= If the current auto adjust is not selected, it works as above. If it is selected, it will figure in the current game data as well.
Summary of Regular League Tool
= This same idea would make a great tool for regular leagues to prepopulate playbooks and AI situations without needing to start at zero.
Now, before I go into my idea in specifics, let me explain what casual means to me. It has nothing at all to do with a casual desire to win games and championships. It has to do with the casual nature in the blocks of time that I can commit to coming up with game plans. I can commit to 20-30 minutes of scouting per game required to do a good plan in casual, but the hours upon hours needed to create your first plan in regular, plus the extra time needed to adjust it to the teams you play is something I can't do.
So, to me, any suggestions for improving casual league play need to come from a perspective where it adds competitiveness and versatility to the game without adding a lot of time to the process.
The Idea
In addition to the current "Automatically Adjust" settings, also have a "Historical Weighted Play Selection". This is a list of all the games you have already played (basically, it's the same as the offensive/defensive scouting page, but without the results printout - just the game selection part). From there I can check any or none of the games to figure into my game plan.
Now, chances are that I haven't played this specific team before, but I probably have played a team that uses the same types of playcall strategies or that is built the same way (e.g. another team with a rushing QB will be a good one to check off on my defensive plan if this game has a rushing QB as well).
Currently, in casual, I can pick what playbooks to call from and how often, but the plays that get picked are generally random - and then I can use the auto-adjust to affect the chances of those plays being called again (or not) in the future.
With this idea, though, it could run scouting report type data from the past games that I've checked off as relevant and have the success percentage and average gain data as a sort of "seed" to the play selection process even before the game starts. As the season goes on and you can check more and more games with relevant matchups, the playcalling gets better because there are certain plays that "never" work - and some that work really well.
A system like this doesn't add any real complexity to the creation of gameplans, but in the long term over the course of each season (especially in playoffs) it would step up the competitiveness of the games because it's starting with some sort of a "seed" rather than starting with nothing.
When a situation comes up, rather than running the numbers to get a set of possible plays and then randomly picking one, it'll have a set of seed values that will roll the dice and favor the ones that worked before over the ones that failed before - but not just for this game. It'll do it based upon the games that I ticked off.
The basic system for this is most likely already in place with the auto-adjust. This just sort of carries it over a longer period of time (multiple games) thus creating more data, and thus creating more efficient play calling choices.
A bonus from this system would be a the possibility cool new tool for OCs and DCs in regular leagues. As they create new packages and AIs there could be an option to tick off the boxes in a similar way and import automatically (or just see effectiveness ratios on the same page - or both) to give yourself a starting package that is populated with the plays that worked and skipping the ones that didn't. Then you can manually go in and adjust and add/remove. Instead of starting from zero, you can get something that is "roughly" a good idea and edit from there.
Summary of Suggestion
=Tickboxes of all past games and scrimmages from the season to date in the casual gameplan. Coordinators can check off which past games to include in the calculations
= At Gametime, when the sim decides what play to use it generates it's pool of plays the same as it always has, but rather than having an equal chance of each play being called, it weights them based upon the scouting data (success rates and yards gained) collected at the beginning of the game from the checked off past games. (Basically it just expands the dataset used to adjust the gameplan and will favor successful plays and unsucessful ones from a larger dataset.
= If the current auto adjust is not selected, it works as above. If it is selected, it will figure in the current game data as well.
Summary of Regular League Tool
= This same idea would make a great tool for regular leagues to prepopulate playbooks and AI situations without needing to start at zero.
Edited by Ali Khaman on Nov 21, 2011 02:02:45
Edited by Ali Khaman on Nov 20, 2011 10:48:25
