|
I've become increasingly bothered by the predictability of the development model. Because Potential ratings act as targets for Overall ratings, if you have a 25 year old player, and his ratings start to drop (after injury, say), you can be pretty sure his Potential is less than his current ratings, and you can expect him to continue to get worse. I'd need to think more about whether it would be a good idea, but perhaps overall ratings should change more suddenly, rather than trending gradually.
More importantly, the AI does not value declining players correctly - the AI doesn't seem to notice that a player is declining, and will happily trade the farm for a good player who is dropping precipitously, or give a contract extension to a player who is likely to be worthless in a year's time. The problem seems to be that the AI treats current ratings as static entities - the AI will give a massive contract to a 28 year old with 80/80/80 ratings, expecting him to be a superstar for a while, even if the player had 100/100/100 ratings a year ago. If the AI were to project trend lines based on past ratings, it could more appropriately value declining and aging players - the AI might anticipate, in the example above, that the player could drop up to 20 points over the next year, and trade or offer contracts accordingly.
|