Quote:
Originally Posted by usa_hank
Pretty simple answer for this - the devs look at these forums. They could answer the question of whether there is a single or multiple hidden attributes that randomly varies performance in season to season (or over chunks of a season) for each player on each team. Or perhaps the first couple of games have a greater influence on the seasonal performance per player? It's not some secret sauce they would be exposing here.
In fact, you can call it conspiracy theory or not, but the fact that they haven't chimed in already tells me it's more than likely true. They could have easily debunked this by saying "nope" and that is that, and it immediately shuts down this conversation. I see them chime in on ephemeral subjects about mission names and other pointless details. But this forum, left untouched.
Of course, they don't have to chime in at all. That's their prerogative. Sometimes a lack of response is quietly a big response. You'd think if it wasn't true they'd try to quell that kind of discussion pretty quickly, as it kind of affects player opinions when a bunch of people are pointing at it and providing performance evidence of it. I certainly can point you to at least 30-40 player-seasons of performance from individual players on my team that shows something affected that player's performance which caused it to vary significantly year to year.
In statistics, outliers don't happen at that rate, and if they do something is wrong with the data or the sampling - which would point to the way the simulation is working. Hence, problem with the sim, not a good look. But if it's an intentional insertion of performance- and period-based random factor, that changes everything.
I'll have to consider a better way to analyze the data and find the right questions to ask. I think I'll start saving off the stats for every player on my team and every "same" player on other teams year to year, to start a true statistical analysis of this. I'm a nerd, I enjoy mathematical analysis of data. Heck, AI is a wonderful tool and could easily help here.
|
I'm a statistician: nothing about the things look wrong. You might disagree with how the distributions work, but there's nothing to lead me to believe that there's something unintended happening. It's entirely a sample size issue.
I think largely you just don't understand how normalization and league totals work, and how in small sample sizes (which are still thousands of PAs, especially with all the different confounding variables happening from game to game) you'll see a lot of difference. League would be far less fun if things were more consistent, and people would be complaining about always losing to the best players if that was the case