Timing the market is hard and any probability model is going to introduce some flaws along with some insight or trends. The game involves a lot of luck so paying attention to trends can be a way of improving your odds, recognizing that there will always be an exception to what usually happens. I think you look at each player separately as well. How do you see that player fitting into a career productivity curve? Knowing when most players at a position will peak can help inform decisions such as when to trade for or trade a player away. Who to draft and in what order (the rankings).
Instead of just throwing darts blindfolded though, we can look at things such as the teams recent performance to form some sense of a projection for a player in that offense, with that supporting cast. The opportunity and situation are a big part of this, so I see your opportunity category taking into consideration things like team stats and what a players role with their team is, or is projected to be.
I'm sure there are a thousand ways of improving my individual ratings. I think there is a ton of merit to what you are saying about looking at the team output and using those outcomes to drive a portion of an "opportunity" or "situation" metric for an individual. What is the 3 year average for WR1 on the Colts? What about WR2? I think a person can make some very informed adjustments to their ratings (whatever they are) using data driven partially by these numbers in conjuction with evaluations of the individual's overall talent. This would be particularly useful in evaluating players who are moving (either team-to-team or up the depth chart).
You are using a 1 to 10 rating of this (perhaps some other things, not sure what all goes into that for you) where I would prefer to use a 3 year average as the basis for a team projection, then player projections from that. Your method is a lot easier to complete than mine which is a process that for me is always in flux and I never really consider done. Projections will often be very wrong, but the process of making them helps me dig deeper into each team and hopefully gain some better understanding of which players are starters, which are back ups, which are prospects. The outcome helps me quantify the value of players which at least helps me identify players that are of similar value to each other, if my projections are not really off anyways.
On that point I usually like to project an upside and downside for players, which would fit into the upside only category you have, where I may draft a player based on their ceiling instead of the more conservative median range projection
Back when I was doing this, I was only playing in one league, so adding complexity resulted in diminishing returns. The work I was putting in was so far beyond what others in my league were doing (and the results pronounced) that I didn't need to take it much further. All of my 1-10 ratings were from my own gut and not drawn from any further analysis. I'm sure the outcomes would be improved the more real world numbers were used to derive any of the ratings. One nice thing about this particular system is that I didn't use "projections" at all when rating these players, though I'll admit that some form of projecting was taking place when evaluating talent and opportunity, the goal wasn't to determine who would score the most points in season N or N+1. FWIW, I think that ultimately strengthened the outcomes here in this particular methodology.
The talent category could be based on things like their draft position, college performance, combine data. What data would you use to measure talent for veteran players however? When these things perhaps become meaningless to how the player has developed and performed in the NFL since being drafted, or do we just carry those data points forward for the players career? How are you measuring talent?
There is one philosophy that the creme rises to the top, so a completely talent based ranking should be quite a bit different than a opportunity or situation based ranking. The perspective of the talent based ranking does sort out starters from back ups in the sense that if you had player A and player B on the same team, which of those players is the most talented of the two? Which player will earn more playing time than the other, just based on the relative talent level. Some times this is very clear, often times it is not, as players competing for playing time may be of similar talent or ability. Some players have holes in their skill set and only suited for specific roles or for specific types of offenses.
For rookie players I am often focused on the talent based rankings over situation, at least until those players have been in the league a few seasons (like 3) when I am mostly only focused on what they have done in the NFL so far.
At the time, I was measuring talent with my eyes on most everything you've mentioned above. Data for veteran players would take into everything previously mentioned, plus a small adjustment based on production to date or what I'd seen with my eyes against NFL level talent. This piece was done in my head, but the one rule that I always kept in mind was that talent doesn't "go away". Opportunity changes, but talent doesn't change much without injury or something else substantial. So I generally tried not to adjust the talent ratings by much, regardless of performance. If someone wasn't living up to their potential it was likely their surroundings or something besides talent holding them back, so opportunity was far more likely to get adjusted before talent.
You are 100% correct about a talent ranking being different than an opportunity-based ranking. The trick would be to have a way of weighting them so that they make sense - then we'd have the ability to pre-emptively select a Tyreek Hill over a Chris Conley (but don't sleep on Conley now!) because maybe if we have talent and opportunity balanced just right, one of them would have popped out differently during the offseason. No system is ever going to be 100% accurate, of course, but all we really need is something that is more effective than what our opponents are doing. I pretty much agree that after some time in the NFL it may be effective to focus on "what have you done" more so than any original talent evaluation, though I would argue that if you're doing a good job maintaining your opportunity side of things, the player will drop without much adjustment to talent since the NFL isn't going to give too many opportunities to someone who's done nothing in 3-4 years. In fact, letting the opportunity rating drive this more so than dropping your talent rating, might allow you to better identify late bloomers through a position change or team change, whereas if you're dropping their talent along with opportunity, they might get lost in the crowd of mediocrity.
In any case, I think if you're managing your team along these three roles (starter, backup, and developmental players) you're going to outperform your opponents long term, simply due to more effective roster management - even if the methodology when rating those players within those roles is imperfect.
An exception to this for me would be Sammy Watkins, who I am still very high on based on pre NFL draft evaluation and what I have seen from him at times in the NFL. He hasn't produced as well as expected. Part of that is because of injuries, which can happen to anyone, part of it is because of the situation with Rex Ryan conservative run based offense. Watkins has shown me enough to think he will reach his potential, but in retrospect I likely should have downgraded him more than I did based on opportunity or his situation in Buffalo. It has been 3 seasons, maybe I should let the early evaluation go. However Watkins is only 24 years old. Rex is gone some hope of an offense that might be friendlier to Watkins, and if he stays healthy I could still see him performing at a WR 1 level in 2017 and make everyone forget about the missed games. This is more of a gut call on my part based on the talent of Sammy Watkins, rather than sticking to a methodology which I know will tell me to downgrade Watkins based on what he has done so far compared to his peers.
I love this and he's a perfect example of why I think that any talent rating should remain more static than any opportunity rating. Remember when Moss went to the Raiders? His talent didn't go anywhere, his opportunity tanked due to surrounding cast and attitude. Opportunity went back up as the #1 WR for the Pats and if we kept talent at appropriate levels, Moss would have jumped up pretty high
as a backup in my ratings. In fact, I think I had him as my #1 rated backup that first season. Imagine if Watkins went to the Packers.
