Posted by hughesjr on 11/23/2016 6:53:00 PM (view original):
Posted by rednu on 11/23/2016 6:45:00 PM (view original):
Posted by hughesjr on 11/23/2016 6:35:00 PM (view original):
Posted by jcfreder on 11/23/2016 4:57:00 PM (view original):
If the overall grades are not going to be even plausibly rational, then they should not be included in the game.
The have been included since the beginning of 1.0 .. and they have NEVER been accurate. Overall means nothing, because different attributes mean different things based on position, etc. So 667 current overall score means absolutely nothing. It also stands to reason, then that a 667 overall score converted to a letter grade also means absolutely nothing.
When you take into account that it also does not take into account potential .. well .. it is worth even less. Unless you use player roles and you define what you think is important for each position. Then the overall grade can mean something as a current score .. but since it does not include potential, it only means something when compared against other freshmen. Because, comparing the current score of a freshman against the current score of a senior (who has had 4 years of growth) is kind of pointless. Now, comparing a Freshman to that senior WHEN he was a freshmen as some merit.
But that 667 HAD value -- it told you the starting point of the recruit. It might not have told you how good the player would be, but it told you the EXACT starting position. If I was a D1 or a D2 or a D3 program, I could scan that and find players that were "about" my divisional level.
A grade of C, at present, tells nothing. Which leaves it in stark contrast to every other item out there. A 'C' in rebounding at Level 2, for instance, means that the player has a minimum of 42 and a maximum of 57 in that category. That's information. It's vague information, but useful.
An Overall score of C, by contrast, means nothing. It doesn't identify that the player lies within a given range. It's based on a "proprietary ratings formula" (seble's words in beta when we had this conversation) that is utterly meaningless to the end user because it is based on an unknown formula that WIS hasn't disclosed, is based on the player's listed position (rather than one said player's ratings better align to), and which clearly, given the distribution pattern of ratings, wasn't remotely calibrated well when placed into use.
It is NOT true that C tells us nothing.
It tells you one of two things, depending on whether you use 'player roles' or not.
We may not know what it means .. but there is a direct numerical range involved. Either a player role grade (like 45.6 or 57.4, etc) .. or an overall score.
You could likely very easily figure this out if you want to.
When players sign, the EXACT numbers and potentials show up in the recruit pool. If you saved the initial players you have scouted at level 1 into a spreadsheet (with their letter grades), you could then on the last day of recruiting save all the level 1 players exact attributes. You could then figure out the exact range for each of those grades. Certainly if you did for more than a couple seasons, you could easily get a range.
WIS did not tell us the range of each color for potential .. people figured out what they are and posted it here .. then others verified it.
Unless you're in possession of some insider knowledge, I'm not sure how you can definitively say there's a "direct numerical range" involved. I am completely unaware of that explanation ever being provided, regardless of how logical it may seem. That said, let's consider the following:
1. That range absolutely does not correspond to the recruit's starting overall rating score. Based simply on data that I have, the C value runs from at least 617 on the top end to 432 on the lower end. It is overlapped at least 50 points on the high side by the B rating and by at least 30 points on the low end by the D rating. Clearly, if there is a fixed numerical value, it is NOT based off of overall score. Which leads us to...
2. The player role grade. We were told by seble that the "proprietary ratings formula" was "like player roles" but "not exactly the player roles" defaults that players have for each position.Here's the problem -- if that is how the rating is determined, then that formula only takes into account the player's listed position, meaning there is not ONE formula determining the overall grades, but actually FIVE separate ones whose data is being lumped together in one output field. Which actually could explain the FUBAR nature of the grade, but is hardly something to be defending or thinking of as good for the playability of the game.
Think of how confusing and chaotic it would be if PG and PF rebounding grades meant two different things. Sure, we could reverse-engineer things and come up with a master list, but that takes time, which in turn impedes game playability for the mass number of users, particularly the casual and/or new users who don't inhabit the forums to access such insider knowledge (in much the same way that many people didn't know you could safely zero out practice time in maxed out categories...some still don't perhaps) and upon whom the viability of the game depends.
BUT EVEN IF THERE IS a fixed numeric range tied to the evaluation produced by this unknown formula, I assert that it is functionally useless because the calibration of the scale converting numbers to letters is hopelessly off in its present form. The evidence is right in front of us in the sheer volume of players that return the same grade. The same "range" is capturing players projected from the Top 100 down to upper level D3. Even if you're comfortable with the level of vagueness which the mask produces, the FREQUENCY with which one SPECIFIC result gets created has to be a harbinger of a problem.
I honestly don't know what overall percentage of the generated recruiting class receives a C Overall evaluation, but if I had to wager money, I'd say probably just shy of 1/2 (45%?) with maybe around 1/3 of the class getting D's and the rest sprinkled between A's and B's (and F's? Do we have an F overall for generated players?). Isn't that something that we are interested in monitoring and seeing addressed as players? As game designers?