Quote: Originally Posted By coach_billyg on 12/28/2009

the 2 biggest flaws:
1) as many have stated, how good a team could be if played a certain way should have no impact on tournament seeding. Agreed, I never said it should have, but to say that OTRs aren't looked at during NT selection without PROOF is unproven. I never talked about how good a team could be if they played a certain way for what its worth either. only how well they played. this is an opinion, but i believe it is a majority one.

2) "An adjusted overall team rating as I've suggested shows the true quality/talent of a team from an HD engine perspective" - you nor any other coach can come up with such a rating. I can come up with that rating, its not hard. You guys see obstacles, I see solutions.your argument is equivalent to me saying, if i had a magic want that could cure cancer, i could cure cancer. well great for me. There's quite a gap between CURING CANCER and adjusting HD OTRs to make them better...yet another ridiculously extreme example/comparison that my haters purport to "put me in my place" it doesn't mean i could actually do it. nor can you come up with a SOS system based on ranking that is truly a quality system. That's your closed-minded opinion. You guys are all stuck in the context of real life and you can't see beyond that....that's what's happening here.

Quote: Originally Posted By fd343ny on 12/28/2009



fd notes - one could go with an average, but as you note that will be complicated - AND averaging the ratings obscures the benefits of the combos - I dont think there is a simple, easy and reasonably accurate answer here

Perhaps just go with each player's highest IQ...there are simple solutions here that work, such as averaging as well. The system doesn't have to be overly complex to be extremely effective and efficient. Those that seek solutions will find them...those that don't think solutions can be found, won't.

fd notes - "his ratings are what they are" misses the point - ratings matter in different ways at different positions and in different systems - as others have explained repeatedly in this thread - I understand that you think ratings are what you are trying to measure because you think ratings="strength" - this is one very clear example of where that theory fails...the glib answer "his ratings are what they are" simply assumes the conclusion desired to be analyzed - which may be fine if what you mean by ratings based rankings is a ranking of team ratings

For ranking purposes, where a player plays, whether a player plays out of position or not DOESN'T MATTER. If his ratings change, what are they? Like I said, you can't measure every last little thing that happens in a game and factor it in to a ranking system...you're guessing at coach's intent, the correctness or not of the decision, etc, and that frankly doesn't belong in a ranking system. This is one of those "THEMS THE BREAKS" situations chalked up due to everyday sports happenings. Am I supposed to go through and find all of the human coaches on hiatus too? Those that missed a game or 2 because they were on vacation? Also, this is the 3rd time at least I'm asking this...why should I make ANY rating adjustment to a 12 man SF roster when they can win titles? It would only dumb their OTR down, and incorrectly portray the team as less talented when they really are. Surprisingly, nobody has addressed that point.

fd notes - weighting the categories is an interesting possiblity - there are TONS of other threads - including Z's excellent guide for new coaches - that qualitatively offer guidance on this - but any weighting effort will be of only partial effectiveness given the further differential effects of usage and system....the glib answer "weight the categories" is like waving a magic wand to change dross into gold - wont work

If you could link me to those threads, I'll take a look. Personally, I am not weighting the categories, I don't really think its necessary, but like I said if someone has a good system out there and can convince me that its worthwhile, then I'm in...but I'm not just going to do something just because the masses say I'm wrong because we have a difference of opinion. As I said earlier, this is not a group project...this is my ranking system, if you don't like it, don't pay attention or start your own....or better yet, start answering MY questions.

fd notes - the large number of micro questions arises from the desire to rank based on inputs to the game engine, rather than using the outputs from the game engines - wins, losses, margins of victory and the similar outputs for a teams opponents and their opponents. Sagarin doesnt analyze the strength of the backup point guard at Florida Atlantic, yet his methodology produces a useful ranking

I am not ranking based on OTR...OTR is the SOS FOR MY RANKING and that means that WINS AND LOSSES still matter more than what your opponents' OTRs are. Last page or a couple pages back I showed you basically how the points would shake out, please take a look. And the Sagarin thing doesn't mean crap here...that's a real life ranking and this isn't. Real human beings don't have concrete numerical ratings like sim players do, thus even if Sags wanted to analyze the strength of FAU's backup point guard, he couldn't do it on anything more than subjection and conjecture. Lastly, you've never seen what my rankings have output, but you've already got your stamp of disapproval on it...sorry, but the only word that pops into my mind is closed-minded.

Quote: Originally Posted By colonels19 on 12/29/2009

Quote: Originally Posted By fd343ny on 12/28/2009




fd notes - one could go with an average, but as you note that will be complicated - AND averaging the ratings obscures the benefits of the combos - I dont think there is a simple, easy and reasonably accurate answer here

Perhaps just go with each player's highest IQ...there are simple solutions here that work, such as averaging as well. The system doesn't have to be overly complex to be extremely effective and efficient. Those that seek solutions will find them...those that don't think solutions can be found, won't.

new text - I like the each player's highest rating answer - a pretty good solution, imperfect but likely close enough for government work and hand grenades

fd notes - "his ratings are what they are" misses the point - ratings matter in different ways at different positions and in different systems - as others have explained repeatedly in this thread - I understand that you think ratings are what you are trying to measure because you think ratings="strength" - this is one very clear example of where that theory fails...the glib answer "his ratings are what they are" simply assumes the conclusion desired to be analyzed - which may be fine if what you mean by ratings based rankings is a ranking of team ratings

For ranking purposes, where a player plays, whether a player plays out of position or not DOESN'T MATTER. If his ratings change, what are they? Like I said, you can't measure every last little thing that happens in a game and factor it in to a ranking system...you're guessing at coach's intent, the correctness or not of the decision, etc, and that frankly doesn't belong in a ranking system. This is one of those "THEMS THE BREAKS" situations chalked up due to everyday sports happenings. Am I supposed to go through and find all of the human coaches on hiatus too? Those that missed a game or 2 because they were on vacation? Also, this is the 3rd time at least I'm asking this...why should I make ANY rating adjustment to a 12 man SF roster when they can win titles? It would only dumb their OTR down, and incorrectly portray the team as less talented when they really are. Surprisingly, nobody has addressed that point.

new notes

a. saying in all caps that where a player plays "doesnt matter", does not make it so - as discussed, different attributes have different effects in different positions, which means that a set of numbers associated with a player who is - say - a SG - but is used at PG will have different effects in the engine than those numbers if he is used at SF. If the goal is to use ratings, adjusted ratings or the like to measure team strength the ratings have to be analyzed on a position of use dependent basis.

b. this same issue applies to an all SF or all anything team (each of which has been done, and no I'm not going back to find them). The well constructed all SF team - for example - had some SFs with guard type skills and some with big skills. One could say these effects are too small to matter - people believe they matter based on what it takes to win games. They could be wrong of course. I'm not sure that I see the negative effect of adjustments.

fd notes - weighting the categories is an interesting possiblity - there are TONS of other threads - including Z's excellent guide for new coaches - that qualitatively offer guidance on this - but any weighting effort will be of only partial effectiveness given the further differential effects of usage and system....the glib answer "weight the categories" is like waving a magic wand to change dross into gold - wont work

If you could link me to those threads, I'll take a look. Personally, I am not weighting the categories, I don't really think its necessary, but like I said if someone has a good system out there and can convince me that its worthwhile, then I'm in...but I'm not just going to do something just because the masses say I'm wrong because we have a difference of opinion. As I said earlier, this is not a group project...this is my ranking system, if you don't like it, don't pay attention or start your own....or better yet, start answering MY questions (hostility noted).

new notes

some good threads are pinned at the top of the forum - see especially Z's guide - you are just as capable as I am of scrolling through the forum - select a wider view than the last few weeks and read away

fd notes - the large number of micro questions arises from the desire to rank based on inputs to the game engine, rather than using the outputs from the game engines - wins, losses, margins of victory and the similar outputs for a teams opponents and their opponents. Sagarin doesnt analyze the strength of the backup point guard at Florida Atlantic, yet his methodology produces a useful ranking

I am not ranking based on OTR...OTR is the SOS FOR MY RANKING and that means that WINS AND LOSSES still matter more than what your opponents' OTRs are. Last page or a couple pages back I showed you basically how the points would shake out, please take a look. And the Sagarin thing doesn't mean crap here...that's a real life ranking and this isn't. Real human beings don't have concrete numerical ratings like sim players do, thus even if Sags wanted to analyze the strength of FAU's backup point guard, he couldn't do it on anything more than subjection and conjecture. Lastly, you've never seen what my rankings have output, but you've already got your stamp of disapproval on it...sorry, but the only word that pops into my mind is closed-minded (not nice)

new notes - Sagarin offered as an analogy - personally, I would favor a Sagarin style, output based ranking system, which is why I mention it - please dont be closed minded about my opinion - and please dont call me closed minded when I offer ideas, praise some of what you suggest and criticize other elements of your concept. I dont like your concept because I think it rests on fundamental flaws, but I have posted before that I think it would be interesting to see a sample for one division in one world

When will you post one division in one world's results?

Quote: Originally Posted By fd343ny on 12/29/2009



new text - I like the each player's highest rating answer - a pretty good solution, imperfect but likely close enough for government work and hand grenades

People can take this the wrong way if they choose, but close enough is good enough for me in this instance. I don't want to get caught up in miniutiae and endless arguments/debates with this IQ thing, which I am aware is incredibly ironic for me to be stating.

new notesa. saying in all caps that where a player plays "doesnt matter", does not make it so - as discussed, different attributes have different effects in different positions, which means that a set of numbers associated with a player who is - say - a SG - but is used at PG will have different effects in the engine than those numbers if he is used at SF. If the goal is to use ratings, adjusted ratings or the like to measure team strength the ratings have to be analyzed on a position of use dependent basis.

It was for emphasis. It does matter in the context of the game, I agree with that, but I don't think it does for the context of basing a SOS rating off of OTRs. The ratings don't visibly change and that's my claim here...of course different ratings matter differently at each position, but to judge/guess why a coach is playing a player OOP if its good, or bad, is a terrible conundrum to get into to begin with. If a guy wants to play a PG as a C or an SF as a PG, etc, or if his players have Misfit Toy-like ratings for their listed positions (PG with high LP) then thems the breaks. A coach's intelligence/idiocy doesn't change the ratings.

b. this same issue applies to an all SF or all anything team (each of which has been done, and no I'm not going back to find them). The well constructed all SF team - for example - had some SFs with guard type skills and some with big skills. One could say these effects are too small to matter - people believe they matter based on what it takes to win games. They could be wrong of course. I'm not sure that I see the negative effect of adjustments.

But from what I've gathered thus far, is that if teams have players playing OOP, then somehow I should account for that, and the only logical accounting of this would be to somehow downgrade the player's ratings because he's not playing in his designated position....and in response to this, I'm logically asking..."Why would I want to do this if teams with 12 SFs etc can and have won championships before?" By downgrading player ratings in this instance, I'm misrepresenting the team's talent, because the team isn't as bad as the OOP adjusted ratings would suggest. So what's the point of penalizing a player for playing OOP if he plays better at that other position? I say leave the player ratings as is, and don't get caught up in all this nonsense...that is my solution. If coaches want to be boobs or the smartest guy in the room in doing these things, then fine, but that doesn't mean that my ranking system has to or should account for these kinds of happenings/judgments. I hope this clarifies my stance in this part of the argument.

new notes some good threads are pinned at the top of the forum - see especially Z's guide - you are just as capable as I am of scrolling through the forum - select a wider view than the last few weeks and read away

Fair enough, but if I still need convincing (which I do) then what am I necessarily going to gather from these threads that makes me think "hey I should use this?" Let's face it FD...what I'm asking people to do isn't tough...I'm asking the coaches that think its worth while to weight the non WE-ST-DU player rating categories as they see fit...I don't really see how that's beyond at most, a 15 minute exercise. I personally have no interest in doing this and am admittedly not qualified to do so.

new notes - Sagarin offered as an analogy - personally, I would favor a Sagarin style, output based ranking system, which is why I mention it - please dont be closed minded about my opinion - and please dont call me closed minded when I offer ideas, praise some of what you suggest and criticize other elements of your concept. I dont like your concept because I think it rests on fundamental flaws, but I have posted before that I think it would be interesting to see a sample for one division in one world

You were attempting to show Sagarin's ratings in the context of the sim when his ratings aren't taking into account the concrete measurable values of human beings because their aren't any. It was a complete apples to oranges scenario and it virtually had no effect here. My adjusted OTR would account for player minutes and usage, thus that FAU backup PG would be weighted and accounted for accordingly. I apologize for the closed-minded comment...the way that I read your post I saw it to say that you were casting it aside before you even see the output, like some have already done.

When will you post one division in one world's results?

Not sure...not even sure its going to be that in depth...I would need a lot of info from WIS and I've yet to contact them. I will be doing rankings for my team and for others that would like to participate, its just impossible to accomplish without some outside assistance, given that close to 1000 games are run a day in the 2 game/day worlds and that the overalls aren't openly listed like W-L, RPI, and SOS are.

My suggestion to you and to everyone out there is that if you really want to see how my 3 rankings will turn out for HD, then you'll do so by participating and submitting the information per how I asked. Thank you for the post.

Quote: Originally Posted By colonels19 on 12/21/2009
I think in games, where team ratings are so incredibly concrete (unlike real life which is a bit of a toss up...more reliance on W-L and perception), that your opponents team ratings should be the main basis for a ranking system.

Personally, I think seble is overthinking the process. Its funny that something I FIRST griped about 5.5 years ago is just now finally getting attention.