All Forums > Gridiron Dynasty Football > Gridiron Dynasty > Disparity between GUESS reports and performance
3/30/2013 12:26 PM
I recently picked up Southern University in D1aa Wilkinson..SU has had good recent success and had a good coach. They went 15-2 last year and ended up ranked #7. Beat some good teams.

Ranked #88 last season overall in GUESS.

When I was in D3/D2 my teams always finished higher than their GUESS ratings would predict, I think primarily because I also use my own system and put a premium on potential, but they never were that low and finished that well.

Love to hear from others who have teams that show this kind of disparity and theories on why they are successful despite low GUESS ratings.
3/30/2013 2:05 PM (edited)
No news here, but a few points to make:
1. As you mentioned guess doesn't include potential at all, so if you value this, then your roster will always be slightly underrated, particularly as the calculations are only make at the beginning of the season, and not updated throughout.
2. It only ranks your roster ratings, not how you use it, so if you're running a good game plan, your players can outperform their ratings.  In other words, maybe its just that you're a good coach!
3. I have a working hypothesis myself that GUESS overvalues GI/Tech for some players, but I'd be interested to hear others' thoughts on this as well....

3/30/2013 4:26 PM
Yes, I had a rant about this about 8 months ago and got hammered on the forums as people got upset. But My "terrible recruiting class at D-3 Willamette in Stagg ended up being top 20 in overall team ratings by the time the guys were seniors so a lot of the growth at D-3 and D-2 is hard to predict for Guess Ratings. Because of potential and work ethic. But Guess is still very fun to look at but it's definitely not everything! 
3/30/2013 5:51 PM
Guess ratings are a fun part of GD.  That being said, they are not that accurate in several cases.  Secondary attributes can really add up and make a player with a horrible core attribute a highly ranked player.  Unless he went and looked at every player and manually corrected their rating when this happened, you just have to accept it this way.  It's really pretty good overall but it's just one of many tools you should look at when recruiting.
 Personally I think core ratings should count for about 80% then the secondary all added up to about 20%.  Great secondary attributes will never make a bad player good.  They will make a good player great though.

3/30/2013 6:47 PM
. And I think we are all in agreement that Bob Hazlewood enriches GD! I wish he had GUESS for Hoops Dynasty!
3/31/2013 3:06 AM
I love the GUESS Ratings. However, I also see them as very overvalued. They are the foundation of my scouting, but not as a stand-alone perspective.

You also need to fit players to your system, IMO, as well as spot outliers.
3/31/2013 5:39 AM
As stated previously, GUESS doesn't include potential, so recruiting classes aren't very accurate; you can easily have the top class season after season, but never have the top team. Also, individual position scores aren't comparable (a top LB will have a score lower than a top DL), so it's skewed towards teams that have more high-scoring positions. Finally, I think that it under-values starters, because if you have two QB with the same GUESS rating, they're worth the same even though only one starts.
Also, GUESS can't take into account how you use your players. A top RB won't help a shotgun team much, and it really sucks to have a bunch of great ILBs and end up running nickel and dime defenses all season.
That said, it's amazing how many times top GUESS teams go deep into the playoffs, compared to lower ranked teams.

3/31/2013 8:44 AM
Agree with opinions already on this post, but having done a small amount of research you don't often see teams outside the top 20 of the Full Roster Guess reports making it into the semi-finals, certainly at divs 2&3. My own opinion is that the Guess ratings do give a good guide to a players effectiveness in his position, if you get 50 players with good guess reports you should have a talented team. How you use them and how you game plan will make the difference between a team of "Almost made it" To a team "That did". I can't see a team well down the full roster guide winning a NC. I may be wrong, I'm sure someone will tell me.
3/31/2013 9:19 PM
Posted by graham1958 on 3/31/2013 8:44:00 AM (view original):
Agree with opinions already on this post, but having done a small amount of research you don't often see teams outside the top 20 of the Full Roster Guess reports making it into the semi-finals, certainly at divs 2&3. My own opinion is that the Guess ratings do give a good guide to a players effectiveness in his position, if you get 50 players with good guess reports you should have a talented team. How you use them and how you game plan will make the difference between a team of "Almost made it" To a team "That did". I can't see a team well down the full roster guide winning a NC. I may be wrong, I'm sure someone will tell me.
I think this is basically right.  Talent is king and the GUESS ratings do reflect that talent level on each team.  Regardless of potential, the #30 GUESS rated team is not going to become the #5 GUESS rated team within the course of a specific season.

Teams that recruit a lot of RBs can artificially inflate their overall GUESS rating because RBs tend to have the higher ratings from the start.  I think that overall GUESS ratings can be looked at in groups of maybe 20 teams.  The top 20 overall can compete for national title and certainly should make the playoffs.  The next 21-40 are bubble teams, talent-wise, and may or may not make the playoffs, but probably don't have enough talent to make a deep run.  After that its just decreasing "levels" of talent.

However, the best way IMO to get high overall GUESS ratings is to recruit the best CORES you can get that FIT YOUR SYSTEM.  If you go and recruit the highest rated players, they probably won't fit your system and will not perform as you might like.  Example:  a D3 OL with 75 GI, but only 30 BLK is probably not worth picking up, even though his overall GUESS rating might be pretty high.  His BLK just won't ever get good enough.  Especially when you then consider WE and potential.  
4/2/2013 1:52 PM
Posted by harriswb3 on 3/31/2013 9:19:00 PM (view original):
Posted by graham1958 on 3/31/2013 8:44:00 AM (view original):
Agree with opinions already on this post, but having done a small amount of research you don't often see teams outside the top 20 of the Full Roster Guess reports making it into the semi-finals, certainly at divs 2&3. My own opinion is that the Guess ratings do give a good guide to a players effectiveness in his position, if you get 50 players with good guess reports you should have a talented team. How you use them and how you game plan will make the difference between a team of "Almost made it" To a team "That did". I can't see a team well down the full roster guide winning a NC. I may be wrong, I'm sure someone will tell me.
I think this is basically right.  Talent is king and the GUESS ratings do reflect that talent level on each team.  Regardless of potential, the #30 GUESS rated team is not going to become the #5 GUESS rated team within the course of a specific season.

Teams that recruit a lot of RBs can artificially inflate their overall GUESS rating because RBs tend to have the higher ratings from the start.  I think that overall GUESS ratings can be looked at in groups of maybe 20 teams.  The top 20 overall can compete for national title and certainly should make the playoffs.  The next 21-40 are bubble teams, talent-wise, and may or may not make the playoffs, but probably don't have enough talent to make a deep run.  After that its just decreasing "levels" of talent.

However, the best way IMO to get high overall GUESS ratings is to recruit the best CORES you can get that FIT YOUR SYSTEM.  If you go and recruit the highest rated players, they probably won't fit your system and will not perform as you might like.  Example:  a D3 OL with 75 GI, but only 30 BLK is probably not worth picking up, even though his overall GUESS rating might be pretty high.  His BLK just won't ever get good enough.  Especially when you then consider WE and potential.  
Since OL stats are combined, this example could either be a good player or a bad player, depending on the stats of the rest of the line. If he's in a line with really high BLK but low GI, he might be just the piece of the puzzle you need. In a line with weak BLK, though, he'd be poison.

It does, however, reinforce the point that individual players win GUESS rankings, but TEAMS win NCs. Strange, but that might be the most realistic aspect of GD.

4/3/2013 1:51 PM
Just a few points to make, some of which have already been made before in this thread:

a) I don't have access to Potential, your Practice Plans, or Game Plans.  I *can't* rate your recruit class on how it's going to look in 4 seasons.  I tried for a while, assuming average potential and average player growth, based on a formula used by some extremely good coaches.  All it did was overly complicate and confuse people, because I couldn't publish that part of the rankings formula.

b) Positions don't all rate equally.  QB's will kill your ratings.  So will K and P.  I've tried a number of different ways to equalize those over the years, but none has been satisfactory.   I think it actually became worse with the addition of DE/DT, ILB/OLB, and S/CB.  

c) Team ratings DO value starters more than bench players.  It's not just an average of all your player ratings.  I'd have to go double-check to be sure, but I think starters (the top players rated at their position) get a 25% bonus added to their impact on team ratings.

d) The current formulas are based on what JConte told me directly was important to the game engine.  I think it's pretty clear now, based on the things Norbert discovered as he started re-coding the engine, that many of our assumptions and things we were told were less than fully accurate.  Frankly, it's amazing that the ratings perform as well as they do.

e) I'm always willing to listen to ideas on how to improve the formulas.  They are, after all, just guesses to begin with.  
 

4/3/2013 4:35 PM (edited)
I have no problums with your ratings at all.  If something didn't match with a #1 player it is only because I will use this player to do some thing else
.  Example would be my TEs.  My best TE is not the one that I start as my #2 is better a RTE compaired to #1. I start serten OLs over others as I like them better as a Pass BLK over a #1. The guess ratings are not F up. They are just basic and is to be used as such.  maddiesdad spread sheet is the same way only a little out of date.


4/4/2013 5:57 PM

Just my opinion,Guess Reports is for entertainment value and helpful for setting depth charts.  YATR is where the money is made.
4/17/2013 3:21 PM

I find the Guess reports to be pretty darn accurate in addition to the most anticipated thing in GD.  It isn't my only tool but it darn sure helps.  I haven't won the big one yet, but that's my two cents worth.  One a side note, I can finally see that the Ol' BluGolds (yep dumbest nickname ever) are elite.  All trhe recruits I go after are taken by Champ, Myers, Harris, BigAlric, Dev, etc.   A few seasons ago, I never even bumped into them on accident.  GUESS helped with that!  PS.  LOL Still can't beat them.

4/19/2013 1:35 AM
I wish Bob would do GUESS for HD!
of 2
All Forums > Gridiron Dynasty Football > Gridiron Dynasty > Disparity between GUESS reports and performance

Terms of Use Customer Support Privacy Statement

Popular on WhatIfSports site: Baseball Simulation | College Basketball Game | College Football Game | Online Baseball Game | Hockey Simulation | NFL Picks | College Football Picks | Sports Games

© 1999-2014 WhatIfSports.com, Inc. All rights reserved. WhatIfSports is a trademark of WhatIfSports.com, Inc. SimLeague, SimMatchup and iSimNow are trademarks or registered trademarks of Electronic Arts, Inc. Used under license. The names of actual companies and products mentioned herein may be the trademarks of their respective owners.