Major issues with new polls (with long-term dynasties)...

[JJSmitty]

[QUOTE=TPBPlayer]The polls are killing me right now, but the other changes are great. A roll back to the 1.03 polls and a re-release of 1.04 would be really great. It is almost a month since release now, and a playable game would be nice. This would do the trick.[/QUOTE] :eek: you cant be serious.....????? You must be playing NCAA 2006 for your PS2, this game is called Bowl Bound and you play it on your PC... Dumb ass...

[TPBPlayer]

[QUOTE=JJSmitty]:eek: you cant be serious.....????? You must be playing NCAA 2006 for your PS2, this game is called Bowl Bound and you play it on your PC... Dumb ass...[/QUOTE] I don't own a PS2, nor an X-box for that matter. Because there are elements of the game, namely the polls, that are broken, and it ruins the experience for me, that makes me a dumb ass? I would say you are the dumb ass for blindly playing a game because it looks pretty on your screen, regardless of whether or not it actually works right. So go **** yourself.

[Arlie Rahn]

As I said before (numerous times), "It will be addressed in 1.1". So, wait until 1.1 if it really bothers you. My point is that I just don't see the main "showstopper" comments. Polls do not effect your budget or recruiting, are very minor in prestige (wins, conf finish, bowl appearance/result and CPU ranking are much more important), and really only impact your bowl bid (along with CPU ranking) if you are one of the wildcard teams in the GDCS. Is it important to fix? Yes, and it will be fixed in the next update. Should it cause people to stop playing the game for fear of gameplay impact? Not in my opinion. But, that's a call each of you have to make. I am going to be working as fast as I can on 1.1, but there are a lot of important additions/changes that I don't want to rush.

[JJSmitty]

[QUOTE=TPBPlayer]I don't own a PS2, nor an X-box for that matter. Because there are elements of the game, namely the polls, that are broken, and it ruins the experience for me, that makes me a dumb ass? I would say you are the dumb ass for blindly playing a game because it looks pretty on your screen, regardless of whether or not it actually works right. So go **** yourself.[/QUOTE] Ohh OK, now I understand your point... You really are a dumb ass if you think the game is broken, Arlie has worked his ass off to get everything corrected. BTW, the problem with the polls seems to work it self out at the end of the season. So I have to ask you,"have you really played the game or are you just being a dumb ass" ;)

[TPBPlayer]

[QUOTE=JJSmitty]Ohh OK, now I understand your point... You really are a dumb ass if you think the game is broken, Arlie has worked his ass off to get everything corrected. BTW, the problem with the polls seems to work it self out at the end of the season. So I have to ask you,"have you really played the game or are you just being a dumb ass" ;)[/QUOTE] Because he worked his ass off, that makes the game fine and above any criticism, especially from a paying customer? I wish my job worked like that. I'm done tossing insults around.

[Rene778]

Ahhh...personal insults. I feel at home now....which sports board is this now

[JJSmitty]

[QUOTE=TPBPlayer]Because he worked his ass off, that makes the game fine and above any criticism, especially from a paying customer? I wish my job worked like that. I'm done tossing insults around.[/QUOTE] cool.... Take your ball and go home...

[Hador]

[QUOTE=Arlie Rahn]As I said before (numerous times), "It will be addressed in 1.1". So, wait until 1.1 if it really bothers you. My point is that I just don't see the main "showstopper" comments. Polls do not effect your budget or recruiting, are very minor in prestige (wins, conf finish, bowl appearance/result and CPU ranking are much more important), and really only impact your bowl bid (along with CPU ranking) if you are one of the wildcard teams in the GDCS.[/QUOTE] You are surprising me more and more with your posts. I understand that you can't fix this overnight but your comments are now bordering on being ridiculous. First a 2-4 team in the Top25 midway through the season is no big deal and now you are trying to tell us that the polls have not much influence since they only decide who will play in the championship game.... :confused: As I said I am all fine with you saying that you'll try to fix it with 1.1 but please stop trying to tell how us how small this problem is.

[JJSmitty]

[QUOTE=Hador]You are surprising me more and more with your posts. I understand that you can't fix this overnight but your comments are now bordering on being ridiculous. First a 2-4 team in the Top25 midway through the season is no big deal and now you are trying to tell us that the polls have not much influence since they only decide who will play in the championship game.... :confused: As I said I am all fine with you saying that you'll try to fix it with 1.1 but please stop trying to tell how us how small this problem is.[/QUOTE] Dude, just use the playoff option if you don't like the polls... It's that simple.... Damn Arlie it must be hard to try and please everybody because it seems some people just like to complain...

[Tim Plum]

[QUOTE=Hador]You are surprising me more and more with your posts. I understand that you can't fix this overnight but your comments are now bordering on being ridiculous. First a 2-4 team in the Top25 midway through the season is no big deal and now you are trying to tell us that the polls have not much influence since they only decide who will play in the championship game.... :confused: As I said I am all fine with you saying that you'll try to fix it with 1.1 but please stop trying to tell how us how small this problem is.[/QUOTE] The 2-4 team is no big deal if you understand exactly what the polls do in the game. Arlie programmed the game so his explanation on the polls having little influence during the season and they decide the championship game is what the polls mean. Hador rather then referring to someone elses poll results what have your polls looked like? Has it changed anything you do in the game?

[Arlie Rahn]

[QUOTE=Hador]You are surprising me more and more with your posts. I understand that you can't fix this overnight but your comments are now bordering on being ridiculous. First a 2-4 team in the Top25 midway through the season is no big deal and now you are trying to tell us that the polls have not much influence since they only decide who will play in the championship game.... :confused: As I said I am all fine with you saying that you'll try to fix it with 1.1 but please stop trying to tell how us how small this problem is.[/QUOTE] Fair enough. Working on the polls is one of my top priorities in 1.1 and I will leave it at that.

[John Comey]

Arlie, I apologize this thread came to this. I wish I had voiced my concern with you personally, either via PM or on a messenger service. My apologies. -John

[Miral]

It's all John's fault. :) Just to add a little more data, here's some rankings for Big East teams: Pitt 6-1 #35 Media, #34 Coaches, #10 CPU, #17 Off, #12 Def WV 6-1 #34 Media, #32 Coaches, #14 CPU, #5 Off, #9 Def Rutg 501 0 votes Media, #45 Coaches, #28Cpu, #30 off, #24 def The descrepancy between the two polls and the computer rankings in each case is significant, like the smaller conferences get no love at all. Pre 1.4, both WV and Pitt would have been in top 25 in both polls by now. West Virginia is 5th overall in offense and 9th overall defense, yet isn't close to even cracking the top 25 yet. Its striking how dissimilar the cpu rankings are to the other two polls though.

[coachcbj]

Arlie, Ive fired shots off about how serious I think the poll issues are too. But on another note. just keep doing what your doing. They game is going to be great. Its only been out a month and better than anything else out there at this time. So I apologize if any of my comments or gripes offended you. Keep working hard.

[srudoff]

Instead of complaining, perhaps we could offer up ideas for Arlie on Poll AI...... Obviously the components need to be Prestige, Wins, Losses, Wins against top 25, Road Wins, Home losses, result of last game Anyone care to pontificate?

[JJSmitty]

[QUOTE=srudoff]Instead of complaining, perhaps we could offer up ideas for Arlie on Poll AI...... Obviously the components need to be Prestige, Wins, Losses, Wins against top 25, Road Wins, Home losses, result of last game Anyone care to pontificate?[/QUOTE] Excellent Idea srudoff, I've noticed that the polls usually work them selves out after the conference championship games. Now during the season there's always alot of strange movements but it seems to work it self out in the end.. BTW does anybody know if your Head Coach Retires..????

[Arlie Rahn]

I've been doing some research on this and it seems the problems are three-fold: 1. Teams that lose do not drop enough for losses 2-4. Teams in the top 20 will usually drop 2-3 spots for their first loss (which is OK, but a little low in some cases). But, they will also only drop 2-5 spots for losses all the way to 4. Plus, it seems like teams in the top 10 drop even fewer spots for losses early on. Now, as the season goes on, some teams may jump ahead. Still, this causes some teams to hang around in the top 25 much longer than they should. 2. Teams outside the top 25 do not move up as fast as they should. This explains why some have seen unranked programs like Syracuse to still be 25-35 at 8-0. It appears that they have such a hill to climb from where they start, they never really get a big enough jump. 3. Teams are sometimes stalled from moving up by a set of teams that win above them. IE, No. 12 Penn State beats No. 3 Ohio State by 17. But, because the No. 11 and 12 teams also won, they sometimes don't move above a No. 8 team that had a subpar win, loss or didn't play. This is a tough one to fix but I think by fixing 1 and 2, this should be resolved.

[bryanellis]

Third season for me and was the 15th seed in the playoff format. I went on to win the whole thing (by reloading over and over just to see what i would be ranked at the end of the season) only to be ranked 40th by cpu and unranked by media/coaches poll. So you can win the playoffs and the title but not be ranked even close to the top 25.

[srudoff]

[QUOTE=Arlie Rahn]I've been doing some research on this and it seems the problems are three-fold: 1. Teams that lose do not drop enough for losses 2-4. Teams in the top 20 will usually drop 2-3 spots for their first loss (which is OK, but a little low in some cases). But, they will also only drop 2-5 spots for losses all the way to 4. Plus, it seems like teams in the top 10 drop even fewer spots for losses early on. Now, as the season goes on, some teams may jump ahead. Still, this causes some teams to hang around in the top 25 much longer than they should. [b]It's my perception that teams should drop 2 to 5 spots for their first loss, unless their prestige is low (see Louisville & Boise St in 2005) - those teams should drop 6-15 spots IMO. BSU dropped from 18 to not ranked and UL from 9 to 24. It also should depend on their foe. OSU lost at home to #2 Texas and went I think from 4 to 9. I would hope that a loss to Miami of Ohio would lower them even more. Teams should drop even more for the 2nd loss and much more for the 3rd and 4th losses. Esp if there are multiple teams with less than 3 and 4 losses.[/b] 2. Teams outside the top 25 do not move up as fast as they should. This explains why some have seen unranked programs like Syracuse to still be 25-35 at 8-0. It appears that they have such a hill to climb from where they start, they never really get a big enough jump. [b]Agreed - and I have no idea how to fix this :)[/b] 3. Teams are sometimes stalled from moving up by a set of teams that win above them. IE, No. 12 Penn State beats No. 3 Ohio State by 17. But, because the No. 11 and 12 teams also won, they sometimes don't move above a No. 8 team that had a subpar win, loss or didn't play. This is a tough one to fix but I think by fixing 1 and 2, this should be resolved. [b]It's almost as if you'd have to create a old school BCS "quality of win" factor. Some wins are > than others.[/b][/QUOTE] Good luck with this - hopefully more people will opine

[Hador]

[QUOTE=Arlie Rahn]I've been doing some research on this and it seems the problems are three-fold: 1. Teams that lose do not drop enough for losses 2-4. Teams in the top 20 will usually drop 2-3 spots for their first loss (which is OK, but a little low in some cases). But, they will also only drop 2-5 spots for losses all the way to 4. Plus, it seems like teams in the top 10 drop even fewer spots for losses early on. Now, as the season goes on, some teams may jump ahead. Still, this causes some teams to hang around in the top 25 much longer than they should. [/QUOTE] I think the opponents record definitely needs to be factored in here as has already been sugested. A loss to a higher ranked team has to have less effects than if let's say the #5 team loses to an unranked opponent. Also margin of victory could factor in here. However this again should be weighed against the oponents rank, record and SOS. [QUOTE] 2. Teams outside the top 25 do not move up as fast as they should. This explains why some have seen unranked programs like Syracuse to still be 25-35 at 8-0. It appears that they have such a hill to climb from where they start, they never really get a big enough jump. [/QUOTE] Again there should maybe be a larger impact for beating a ranked opponent. Also maybe SOS should propably have less impact. [QUOTE] 3. Teams are sometimes stalled from moving up by a set of teams that win above them. IE, No. 12 Penn State beats No. 3 Ohio State by 17. But, because the No. 11 and 12 teams also won, they sometimes don't move above a No. 8 team that had a subpar win, loss or didn't play. This is a tough one to fix but I think by fixing 1 and 2, this should be resolved.[/QUOTE] I wouldn't dwell too long on this. This happens in real life all the time. I agree, the important thing here is to increase the impact of a loss because this way room will be freed higher up in the polls. I am not sure what factors you already look at for calculating the polls but from what I am seeing I would suggest the following: 1. Influence of SOS should be decreased: It seems that SOS is way overrated right now which kind of results in a positive feedback effect for teams from high-class conferences. SOS should IMO be used mainly as a secondary factor instead of a direct one. I think SOS could be used to calculate quality wins. Let's say Team A beats Team B (7-1) which has run up the 7-1 record against quality opponents. In that case the win should amount for quite a lot. If team B ran up it's record by beating cup cakes the win shouldn't amount for much. I hope you get what I mean. Kinda like the quality win factor. It may not be perfect but I think if done right it can be a great help. 2. The impact of a loss should depend heavily on how the winning team is ranked. If a Top 5 team loses to an unranked this should have bigger consequences than #25 losing to #28. 3. Multiple losses should have more impact. Right now it seems the actual number of losses doesn't make much difference as long as your SOS is high enough and you have been ranked high at the start of the season. 4. Other components: Home Loss vs. Road Loss, Margin of victory, Opponents prestige, maybe one could even take into account how the game was lost. Did the team beat itself because of penalties and turnovers or were they simply dominated by the opponent. Still this might be very difficult to implement.

[srudoff]

I think SOS should be replaced if possible with strength of last opponent. Arlie - does your logic recreate the polls every week or does it start with prior week's rankings and adjust after the games are run?

[MirageAg99]

[QUOTE]I think SOS should be replaced if possible with strength of last opponent.[/QUOTE] That’s the ticket right there, IMO. I think all pollster look at are records and the last opponent played. If you are ranked below a team with a worse record (happens all the time), and you both win, the strength of the last opponent should count for a lot. If the team above you has the same record, the last opponent counts for some, but not as much. A case in point from this year. ND lost to a bad Mich State team, but after the USC game and some wins later, no one cared about Mich State. People focused more on the last game (ND closely losing to USC) than the bad loss at home to MSU and thus proclaimed Wies a genius (not that he isn't). My suggestion for a ranking algorithm is start off with prestige and expected conference place (however you figure that out) and move on from there. After the week is played do the following: 1st, sweep by record and previous ranking. This means if the team didn't lose, it is very hard to move them down the ranking. With out the previous ranking flag, you would have seen USC fall a few poll positions after the Fresno State game. 2nd, rate teams with respect to their last game only. give ratings for strength of opponent, margin of victory (or loss), ugliness (penalties and turnovers, winning an ugly, close game from a power doesn't have the impact of winning a close, hard fought game), and any other factors you can think of. This way you reward the equivalent of the UT-OSU game. Anyway, that's my 2 cents (or more).

[Miral]

Ouch. I just finished my season with Rutgers with an 11-1 record. 9th best offense in nation, 15th best defense, and while the coaches voted me 17th, my 11-1 team didn't even crack the media top 25.:mad: Granted I had very very weak sos, but still how can an 11-1 team miss the top 25 on any poll.

[Toddzilla]

Silly question - has anyone gone through the real coaches and media polls week by week to see how the teams move up and down in real life? I took a glance at them tonight - I'll probably write up an analysis sometime - and it seems that the following hold true for ranked teams: Beating another ranked team means more than beating an unranked team. In many instances, teams with big wins in the top 25 will jump other teams who also win, but not many. A one-loss Miami jumped over undefeated Alabama after Miami beat #3 Virginia Tech. Losing to a top 25 team will drop you, but not as much as losing to an unranked team. The top 5 seems to be much more static than the second 5, and they are more static than the next 5, and so on. In other words, movement in the bottom of the polls is more frequent than at the top. It is like a pyramid distribution of talent, where you need to do more work to get you from 5 to 4 then you do from 6 to 5, and so on. 4 losses were the most for any ranked team at the end of the year (prior to the bowls), and all of those teams had at least 7 wins. Going back to week 12, there were no 4-loss teams, and only 5 3-loss teams. With the exception of Tennessee, no team with a 500 record or below was in the top 25 after week 8. In fact, no team that was less than 3 (!) games over 500 were in the top 25 after week 9. Michigan at 5-3 was ranked 25 in week 9.

[srudoff]

[QUOTE=Toddzilla]Silly question - has anyone gone through the real coaches and media polls week by week to see how the teams move up and down in real life? I took a glance at them tonight - I'll probably write up an analysis sometime - and it seems that the following hold true for ranked teams: Beating another ranked team means more than beating an unranked team. In many instances, teams with big wins in the top 25 will jump other teams who also win, but not many. A one-loss Miami jumped over undefeated Alabama after Miami beat #3 Virginia Tech. Losing to a top 25 team will drop you, but not as much as losing to an unranked team. The top 5 seems to be much more static than the second 5, and they are more static than the next 5, and so on. In other words, movement in the bottom of the polls is more frequent than at the top. It is like a pyramid distribution of talent, where you need to do more work to get you from 5 to 4 then you do from 6 to 5, and so on. 4 losses were the most for any ranked team at the end of the year (prior to the bowls), and all of those teams had at least 7 wins. Going back to week 12, there were no 4-loss teams, and only 5 3-loss teams. With the exception of Tennessee, no team with a 500 record or below was in the top 25 after week 8. In fact, no team that was less than 3 (!) games over 500 were in the top 25 after week 9. Michigan at 5-3 was ranked 25 in week 9.[/QUOTE] Michigan is ALMOST always the top ranked team for however many losses they have in real life (ie top ranked 1 loss team, top ranked 2 loss team, etc)