FOF2K7 Scouting Error: Is it too high?

[Icy]

Umm in my game and i guess it will happen always, the level of the FA's is way better than the default players in most of the teams. In the first week, there were a lot of undrafteed rookies with ratings like 23/63 18/62 18/58 24/64 and i mean dozens of players like those, so the AI has fired all their players under 30 to sign those.

I guess is the attempt of the game to have a good talent pool but it's really unrealisitic to have all those awesome players as undrafted FA's that are way better than most of the vets in the default roster, even more being in the first season. It makes the real roster totally unuseful unless for the top players in every team, as like 30% of the original players are fired to sign all those superstars undrafted players sitting in the FA pool.

If i start to post screen shoots of those exceptional undrafteed rookies at the start of the game you would be amazed at how good they are. I would choose a lot of them in a 1st round in the draft.

Please guys, start a new game and check the transaction log before doing anything else. Click on a few of those recent signing and check how good they are.

[Icy]

Dola, just started a few games more and it happens always. Check the transactions log before doing anything else, there are dozens of signings of players with 60's in potential before the first week is finished.

[Icy]

Triple dola, Here is an example of a guy who was signed as undrafted rookie from the FA pool. His future is too good to be an undrafted rookie sitting in FA.

Also, how the heck is this guy 8/43 with theese ratings? His potential is more close to 60s if not higher.

http://www.prodeportes.com/fof2k7/player1.jpg

[Icy]

Look at more undrafted rookies that were in FA and signed by the AI teams just after starting a new game. Those players can't be undrafted rokies in the FA pool as they are even good enought as second rounders. I think this defeats the purpose of having a real roster set as you can create a new team only with them.

New starting LT for the Jets
http://www.prodeportes.com/fof2k7/player2.jpg

New starting LDE for the Bills:
http://www.prodeportes.com/fof2k7/player3.jpg

New starting LG for the Bills
http://www.prodeportes.com/fof2k7/player4.jpg

Great backup for the Panthers
http://www.prodeportes.com/fof2k7/player4.jpg

[Deattribution]

Look at more undrafted rookies that were in FA and signed by the AI teams just after starting a new game. I think this defeats the purpose of having a real roster set.

New starting LT for the Jets
http://www.prodeportes.com/fof2k7/player2.jpg

New starting LDE for the Bills:
http://www.prodeportes.com/fof2k7/player3.jpg

New starting LG for the Bills
http://www.prodeportes.com/fof2k7/player4.jpg

Great backup for the Panthers
http://www.prodeportes.com/fof2k7/player4.jpg

I'd say some of these could be caused by the computer passing on players due to lack of workouts, or being considered overrated or underrated.

[Icy]

I'd say some of these could be caused by the computer passing on players due to lack of workouts, or being considered overrated or underrated.

This is at the start of a new game so not any real draft has been run. I guess those plaeyrs ahve been autocreated by the game engine to have more FA's in the pool, but at the start of a new game, that starts in the preseason week 1 and not just after the draft, those players shouldn't be available in the FA pool as undrafted rookies. The AI teams are signing up to 10 of those players in the first week so the default roster set veteran players are being replaced by fake ones than are rookies. Unrealistic in my opinion.

At the start of the first season, the FA pool players should be some average veterans, some average to poor undrafted rookies and maybe one of two hidden gems, but not that much great players.

[yabanci]

...
1st- Just after starting the game i took a look at the transactions report and i see it's full of movements. It includes a lot of FA signups, some releases and also a lot of players positions changes. It's nice to see the AI teams active, but if a soon as i start the game all the good FA's are signed already by the teams... well it's not so fair with the human GM. We all should start to make offers at same time. I would prefer it to work like during the FA stages, i make an offer, the AI makes his offers and then the FA's decide after i advance one week.

Anyway this is not a big issue for me as i like the AI to have some advantage but it looks that the AI teams are one sim stage ahead of you, i guess it's a design decision and i think it was that way too in 2k4.....

This happens because there aren't enough players in the original file to fill 60 man rosters for each team. If you don't want to get shut out on filling your roster with these initial free agents, what you want to do is -- before you start the new career -- set your global options to "scouts automatically sign players to fill roster." You can then turn the option off after the game starts. It might not be the perfect fix, but at least you don't left out of the party.

[yabanci]

Look at more undrafted rookies that were in FA and signed by the AI teams just after starting a new game. Those players can't be undrafted rokies in the FA pool as they are even good enought as second rounders. I think this defeats the purpose of having a real roster set as you can create a new team only with them...

I agree with this 100%. Although not a bug, the game generates initial undrafted rookie free agents with way too much upside potential, many of the 1st and 2nd round quality. The game should create these players as roster filler, which is what they are. It would be great if Jim could tweak this.

[Icy]

This happens because there aren't enough players in the original file to fill 60 man rosters for each team. If you don't want to get shut out on filling your roster with these initial free agents, what you want to do is -- before you start the new career -- set your global options to "scouts automatically sign players to fill roster." You can then turn the option off after the game starts. It might not be the perfect fix, but at least you don't left out of the party.

Thanks for the tip, and yes, probably the problem is not enought players in the default roster file, that makes the game to create a lot of fake undrafted rookies to fill those teams. The problem is that those fake rookies created are too good to be just fillers, in fact there are some of 2nd round value as i said before.

[Icy]

Dola, you beat me to explain it yabanci :)
Potential for the roster fillers should be more around 30 max instead of over 60, i hope Jim can tweak it.

[Icy]

Here is the last proof i post about this. Look at the undrafted rookies in the FA pool when you start a new game. I have sorted it by potential and it looks like a full draft file with all those great rookies.

http://www.prodeportes.com/fof2k7/fas.jpg

[Solecismic]

On this one, believe it or not, it is working as designed. There was no change made to the algorithm generating undrafted rookies. However... the scouting error is just a little bit higher.

If you open a new game in Main Street mode, that screen would range more from 45-50 in potential.

I just ran the test in '04 and the range was 40-45 there on Main Street and 44-52 on Wall Street. In '07, the scouting error on Main Street is about the same as the Wall Street level in '04.

The biggest adjustment from 04 to 07 is going to be the scouting error. Did I make it too high? That may well be the case. This is probably worth some separate discussion. Odds are good that not a single one of those players in your screen shot will ever be worth starting.

I think this issue should be split out into a new item, where it can be the focus of discussion.

[Icy]

Umm i like a big scouting error to base my decissions more in stats and not only in ratings... the only problem is if the AI is also seing those inflated ratings and in a hurry to sign them while they may end being crap. If the AI is capable as we humans of taking the right decissions based also on stats, then i like it.

[AlexB]

If these guys do turn out to be overrated duds, then yes the scouting error is too high IMHO - other wise attributes are worthless, and we need a few years stats history on any player.

The draft should now be based on the combine numbers, ratings and scouts impressions, with little or no regard to the attribute bars. This explains why some of the top rated guys look useless when scouted maybe?

If players went undrafted, it is reasonable to assume scouts didn't think much of them at the time: all of a sudden after the draft scouts suddenly see them all as potential pro-bowlers? Doesn't make sense...

edited to say 'potential pro-bowlers'

[Groundhog]

So, when Vince Young went from ~26/68 as a QB to 34/80 as a SE it was not such a big deal? He might never come close to living up to his potential?

[RedKingGold]

I think this might actually be a good thing if it applies to younger (R-3rd year players).

Let's say I take a rookie who ranges between 41-73 on future ability. As he progresses through his career, his statistics give an indication of the true value of the player as the future ability range narrows. By his 6th or 7th year; we should have a pretty good indication (by the red bars) of his actual value.

This is very similar to how it works in real-life professional football. There are some players who scouts thought had all the right tools, but never came close to their realized potential.

However, if this scout variance exists with veteran players; that is not too realisitic because by the 8th year of a player's career, he usually is what he is (with certain exceptions like Randal Cunningham and Kurt Warner; but they are the exception rather than the rule and are covered by the unexplained veteran breakout)

[Caratacus]

I'm not sure on this one, If those players are not worth starting my guess would be they are actually 30 something at best in real terms but are showing as high as 68 and there are so many of them. I don't want ratings to be so far off they are almost meaningless.

If (not including busts) players can be 30 points out ratings wise I can't see the point having any rating at all?

[Eaglesfan27]

I've also wondered about the scouting error. Bledsoe is a 32/32 guy in my league yet he is the top QB statistically after half of a season. I thought that was mostly due to having Terrell and Glenn who are very good receivers, but maybe Bledsoe is quite a bit better than my scouts see him.

Quite a few other cases that made me wonder about scouting error and how much larger it is in this version. Glad Jim commented on that.

[highfiveoh]

What kind of ranges would most guys be happy with? Scouts were pretty worthless in 2004. I definitely don't want to go back to that.

[Caratacus]

What would be nice is if by giving them playing time in preseason & spot duty in regular season the scouting report updated as they played more. So if you started a player in all 4 ps games you would have atleast some idea if he was living upto his ratings.

Going on a young players stats alone for preseason seems a bit dubious as he could be playing against the opponents second string and if he is very under-developed he isn't going to show much anyway. I guess I just don't like the thought of starting a player for a couple of years before finding out he is no better than me at playing football, fair enough if he is a high draft pick with all the physical tools but just doesn't live up to his potential, but an unsigned rookie free agent that my scout thinks is a quality 2nd rnd pick? Hmm

[aran]

Different scouts should see different guys as break-out, high potential players. One scout shouldn't seem them ALL that way unless he's terrible with young talent. I could understand if there are a few 18/65s that go undrafted according to my scout, but the population of players like that should be no greater than 10--EVER--and which players those ten are should vary depending on where the scout is weakest.

[yabanci]

I just want to reiterate on one specific point. The issue raised by Icy is about the quality of the free agent roster filler when you start a new game.

These are players who went unselected in the previous year's draft. For example, if you use real players, they are fictional players left unselected from the 2005 NFL draft. Consequently, these players should be of the quality of players that you typically see go unselected in FOF drafts. However, what we see are many players with ratings so high that they would never make it through a draft without getting selected, players with future potential in the 60s. We are seeing the game start out with undrafted rookie free agent quarterbacks with ratings often higher than guys like Matt Leinart or Jay Cutler.

That's the problem raised by Icy: this initial roster filler should be of the quality of players typically unselected in an FOF draft, not quality of first round picks with future potentials in the 60s who are signed and become immediate starters.

Jim noted that it may be scouting error and raised the issue of whether scouting error is too high in the new game. It may or may not be, and that is a good topic for discussion, but either way the initial undrafted rookie free agents should not be perceived by your scouts as substantially better than undrafted rookie free agents in any other FOF draft. You should not see an initial pool of undrafted free agents rated in the 60s and 50s.

Regardless of any scouting error, you never finish a draft and see a pool of undrafted players who look like this:

http://www.prodeportes.com/fof2k7/fas.jpg

It doesn't happen. That would be a pretty good draft class, not a group of players who fell through the cracks. I don't think this is an issue of scouting error.

So the specific change that is being requested is to lower the quality of the initial pool of undrafted rookie free agents generated by the game. Instead of potentials in the 60s and 50s, they should have potentials of about 40 and below.

[aran]

It doesn't happen. That would be a pretty good draft class, not a group of players who fell through the cracks. I don't think this is an issue of scouting error.

So the specific change that is being requested is to lower the quality of the initial pool of undrafted rookie free agents generated by the game. Instead of potentials in the 60s and 50s, they should have potentials of about 40 and below.

Perhaps a course of action would be to have ALL undrafted players take a ratings hit in the eyes of your scouts simply because they went undrafted.

[Ben E Lou]

On this one, believe it or not, it is working as designed. There was no change made to the algorithm generating undrafted rookies. However... the scouting error is just a little bit higher.

If you open a new game in Main Street mode, that screen would range more from 45-50 in potential.

I just ran the test in '04 and the range was 40-45 there on Main Street and 44-52 on Wall Street. In '07, the scouting error on Main Street is about the same as the Wall Street level in '04.

The biggest adjustment from 04 to 07 is going to be the scouting error. Did I make it too high? That may well be the case. This is probably worth some separate discussion. Odds are good that not a single one of those players in your screen shot will ever be worth starting.

I think this issue should be split out into a new item, where it can be the focus of discussion.The one thing that jumps out at me from your post is that it seems to indicate that scouting error is always on the high side, based on the 40-45 moving up to 44-52, and from your comment ("...not a single one of those players in your screen shot..."), the scout error is *always* the scout valuing the players too highly. Scout error should be a +/- proposition. I'd say that for rookies, something along the lines of +/- 5 at Main Street, +/- 10 at whatever the middle setting is called, and +/- 15 at Wall Street would be appropriate, with the potential for error decreasing each season. Under this system, if my scout sees a 20/50 rookie at Wall Street, he might be anywhere from a true 35 to 65. Scouts don't always overvalue rookies, do they? I would think that they just as often undervalue them.

[molson]

The one thing that jumps out at me from your post is that it seems to indicate that scouting error is always on the high side, based on the 40-45 moving up to 44-52, and from your comment ("...not a single one of those players in your screen shot..."), the scout error is *always* the scout valuing the players too highly. Scout error should be a +/- proposition. I'd say that for rookies, something along the lines of +/- 5 at Main Street, +/- 10 at whatever the middle setting is called, and +/- 15 at Wall Street would be appropriate, with the potential for error decreasing each season. Under this system, if my scout sees a 20/50 rookie at Wall Street, he might be anywhere from a true 35 to 65. Scouts don't always overvalue rookies, do they? I would think that they just as often undervalue them.

But, the only ones you're seeing when you search the free agent pool are the over-rated rookies, correct? (And that's the screenshot Jim's referring to) You're sorting by by future rating, so you're not going to get the guys that are true 55s, but rated 35.

[Ben E Lou]

But, the only ones you're seeing when you search the free agent pool are the over-rated rookies, correct? (And that's the screenshot Jim's referring to) You're sorting by by future rating, so you're not going to get the guys that are true 55s, but rated 35.Ah. Good point. It may be that there is +/- error then.

[Warhammer]

My issue with this is why aren't these guys in the initial draft pool. The reason why this is sticking out is we are looking at it after the fact and saying, "Gee where was this guy during the draft? I would have taken him at X spot."

I would think that the simple solution would be to change the point at which the extra players are generated, if possible. If you change it to prior to the draft, it should be fine.

[JPhillips]

I signed about a dozen rookie free agents at the start to see what happened. Going into year two none of them busted. There is more of a problem here than scouting error.

[TroyF]

I signed about a dozen rookie free agents at the start to see what happened. Going into year two none of them busted. There is more of a problem here than scouting error.

1) did you view them with the same scout in year 2?
2) have any of them played well for you off of stats, not just ratings?

These two questions are huge. If your scout who rated them high in year one is the same, why would that change in year two? And if the players all have nice ratings but play like garbage, maybe the scouting error is coming into fruition.

Not denying there isn't a problem here, but without more information your test above doesn't prove it.

[mhass]

edit: TroyF got my point first

[primelord]

I signed about a dozen rookie free agents at the start to see what happened. Going into year two none of them busted. There is more of a problem here than scouting error.

Players whose ratings are off by scouting error generally don't just outright bust the next year. Many of them will just slowly get worse over the years. It's my understanding as well that their green bars may not always shrink, but they just may never actually get to fulfilling them.

[JPhillips]

Its a different scout and the ratings are similar. The current scout is good with young talent and good or average everywhere else.

Here are some ratings going into the preseason for year 2.

Steve Hopkins C 28/63
Jerome Hitchcock T 27/59
Rod Rodriguez P 40/59
Wes Bridges DT 31/56
Jason Navarro MLB 19/51
Wayne Battle DE 22/51
Malcolm Pritchett SE 30/51
Willie Fisher G 24/50
Lorenzo Osborne WLB 16/49
Chad Atkins RB 35/47
Derrick Lockhart FB 21/46

And some stats for those who played much in year 1.

Atkins RB 180 att 681 yds 3.78 ypc 7 td

Willie Fisher 16 games 0 starts 18 KRO 8 KRB

Rod Rodriguez 72 punts 41.7 ave 22 In20

Wayne Battle 16 games 0 starts 14 tkl 2.5 sck

Wes Bridges 16 games 0 starts 21 tkl 2 sck

No superstars yet, but I'd say they played roughly at the level the scout said they should.

I should also note that the AI had
Hopkins C
Rodriguez P
Bridges DT
Battle DE

as starters and Atkins almost splitting time at RB in year 2.

[JPhillips]

Damn, I was going to go through a few seasons, but I just had a crash and lost the game. I'm going to try to set up a career where I sign a number of these first year guys and track them for several years. I'll post the results ASAP.

[yabanci]

Saying they might not end up being as good as they look is totally missing the issue. The AI teams treat a 20/65 guy as a 20/65 guy, which results in the problem that is at issue here:

Umm in my game and i guess it will happen always, the level of the FA's is way better than the default players in most of the teams. In the first week, there were a lot of undrafteed rookies with ratings like 23/63 18/62 18/58 24/64 and i mean dozens of players like those, so the AI has fired all their players under 30 to sign those.

I guess is the attempt of the game to have a good talent pool but it's really unrealisitic to have all those awesome players as undrafted FA's that are way better than most of the vets in the default roster, even more being in the first season. It makes the real roster totally unuseful unless for the top players in every team, as like 30% of the original players are fired to sign all those superstars undrafted players sitting in the FA pool.

[henry296]

After my first draft, I took a look at the rookie free agent pool. While no position player has a future value over 60, the first page is almost all over 50. There is a kicker with a future value of 75. The highest potential is a tackle who is 10/58.

I guess the initial pool is slightly higher than the draft and could be tweaked down a little bit.

[Icy]

The one thing that jumps out at me from your post is that it seems to indicate that scouting error is always on the high side, based on the 40-45 moving up to 44-52, and from your comment ("...not a single one of those players in your screen shot..."), the scout error is *always* the scout valuing the players too highly. Scout error should be a +/- proposition. I'd say that for rookies, something along the lines of +/- 5 at Main Street, +/- 10 at whatever the middle setting is called, and +/- 15 at Wall Street would be appropriate, with the potential for error decreasing each season. Under this system, if my scout sees a 20/50 rookie at Wall Street, he might be anywhere from a true 35 to 65. Scouts don't always overvalue rookies, do they? I would think that they just as often undervalue them.

Knowing that the starting point is Preseason week 1, when most of the teams already have their rosters filled in real life, there shouldn't be overvaluated players sitting in FA. If a player is overvaluated, it means that the scouts think he is a good player, so he would have been signed already. What should be in the FA is the opposite, players rated average or low that can be sleepers not valued correctly by the scouts, and if you gamble and sign one of them, you could be lucky and he could end better than average.

It should work like the late draft rounds (of course, if they are undrafted it's because they looked really bad in the draft no matter how good they are). In the late draft rounds, we can only draft crap guys with the hope of them booming, but there is not any player rated in the 60's who would fall to the 7th round or end undrafted, no matter how bad he really is (that we can't know), as he is perceived as 2nd rounder potential.

As yabanci also said, the problem is that the AI also preceives them as 60's so it signs them and even asign them starting spots, firing their own players. I would prefer all the scouts undervaluating them and that some if given a chance, could end being above average. But players evaluated from 40s to 60s shouldn't sit in FA. Even less when we have seen that in FOF2k7, a player rated around 40 is a good player and a 50 to 60 one is really good.

Another issue i want to raise is that most of the FA's at the start of the 1st season are undrafted players, while i think there should be a good amount of vets and average 3 years experience guys who end not being as good as their teams thought when they were drafted.

I thin this issue would only raise when starting a new season, as for the second season, when we would have run a real draft already, the FA pool will be really undrafted players and not so good vets. Not any 60's rated player will end undrafted by the AI or humans after the 1st draft is ran.

[Narcizo]

Ah. Good point. It may be that there is +/- error then.

Except experience from FOF 2004 suggests that the scouts are always over-estimating players. How many times have you traded for a player who turns out to be better than you expected? Or signed a free agent?

Call me a pessimist but I really don't think this will have changed in 2007.

[Warhammer]

Except experience from FOF 2004 suggests that the scouts are always over-estimating players. How many times have you traded for a player who turns out to be better than you expected? Or signed a free agent?

Call me a pessimist but I really don't think this will have changed in 2007.

I've done it but most of those are the late career booms.

[Celeval]

Except experience from FOF 2004 suggests that the scouts are always over-estimating players. How many times have you traded for a player who turns out to be better than you expected? Or signed a free agent?

Call me a pessimist but I really don't think this will have changed in 2007.

Plenty of times. Scouts both over- and under- estimate players; I've benefitted from this in plenty of leagues (including MP). Not talking about booms either, but the slow, +2, +3, +4 potential increase over a few seasons.

[molson]

I think everyone's assuming that the rookie FAs that have high (60s) potential are also seen by all other AI souts as having that same potential. Perhaps your scout is over-rating these players, but most of the other AI scouts are not - which is why they are not being drafted.

[AlexB]

I think everyone's assuming that the rookie FAs that have high (60s) potential are also seen by all other AI souts as having that same potential. Perhaps your scout is over-rating these players, but most of the other AI scouts are not - which is why they are not being drafted.

To add to this, the very same scout does not overrate these players before and during the draft.

He only overrates them after the draft, which makes no sense whatsoever.

Or to put it differently, the undrafted FA rookies listed post-draft in seasons 2 onwards are not ranked anywhere nearly as highly in the draft preview and rookie combine screens.

If more URFAs are added post draft to make up numbers, they should also be in the draft as the scouts rate them higher than the majority of the draft class.

This is my beef - I don't necessarily mind overrating players all of the time, but the massive shift in perceived ability is a major fault IMHO.

Having said that, other than this I love the new improved game.

[Icy]

I think everyone's assuming that the rookie FAs that have high (60s) potential are also seen by all other AI souts as having that same potential. Perhaps your scout is over-rating these players, but most of the other AI scouts are not - which is why they are not being drafted.

The AI temas are signing them and putting some of them as starters so yes, the AI see them also as 60's players.