J.C. Moln (7) and Glenn Coleman (9) told me to tell you to SUCK IT! :D

Oh yes, Ben making the best use of his players. Shocking. :devil:

Coleman belongs to Quik. Equally shocking, I know.

He doesn't yet have the stats, but Fonzie hasn't shied from noodle-armed guys either.

'Belongs' to? And at only 15 million? SLAVERY!

I thought i should bring this up

This is from a Q&A Ben did with Jim found here
February 2008 Solecismic Q&A: Drafting Tips - Front Office Football Central

So while height/weight will not affect a guys current/potential numbers it could have an effect on his ability to actualy play at that level if he is undersized it seems.

Be careful. That was Feb. 2008. Multiple significant game updates have occurred since then. I suspect things have changed since then in that arena.

So height/weight now applys more directly to bars and ratings?

I can see the merits of both ways. Would be nice to know which it is though.

I didn't say that either. I said that I *suspect* that things have changed. And if they have, I don't think it applies directly to bars and ratings. I'd think it would be applied on top of the existing bars/ratings.

All,

Sorry I haven't had more time to work on this. Life's gotten fairly busy.

Simple version:


Okay, the above is a variant on the original idea. For each position and each measureable there are two values.

The top value is the "good value." Any value above that Solecismic, Bench, Broad Jump, or Position Drill indicates that value is a 575, 600, or 625 level skill. Any value below the top value for 40 Yard and Agility represents the same.

The reverse is true for the bottom value. Any value below the bottom value for Solecismic, Bench, Broad Jump, or Position Drill indicates that value is a 375, 400, 425, or 450 value player.

This is a quick, simplistic, and dirty way to find outliers.

But we've always known this.... What about the difficult way?

That will be coming in posts as I have time to put them up.

Solecismic Scores

READ THIS FIRST!!
This data is currently untested and is derived from statistical analysis detailed above.

To get the numbers to line up in a row perfectly there had to be a little fudging, especially at the periphery (really low and really high) due to small sample size at these extremes.

Use these values at your own risk. Not FDIC insured. People can and do lose money.


HOW TO INTERPRET THE DATA
Suppose you have a player with a 17 Solecismic score. If that player is a QB, then the Solecismic score tends to indicate a 375 level player. However, if that player were a DT, that would indcate a 425 level player.

Yes, some scores you'll never see. I've never heard of a WR with a 54 Solecismic score. I populated the entire table just in case some wacky scores get included in a future update.

You might be saying, "There isn't a 387.5 score in the game. What the heck?" Some scores were equally balanced between two values (in this case 375 and 400). I couldn't make a determination of a single value so the average of the two was placed here. If you want to use them as is, great. You can also round these values up or down, depending on your preference.

Coming next...40 Yard Dash.

40 Yard Dash Times

READ THIS FIRST!!
This data is currently untested and is derived from statistical analysis detailed above.

To get the numbers to line up in a row perfectly there had to be a little fudging, especially at the periphery (really low and really high) due to small sample size at these extremes.

Use these values at your own risk. Not FDIC insured. People can and do lose money.


HOW TO INTERPRET THE DATA
Suppose you have a player with a 4.87 40 yard dash. If that player is a QB, then the score tends to indicate a 450 level player. However, if that player were a DT, that would indcate a 600 level player.

Next...Bench Press.

Bench Press

READ THIS FIRST!!
This data is currently untested and is derived from statistical analysis detailed above.

To get the numbers to line up in a row perfectly there had to be a little fudging, especially at the periphery (really low and really high) due to small sample size at these extremes.

Use these values at your own risk. Not FDIC insured. People can and do lose money.


HOW TO INTERPRET THE DATA
See above.

Next...Agility.

Agility

READ THIS FIRST!!
See Above for warnings

Use these values at your own risk. Not FDIC insured. People can and do lose money.


HOW TO INTERPRET THE DATA
See above.

Next...Broad Jump.

Broad Jump

READ THIS FIRST!!
See Above for warnings

Use these values at your own risk. Not FDIC insured. People can and do lose money.


HOW TO INTERPRET THE DATA
See above.

Lastly...Position Drills

Position Drill

READ THIS FIRST!!
See Above for warnings

Use these values at your own risk. Not FDIC insured. People can and do lose money.


HOW TO INTERPRET THE DATA
See above.

And done.

Let me know if this data is useful by leaving a comment in this thread or sending me a PM.

Bravo, stretch. I look forward to putting these through some drafts.

I've used them in my last two drafts. My draft 2 years ago wasn't bad but this year I nabbed a 35/70 DE in the 2nd round.

How does this differ from just using the red/blue/green numbers in-game and their equivalents in Draft Analyzer / Conscriptor?

For the first post (the simple version) probably not much.

For the players that aren't color coded, I think it is very valuable to differentiate between a 425 player and a 550 player (for example).

Great combines are really easy to spot; as are really bad ones. It's the unwashed masses in the middle that I think can be differentiated using the more detailed analysis here.

Your forty speed post just highlights how important the forty is to the WR and CB positions, more so than any other position. These positions ratings starts to dip in the milliseconds while others stay pat.

...and now we have a fascinating cat in the CCFL draft class. Meet Bryce "Noodle-Arm" Strom







The ultimate short/medium only guy?

If I were still in CCFL, I'd give my left and right nut for that guy.

What's with the "QB G" in that screenshot, by the way? I don't see that in my game.

Where in the first round did that guy go?

Some people use some add-ons with FOF, including graphics to make some of the in-game stuff more meaningful. I don't recall whether the letter grades are part of the Stelmack suite (I seem to recall that using a numeric system) but that's the main idea.

It's part of the Stelmack Suite. Numbers and letters are both options there.


He hasn't been drafted yet.

Its part of resource patcher. I think you choose the Icy stuff.

Wow, I've never seen this thread before. Can't believe I missed it. Looks way useful!

What exactly is an outlier?

In statistics, an outlier is an observation that is numerically distant from the rest of the data.

An RB with a bench score of 27 but average combine scores otherwise is a good example of an outlier. A lot of these guys are really good at one thing, which generally makes them potentially more valuable overall than a guy who is average or slightly above average across the board. The chances of a guy with similar extreme outlier numbers in at least one category being worthless are very low.

Aaahhh Dozer McCarthy in OSFL.

I have a different view. Sometimes this is true but in general an outlier combine score is only indicative that the player will be outstanding at certain bars and not others.

I really think people should start getting away from looking at combine scores ON THEIR OWN and start to take a more global view. I used to look at this Malcpow thread a lot but only when I starting looking more globally did my drafting take a quantum leap. Now I do not even bother coming to this thread.

Umm, I agree?

I think there is a slight difference there.

Take the example of the crazy 27-bench Dozer McCarthy from the OSFL, mentioned earlier. This was a guy that had ho-hum, but not bad, combines across the board and a bench that was about as good as it possibly gets. He wasn't just some Power Inside specialist, he was a solid creeper that rose across the board -- and the decent combine plus the extreme outlier cemented that. Goes back to this:


So yes, do look for big-time outliers as indicators of guys who are simply good all around. On the other hand, Malcpow is pretty spot on when he says about the QBs, "even one score outside the constraints can be a bad sign." I think this applies similarly to other positions, maybe not to the same extent. A mid 7.4s guard, for example, who had a very weak dash score -- the agility was red and easily in one of the top brackets, while the dash was green. This guy was not a creeper and just a pass blocking specialist.

The disagreement is just that the chance isn't very low. Even with an extreme outlier, look for other signs to make sure you are actually getting something special and not a one-trick pony. Neither of which would be really bad, I guess -- although if the one-trick pony is a WR with massive courage or an OL who has nothing but BS, those are cases where the utility is going to be pretty limited.

I think you did not get what I was saying. Aston gets more to the meat of it.

I actually think pure combine drafters are missing the boat. You can be a good drafter if you only look at combines but you are missing a whole other world. It has been proven recently that top value players can have very average combines (think all black). In most multi-player leagues, people can spot those guys now and take them early on. If you are drafting combines only, you will miss these guys and fall behind.

I really don't think anyone drafts purely on combines anymore.

I have a doubt. I saw a RB that ran a 4.44 40 yard dash, but had a bad combine performance, like a 9 bench press. Look at his combines:

Rating Adjusted Solecismic Dash Bench Agility BJ Position Drill
3.50 3.70 20 4.44 9 7.27 9' 3" 11

Look at his bars:
Front Office Football Conscriptor version 1.28.7c

Is he worthy a try as a late round pick? What do you guys think?

Those bars seem to match the combines: he is fast but not much else. What's the question here?

7th round pick maybe. but overall bad combine. id take a flyer on him

I know this thread is long dead but I just got back into FOF after a 2 year layoff and I wanted to answer a couple of questions:

The method I used to derive the data involved creating a CSV file of players with all the same value for every attribute (375, 400, 425, etc. through 625). I then used the FOF 2004 draft file utility to create a draft file for those players, imported it into FOF 2007 and then used GStelmack's Extractor to pull the combine data for the entire class. Wash, rinse, repeat until I had a full draft class for each position and rating.

I then pulled the data into Excel and started analyzing frequency of each combine value for each attribute (how often a bench of 7 for 625 QBs? None).

At the extremes there were some outliers who skewed the data a bit so I had to fudge those out of existence. For example, there was a 475 CB that had a higher Wonderlic than all 3,000 CBs that were 500, 525, and 550 rated.

It looks like there is, very occasionally, a "workout warrior" who has a particular combine value well above (or below in the case of 40 time and agility) his peers with the same value. These are so rare (3 in 1000 or so), however, that if you find one, it's just horrible luck.

At any rate, I'll go dig up the raw, unfudged data and post it here.

This thread is hardly dead, it's still a great resource even if posts aren't flowing.

I'm actually glad it works that way, but it keeps some uncertainty in the game.

That'd be fantastic, stretch. Please do.
(I hope you have TCY by the way -- I suppose an easy way to verify is to see if your 625 BPR players have 100 BPR or only 78)

If it hasn't been said before, this thread should be a required read for any FOF 2007 player.

+1. This thread has helped me so much over the years, from when I first read it when I started playing MP, to all the different insights and understandings I drew from it over time. Still got it favorited* and refer to it every draft.

Thanks, MalcPow :)

* actually saved a desktop version too, in case it ever disappears on us again ;)

One thing I would look at is the mean and standard deviations of the dataset. I've had some luck with that looking at a 95% interval (mean + 1.96 Stddev) to give myself a confidence interval.

Did you have TCY loaded and licensed (Stretch and other).

I've been able to get consistently 90+ with 625 BPR but don't see the 100BPR.

My guess is the 625 to 100 is a correlated but not exact. Not a true ((x-375)/25)*10.

Would still expect some 100's

I don't have TCY loaded and I consistently get 78-79 BPR out of 625. One of the issues with it. I tried installing TCY, but it appears not to work unlicensed (I don't own the game)....which is strange, and makes me feel that maybe there are other issues at play. I mean, surely the draft import function doesn't also check for TCY licenses.

Just to pop in here, the original research posted was definitely done on a machine with TCY licensed and installed. Plenty of 100 BPRs in that testing with a 625 value entered. You definitely want the X-factor setting off for this kind of work.

Without having looked too closely at the work stretch did, I think it's clear he took the time to peel the curtain back even further in the data he displayed. Good stuff.