New Roster Project Planned: I Need Your Help! First: Ratings

[wudl83]

Hi guys. I am planning on a massive roster project that focusses on "real" ratings. Since 2k12 seems to be the last game of the series my roster is also meant for future purposes.

I want to work on nearly all ratings - as far as possible, to create the greatest piece of realism one could get.

I want to use the real numbers, real statistics, real life performance so that reality is represented in the ratings. But I have trouble finding adequate numbers/stats/formulars for some ingame ratings. Also I am undecided if I should only use last seasons (I mean 2012, since I do not expect to finish the roster before the end of the regular season) or a formula where I way 2012 three times, 2011 two times, 2010 one time end divide the sum by 6.

Here is what I already got. I thought of using fangraphs.com, since it offers a bunch of statistics.

At the left you see the relevant Fangraphs statistic name, at the right the ingame ratings name.

If a player has a certain statistical number (e.g. someone has a clutch rating of 2.50) the player's ingame rating would be 99. You get it, right?

I have not decided how I will work down from the highest to the lower ratings ingame, so I only used the highest numbers possible first.

When I use the term "fan" I mean the Fangraphs.com Fan Scouting Report numbers, which seem pretty accurate to me.

Everywhere you see the question marks I have no idea what statistics, formulas, whatever, I could use for the certain ratings. Here YOU come into play. If you have any idea, let me know it. Every hint is very appreciated.

Here are two screenshots. The first one is for the position players, the second one for pitchers.

Thanks for your help guys!

[darth_mickrig]

IF you are willed to analyze them, Pitchf/x data might be the correct thing for movement ratings.

[wudl83]

Yes I already thought of it, but what is good, what is mediocre, what is bad? ^^

Edit:

I found out that fangraphs.com restructured their Pitch/FX data, now it is quite easy to see if a pitch has good or bad movement.

Edited September 6, 2012 by wudl83

[MissionMaximus]

IF you are willed to analyze them, Pitchf/x data might be the correct thing for movement ratings.

Yes I already thought of it, but what is good, what is mediocre, what is bad? ^^

Edit:

I found out that fangraphs.com restructured their Pitch/FX data, now it is quite easy to see if a pitch has good or bad movement.

In regards to movement: This site will give you the horizontal/vertical movement league avg. of all the pitches from 2010,2009,2008

http://pitchfx.texas...<br /><br />FanGraphs is the best. When you create an account you can customize what you want it to show plus the pitch identification has gotten alot better now seperating 4-seamers from 2-seamers. Also check the glossary on FanGraphs they have tables of each stat category that gives you an idea of whats good to excellent to average.

Edited September 7, 2012 by MissionMaximus

[MissionMaximus]

In regards to Pitching or more specific Relief-Pitchers my suggestion would be to go with the Shutdowns/Meltdowns (FanGraphs;check glossary for explanation). For pitchers that would be my choice to use with Clutch rating for Relief-Pitchers it's alot more indepth than blown saves or holds.

Edited September 7, 2012 by MissionMaximus

[wudl83]

Thanks for the input guys.

Has anybody an idea how I could separate the bunt rating from the drag bunt rating?

On fangraphs.com you can see how many infield hits and sacrifice bunts (I think success of them, too) but I have no clue how to separate the two bunt variations.

[wudl83]

I started to create some formulas and calculated the new hitting ratings. I started with the Giants.

I used the customize function (thx for the hint) of fangraphs. I used the season stats of 2010 and 2011 and the updated ZIPS-projections for 2012 and weighed the seasons differently. Here is what came out for the most important Giants hitters:

B. Posey 91 OVR, 85 ConR, 90 ConL, 70 PowR, 75 PowL, 72 Speed

M. Cabrera 86 OVR, 85 ConR, 80 ConL, 63 PowR, 58 PowL, 73 Speed

P. Sandoval 86 OVR, 82 ConR, 78 ConL, 76 PowR, 71 PowL, 67 Speed

H. Pence 86 OVR, 78 ConR, 79 ConL, 70 PowR, 71 PowL, 82 Speed

A. Pagan 86 OVR, 79 ConR, 75 ConL, 61 PowR, 57 PowL, 89 Speed

B. Belt 81 OVR, 72 ConR, 68 ConL, 67 PowR, 63 PowL, 75 Speed

M. Scutaro 80 OVR, 80 ConR, 76 ConL, 46 PowR, 45 PowL, 72 Speed

G. Blanco 77 OVR, 68 ConR, 61 ConL, 53 PowR, 46 PowL, 85 Speed

J. Arias 75 OVR, 65 ConR, 75 ConL, 45 PowR, 55 PowL, 80 Speed

B. Crawford 75 OVR, 64 ConR, 59 ConL, 54 PowR, 49 PowL, 82 Speed

A. Huff 74 OVR, 61 ConR, 63 ConL, 62 PowR, 64 PowL, 65 Speed

H. Sanchez 73 OVR, 73 ConR, 68 ConL, 52 PowR, 47 PowL, 55 Speed

R. Theriot 73 OVR, 69 ConR, 75 ConL, 37 PowR, 43 PowL, 70 Speed

B. Pill 71 OVR, 59 ConR, 65 ConL, 60 PowR, 66 PowL, 68 Speed

C. Gillaspie 68 OVR, 65 ConR, 55 ConL, 61 PowR, 51 PowL, 64 Speed

E. Burriss 66 OVR, 57 ConR, 59 ConL, 32 PowR, 34 PowL, 88 Speed

F. Peguero 66 OVR, 50 ConR, 60 ConL, 35 PowR, 45 PowL, 85 Speed

J. Christian 65 OVR, 45 ConR, 55 ConL, 39 PowR, 49 PowL, 90 Speed

E. Whiteside 63 OVR, 50 ConR, 53 ConL, 51 PowR, 54 PowL, 67 Speed

I think those numbers are quite solid. Funny thing that four players are rated 86 overall.

I worked on all hitting and fielding relevant data for the players noted above (+ the pitchers but their hitting is quite irrelevant). I really would like to show you more but forgive me - this would be too much data to write down.

And note: this is not biased or anything. This is all calculated data.

EDIT

Some Dodgers guys:

M. Kemp 93 OVR, 80 ConR, 92 ConL, 79 PowR, 91 ConL, 90 Speed

A. Gonzalez 90 OVR, 87 ConR, 83 ConL, 74 PowR, 70 PowL, 59 Speed

A. Ethier 85 OVR, 85 ConR, 70 ConL, 75 PowR, 60 PowL, 71 Speed

S. Victorino 83 OVR, 67 ConR, 74 ConL, 61 PowR, 68 PowL, 93 Speed

H. Ramirez 83 OVR, 71 ConR, 73 ConL, 70 PowR, 72 PowL, 85 Speed

M. Ellis 79 OVR, 69 ConR, 74 ConL, 49 PowR, 54 PowL, 76 Speed

A.J. Ellis 79 OVR, 70 ConR, 76 ConL, 51 PowR, 57 PowL, 68 Speed

L. Cruz 77 OVR, 74 ConR, 72 ConL, 60 PowR, 58 PowL, 73 Speed

N. Punto 76 OVR, 61 ConR, 64 ConL, 48 PowR, 51 PowL, 77 Speed

J. Rivera 73 OVR, 64 ConR, 69 ConL, 56 PowR, 61 PowL, 64 Speed

A. Kennedy 73 OVR, 69 ConR, 61 ConL, 53 PowR, 49 PowL, 67 Speed

J. Uribe 70 OVR, 57 ConR, 56 ConL, 56 PowR, 55 PowL, 71 Speed

M. Treanor 67 OVR, 53 ConL, 51 ConL, 53 PowR, 51 PowL, 66 Speed

Edited September 8, 2012 by wudl83

[Qbrick808]

Has anybody an idea how I could separate the bunt rating from the drag bunt rating?

On fangraphs.com you can see how many infield hits and sacrifice bunts (I think success of them, too) but I have no clue how to separate the two bunt variations.

There's no perfect way, but this is what I do to generate a success rating for each category:

For the Bunt ranking: Sac Bunts / (Total Bunts - Bunt Hits)

For the Drag Bunt ranking: Bunt Hits / (Total Bunts - Sac Bunts)

[olipoo]

i like what you're doing, a very good idea for the game. 2K has become WAYYY too generous when it comes to giving out ratings these days. no one should deserve a 99 overall rating, only the Cooperstown players should be allowed to have that.

i wish you luck on this project, won't be easy. Something i've always disliked in video games, is there is no way to measure intangibles. for example: by no means was Pete Rose a more talented baseball player than the next guy, but he played harder than anyone, and has the stats to show for it. how do you rate those grind-it-out players who just simply play hard?

a good example i see in baseball right now is the difference between Jon Jay and Carlos Beltran. Beltran is a great player, but looks extremely lazy out there if you watch him play everyday. Jon Jay on the other hand looks like he's trying to win the gold glove playing centerfield the way he is. He plays hard, and keeps his job for a reason.

[wudl83]

There's no perfect way, but this is what I do to generate a success rating for each category:

For the Bunt ranking: Sac Bunts / (Total Bunts - Bunt Hits)

For the Drag Bunt ranking: Bunt Hits / (Total Bunts - Sac Bunts)

Thx for this hint. I will definately give it a try.

i like what you're doing, a very good idea for the game. 2K has become WAYYY too generous when it comes to giving out ratings these days. no one should deserve a 99 overall rating, only the Cooperstown players should be allowed to have that.

i wish you luck on this project, won't be easy. Something i've always disliked in video games, is there is no way to measure intangibles. for example: by no means was Pete Rose a more talented baseball player than the next guy, but he played harder than anyone, and has the stats to show for it. how do you rate those grind-it-out players who just simply play hard?

a good example i see in baseball right now is the difference between Jon Jay and Carlos Beltran. Beltran is a great player, but looks extremely lazy out there if you watch him play everyday. Jon Jay on the other hand looks like he's trying to win the gold glove playing centerfield the way he is. He plays hard, and keeps his job for a reason.

First I want to say thank you. And I agree with you that 99 is a number that should not be awarded on a regular basis.

I think things like the one with Jay/Beltran could only be done if we had a rating for things like that.

They have brought in a "clutch rating" at least, but that does not go far enough.

In my opinion you are absolutely right. Perhaps something like "temper" or "attitude" with different options could be a thing that would make sense regarding your topic.

But you have problems to rate this adequately.

And since 2k12 seems to be the last baseball game by 2k anyway, it can be discussed but it won't be heard. :(

[wudl83]

Note: since I weighed in the performances of 2010 and 2011, it maybe that you think "this guy is rated too high" or "this guy is rated too low". Examples would be Ramon Hernandez or Aubrey Huff. Also, when a rookie like Josh Rutledge has not many ABs until now but has elite performance, I feel free to put his numbers down by 10. That leaves him still a respectable player, but not an elite one which he sure isn't by now.

Here are some Rockies:

C. Gonzalez 95 OVR, 86 ConR, 81 ConL, 83 PowR, 78 PowL, 90 Speed

D. Fowler 88 OVR, 74 ConR, 79 ConL, 66 PowR, 71 PowL, 92 Speed

T. Colvin 79 OVR, 65 ConR, 60 ConL, 82 PowR, 77 PowL, 73 Speed

R. Hernandez 78 OVR, 70 ConR, 71 ConL, 63 PowR, 64 PowL, 60 Speed

J. Rutledge 78 OVR, 68 ConR, 73 ConL, 64 PowR, 69 PowL, 82 Speed

C. Nelson 77 OVR, 68 ConR, 73 ConL, 57 PowR, 62 PowL, 79 Speed

J. Herrera, 75 OVR, 69 ConR, 65 ConL, 48 PowR, 43 PowL, 78 Speed

J. Pacheco 74 OVR, 68 ConR, 83 ConL, 39 PowR, 54 PowL, 72 Speed

D. LeMahieu 74 OVR, 66 ConR, 71 ConL, 48 PowR, 53 PowL, 72 Speed

W. Rosario 72 OVR, 62 ConR, 50 ConL, 88 PowR, 76 PowL, 65 Speed

C. Blackmon 71 OVR, 68 ConR, 63 ConL, 48 PowR, 43 PowL, 81 Speed

A. Brown 67 OVR, 56 ConR, 61 ConL, 59 PowR, 64 PowL, 70 Speed

[wudl83]

The only real problems left are the following:

- aggression for baserunning

- fielding range

- fielding ability at a certain position

and

- pitch ratings

Aggression, fielding range, fielding ability can somehow be read off from the stats, but pitch ratings are very difficult to come up with.

1) The game eninge seems to use some formulas for rating the pitches. Every pitch is rated by a formula using movement, control and speed. But that does not mean that for every pitch it is the same formula. That is okay compared to reality since in real life movement matters more for some pitches than it does for others.

2) I thought of using the max speed -1 the guys over fangraphs have listed it for the speed rating of the pitches ingame. When I do this I have to adjust the control and movement analog (more or less manual) that the pitch rating fits the pitch value/100 stat of fangraphs.

When I do this the OVR are affected because OVR is influenced by control, which I have to use for rating the pitches.

It is really difficult to come up with 100% pitcher's ratings, but I will do my best. When I have finished the first Giants pitchers I will post them here.

[wudl83]

Okay here come the pitcher ratings for the Giants:

S. Romo 90 OVR, 86 Control, 36 Stamina, 90 Composure, 91 ConR, 87 ConL, 85 PowR, 81 PowL

M. Cain 90 OVR, 85 Control, 99 Stamina, 82 Composure, 91 ConR, 87 ConL, 86 PowR, 82 PowL

M. Bumgarner 88 OVR, 96 Stamina 81 Control, 80 ConR, 86 ConL, 80 PowR, 86 PowL

R. Vogelsong 87 OVR, 79 Control, 94 Stamina 87 Composure, 86 ConR, 85 ConL, 83 PowR, 82 PowL

S. Casilla 85 OVR, 73 Control, 43 Stamina, 85 Composure, 89 ConR, 79 ConL, 84 PowR, 74 PowL

T. Lincecum 85 OVR, 76 Control, 92 Stamina, 78 Composure, 82 ConR, 82 ConL, 83 PowR, 83 PowL

J. Affeldt 80 OVR, 76 Control, 43 Stamina, 78 Composure, 73 ConR, 81 ConL, 81 PowR, 89 PowL

J. Mijares 79 OVR, 77 Control, 35 Stamina, 68 ConR, 78 ConL, 73 PowR, 83 PowL

B. Zito 79 OVR, 71 Control, 85 Stamina, 77 Composure, 76 ConR, 80 ConL, 70 PowR, 74 PowL

J. Lopez 77 OVR, 73 Control, 30 Stamina, 81 Composure, 68 ConR, 80 ConL, 84 PowR, 96 PowL

G. Kontos 77 OVR, 66 Control, 50 Stamina, 73 Composure, 79 ConR, 74 ConL, 79 PowR, 74 PowL

G. Mota 75 OVR, 65 Control, 57 Stamina, 69 Composure, 72 ConR, 67 ConL, 75 PowR, 70 PowL

Dodgers pitchers:

C. Kershaw 92 OVR, 89 Control, 99 Stamina, 85 Composure, 93 ConR, 98 ConL, 81 PowR, 86 PowL

K. Jansen 90 OVR, 88 Control, 46 Stamina, 84 Composure, 91 ConR, 87 ConL, 86 PowR, 81 PowL

R. Belisario 85 OVR, 78 Control, 46 Stamina, 77 Composure, 90 ConR, 80 ConL, 83 PowR, 73 PowL

J. Beckett 84 OVR, 73 Control, 90 Stamina, 79 Composure, 82 ConR, 78 ConL, 77 PowR, 73 PowL

C. Capuano 82 OVR, 73 Control, 87 Stamina, 81 Composure, 74 ConR, 79 ConL, 71 PowR, 76 PowL

B. League 82 OVR, 80 Control, 44 Stamina, 72 Composure, 78 ConR, 73 ConL, 85 PowR, 80 PowL

S. Elbert 80 OVR, 70 Control, 39 Stamina, 88 Composure, 82 ConR, 83 ConL, 82 PowR, 83 PowL

A. Harang 80 OVR, 68 Control, 85 Stamina, 81 Composure, 75 ConR, 75 ConL, 80 PowR, 80 PowL

R. Choate 79 OVR, 81 Control, 34 Stamina, 76 Composure, 75 ConR, 90 ConL, 69 PowR, 84 PowL

J. Wright 79 OVR, 70 Control, 49 Stamina, 74 Composure, 75 ConR, 75 ConL, 78 PowR, 78 PowL

J. Blanton 77 OVR, 62 Control, 89 Stamina, 71 Composure, 63 ConR, 63 ConL, 73 PowR, 73 PowL

S. Tolleson 71 OVR, 64 Control, 46 Stamina, 72 Composure, 75 ConR, 60 ConL, 75 PowR, 60 PowL

Some Rockies pitchers:

R. Betancourt 86 OVR, 82 Control, 40 Stamina, 80 Composure, 90 ConR, 80 ConL, 80 PowR, 70 PowL

M. Belisle 82 OVR, 78 Control, 47 Stamina, 76 Composure, 70 ConR, 69 ConL, 87 PowR, 86 PowL

J. Roenicke 78 OVR, 68 Control, 59 Stamina, 75 Composure, 82 ConR, 79 ConL, 87 PowR, 84 PowL

D. Pomeranz 78 OVR, 72 Control, 76 Stamina, 73 Composure, 64 ConR, 84 ConL, 64 PowL, 84 PowL

J. Francis 77 OVR, 70 Control, 80 Stamina, 76 Composure, 60 ConR, 66 ConL, 71 PowR, 77 PowL

M. Reynolds 77 OVR, 66 Control, 36 Stamina, 82 Composure, 78 ConR, 79 ConL, 68 PowR, 69 PowL

J. Chacin 77 OVR, 63 Control, 89 Stamina, 84 Composure, 81 ConR, 73 ConL, 77 PowR, 69 PowL

R. Brothers 76 OVR, 69 Control, 40 Stamina, 77 Composure, 69 ConR, 81 ConL, 72 PowR, 84 PowL

C. Torres 73 OVR, 63 Control, 60 Stamina, 69 Composure, 63 ConR, 63 ConL, 77 PowR, 77 PowL

T. Chatwood 73 OVR, 56 Control, 75 Stamina, 75 Composure, 65 ConR, 59 ConL, 79 PowR, 73 PowL

A. White 69 OVR, 52 Control, 73 Stamina, 77 Composure, 65 ConR, 63 ConL, 66 PowR, 64 PowL

W. Harris 68 OVR, 56 Control, 49 Stamina, 69 Composure, 63 ConR, 61 ConL, 71 PowR, 69 PowL

Regarding the OVR I feel quite fine.

If you have any comments, opinions, whatever on the posted ratings, let me know it. Perhaps I can tweak some formulas for better results.

Edited September 11, 2012 by wudl83

[Qbrick808]

The only real problems left are the following:

- aggression for baserunning

- fielding range

- fielding ability at a certain position

BASERUNNING AGGRESSIVENESS -- you might try this formula, or perhaps tweak it a bit to something which you prefer.

The logic is to measure SB attempts (with more credit applied for trying to steal 3B or Home) against SB Opportunities. Then you add in some more credit for Triples relative to total Hits, and finally you add the percentage for Extra Bases Taken.

-------- (((SB2 + CS2) + ((SB3 + SBH + CS3 + CSH) * 2)) / SBO) + (3B / H) + XBT%

To see it in action, let's compare 3 Angels in 2012 to see how this looks when you do the math:

ALBERT PUJOLS--

-------- ((( 7 + 0 ) + (( 1 + 0 + 1 + 0 ) * 2 )) / 189 ) + ( 0 / 151 ) + 47%

-------- ( 11 / 189 ) + 0 + 0.47

-------- 0.0582 + 0.47

-------- Albert Pujols has a relative Baserunning Aggressiveness Rating of 0.5282

ERICK AYBAR--

-------- ((( 13 + 3 ) + (( 1 + 1 + 0 + 0 ) * 2 )) / 181 ) + ( 5 / 127 ) + 55%

-------- ( 20 / 213 ) + 0.0394 + 0.55

-------- 0.0939 + 0.0394 + 0.55

-------- Erick Aybar has a relative Baserunning Aggressiveness Rating of 0.6833

MIKE TROUT--

-------- ((( 39 + 3 ) + (( 6 + 1 + 0 + 0 ) * 2 )) / 217 ) + ( 6 / 158 ) + 62%

-------- ( 56 / 217 ) + 0.0380 + 0.62

-------- 0.2581 + 0.038 + 0.62

-------- Mike Trout has a relative Baserunning Aggressiveness Rating of 0.9161

As you can see, the numbers bear out the truth when comparing these three players. If you apply this formula to every player, you can then scale the results on some form of curve to translate these rating numbers into the numbers you actually want to assign for the game.

FIELDING RANGE -- This is always a tough one to try and quantify with straight numbers.

To keep things simple, you might want to just use a comparison of Range Factor per 9IP and the league average per position.

For each position a player plays, use the formula RF/9 - lgRF9.

If we go back to our three Angels, we can see the results:

ALBERT PUJOLS--

---- as a 1B: 9.78 - 9.31 = +0.47

---- as a 3B: 2.30 - 2.59 = -0.29

ERICK AYBAR--

---- as a SS: 4.55 - 4.45 = +0.10

MIKE TROUT--

---- as a CF: 2.70 - 2.58 = +0.12

---- as a LF: 2.06 - 1.96 = +0.10

---- as a RF: 1.50 - 2.03 = -0.53

Since the interface doesn't allow you to input different numbers for the Range at each position, you can apply the formula to the total defense for each player, but this leads right in to your question about...

FIELDING ABILITY AT SPECIFIC POSITIONS -- This may lead to a lot of work, but that seems to be what you're dealing with anyway so I'm sure your attitude is that you might as well make it as accurate as possible.

Let's look at a guy who plays multiple positions with differing results.

Scott Hairston has played all three OF spots for the Mets this season; 51 games in LF, 37 in RF and 10 in CF. It's worth pointing out that when factoring in how much a guy plays at a given position, it is probably a better statistical approach to use the number of Chances he received rather than the game total; Hairston has had 57 Chances in LF, 58 in RF and 11 in CF (you see how deceiving those numbers can be if you just go by games). For each of the defensive categories (Glove, Arm Str, Arm Acc, Rng and Ant) you're going to have a formula or method of arriving at a representative number. So let's take the example I just gave for Range and apply it to Scott Hairston in the three OF spots.

---- as a LF: 1.52 - 1.83 = -0.31

---- as a RF: 2.11 - 2.01 = +0.10

---- as a CF: 1.77 - 2.51 = -0.74

So very clearly when it comes to Range, Hairston deserves a much better rating as a RF than anywhere else. If you did something like this with the other four defensive categories, you could merge them into one number with an appropriate formula. Then you could use the number of Chances each player had at each position as some form of multiplier to determine how significantly his above/below the league average standing should affect the number you give him for the position (which is then used as a multiplier against all of the other ratings by the program's original coding).

Edited September 11, 2012 by Qbrick808

[olipoo]

wait a second. can we define "BASERUNNING AGGRESSIVENESS" ??

from that formula, i'm seeing stolen bases at 2nd, stolen bases at 3rd, stolen bases at home, caught stealing, and triples over hits (i think?), which, if i'm understanding it correctly, does not take into account runners who are already on base, for example going first to third, or going first to home, or second to home.

Albert Pujols by no means is a good base stealer or a fast runner, but i will say after watching him for 11 years in St. Louis, he is a very aggressive, and very savvy baserunner. Jose Oquendo (third base coach) meant NOTHING to Albert rounding third, he had made up his mind already. i've seen Albert run through so many signs, and take more chances than ever.

sorry to make it complicated, but there is no stat for that. Albert is as aggressive of a baserunner as i've seen, and i think there are more things that go into baserunning aggressiveness.

but now that i really think about it more, in terms of the video game trying to think correctly, forget what i said maybe. it's really just trying to make the best simulation of how many bases he would steal, or how many triples he'll collect. i'm a big believer that baseball is purely a game of stats at the end of the day, but i'm a bigger believer that baseball is a game of intangibles at the end of the day.

[Qbrick808]

wait a second. can we define "BASERUNNING AGGRESSIVENESS" ??

from that formula, i'm seeing stolen bases at 2nd, stolen bases at 3rd, stolen bases at home, caught stealing, and triples over hits (i think?), which, if i'm understanding it correctly, does not take into account runners who are already on base, for example going first to third, or going first to home, or second to home.

Albert Pujols by no means is a good base stealer or a fast runner, but i will say after watching him for 11 years in St. Louis, he is a very aggressive, and very savvy baserunner. Jose Oquendo (third base coach) meant NOTHING to Albert rounding third, he had made up his mind already. i've seen Albert run through so many signs, and take more chances than ever.

sorry to make it complicated, but there is no stat for that. Albert is as aggressive of a baserunner as i've seen, and i think there are more things that go into baserunning aggressiveness.

but now that i really think about it more, in terms of the video game trying to think correctly, forget what i said maybe. it's really just trying to make the best simulation of how many bases he would steal, or how many triples he'll collect. i'm a big believer that baseball is purely a game of stats at the end of the day, but i'm a bigger believer that baseball is a game of intangibles at the end of the day.

No, you're completely missing it. What you're describing is by far the single largest component of that formula.

XBT% = Extra Bases Taken Percentage

It's the reason a guy like Pujols has just 8 Stolen Bases and 0 Triples yet shows a very respectable number.

XBT% is defined as the Percentage of times a runner advances more than one base on a single, or more than two bases on a double, when it is possible for him to do so.

That means going from 1st to 3rd on a single, scoring from 1st on a double, or even scoring from 2nd on a single.

This is the exact stat for what you were describing but I guess you simply didn't realize it exists.

And like I said, in this formula I mentioned above, XBT% alone significantly outweighs all of the other components combined.

Look at the breakdown for each player I used as examples:

----Albert Pujols gets a rating of 0.53. He's getting 0.06 of that from SBs, nothing from Triples, and 0.47 from XBT%.

----Erick Aybar gets a rating of 0.68. He's getting 0.09 of that from SBs, 0.04 from Triples, and 0.55 from XBT%.

----Mike Trout gets a rating of 0.92. He's getting 0.26 of that from his many SBs, 0.04 from Triples, and 0.62 from XBT%.

I think if anything, some people will probably feel the taking of an extra base is too dominant for this rating, but I think it's pretty fair.

[olipoo]

ahhh gotcha. i didn't read it correctly, and was under the impression it was Extra Base Hits not Extra Bases Taken.

[wudl83]

Qbrick, this is serious business! I thank you so much!

I am not at home right now. But when I come home I will definately try it out. Will cost some time, since it can not be read out with fangraphs but has to be put in via baseball-reference, nevertheless it looks good.

I quickly scanned some fielding ratings for Giants players and I think for fielding range I will simply put a certain base number for every position, e.g. 65 for a regular first baseman, 75 for a regular 2nd baseman, 80 for a regular center fielder, etc. and depending on the difference between the RF/9 and the lgRF/9 I will rate the player's range up or down. When I calculate some ratings mental they seem to represent the reality really well. The Giants are my bread and butter, so this is quite safe to say. ;)

Edited September 12, 2012 by wudl83

[wudl83]

Too bad it seems that there occur problems regarding my first thought of transfering RF/9 into a range rating.

If I did it this way then I would get players like Scutaro having a way higher range rating than Brandon Crawford. Or Dexter Fowler having a way smaller fielding rating than Hunter Pence.

But I think the RF/9 stat is pretty good to rate a players primary, secondary, and so on position.

Perhaps I got to find a formula which translates speed and instincts into the range rating? Something like: the higher the speed of a player and the better his instincts the more range he gets?

[Qbrick808]

Too bad it seems that there occur problems regarding my first thought of transfering RF/9 into a range rating.

If I did it this way then I would get players like Scutaro having a way higher range rating than Brandon Crawford. Or Dexter Fowler having a way smaller fielding rating than Hunter Pence.

But I think the RF/9 stat is pretty good to rate a players primary, secondary, and so on position.

Perhaps I got to find a formula which translates speed and instincts into the range rating? Something like: the higher the speed of a player and the better his instincts the more range he gets?

I think the issue you bring up regarding Pence and Fowler should be somewhat mitigated by the positions they play. There should be an inherently higher range base number for CF as opposed to RF and LF.

The aspect which needs to be compensated for, and the Crawford versus Scutaro example points it out very well, is that these numbers are dealing with such small and subtle differences which means there are going to be some instances which simply don't fit the mold. Crawford has properly handled 97.2% of his chances at SS this season while Scutaro has been good on 97.7% of his opportunities at the position. When dealing with such small margins between numbers, 0.5% is going to translate into a lot.

You mentioned adding speed into the equation, and while I don't really love the idea for infielders, it's probably the best method you can go with. In the OF, speed and range are obviously very related, but the range of an infielder is much more about quickness than speed, and as I'm sure you know those two traits are not the same thing at all.

Regardless, I could see tweaking the formula by determining an appropriate percentage of the player's speed rating and introducing that into the equation in a way which seems to work without throwing things out of proportion.

[MissionMaximus]

I was figuring out a formula for pitchers. still working on some things but will post kind a what I got as of now. This is in regards to pitchers.

Pitch Speed (velocity)> Fangraphs list all the pitchers pitch velocity average. So add +2 to each pitch average to get your max mph velocity.

Pitch Tendency>Fangraphs has the % number of each pitch thrown. If a pitcher has anything below 1% disregard that pitch.

Base on balls allowed> Multiply BB% by 5 then subtract by 100. example from fangraphs they have 4% base on balls listed as excellent so you would say a pitcher has a 4% base on ball rate on Fangraphs> 4 x 5=20, then 100 - 20=80 so that particular pitcher would have a 80 base on ball rating. Another example 10% is listed as awful base on balls on Fangraphs. 5 x 10=50, 100-50=50 so that pitcher would have a base on ball rating of 50.

Strikeout Tendency>Multiply K% by 3.3 Example: Pitcher has a K% of 25% you would multiply 25 x 3.3=82.5 round that off to 83, so that pitcher would have a 83 Strikeout Tendency. Another example 10% listed as awful on Fangraphs. 10 x 3.3=33 so that picture would have a 33 Strikeout Tendency.

Groundball Tendency> Go by GB% so if a picture has a 40% Groundball rate that would be his groundball tendency. 40 Groundball Tendency.

Contact LHB/RHB>Going by fangraphs pitching stats against LHB/RHB Subtract batting average by 100. Example .200 LHB Average 100 - 20=80 so that pitcher would have a 80 contact rating against LHB's.

Durability>Average games per year on DL last 3 years. Subtract from 100.

Stamina>Average pitches Per Game.

Still working on Composure, Clutch, etc.

Edited September 13, 2012 by MissionMaximus

[wudl83]

I think the issue you bring up regarding Pence and Fowler should be somewhat mitigated by the positions they play. There should be an inherently higher range base number for CF as opposed to RF and LF.

The aspect which needs to be compensated for, and the Crawford versus Scutaro example points it out very well, is that these numbers are dealing with such small and subtle differences which means there are going to be some instances which simply don't fit the mold. Crawford has properly handled 97.2% of his chances at SS this season while Scutaro has been good on 97.7% of his opportunities at the position. When dealing with such small margins between numbers, 0.5% is going to translate into a lot.

Yap, you are right.

And I tried it out with higher base numbers for CF than for the other OF, but let's take a look at Dexter Fowler (CF):

2010: RF/9 2.29 - lgRF/9 2.59 --> -0.30 (*1)

2011: RF/9 2.64 - lgRF/9 2.55 --> +0.09 (*2)

2012: RF/9 2.27 - lgRF/9 2.51 --> -0.24 (*3)

Since I weighed the different years one time, two time and three times (I said that earlier) we get:

((-0.30)+(0.18)+(-0.72))/6 = -0,14

This means that (using the weighed average) difference between Dexter Fowler's RF/9 and the lgRF/9 he is -0,14.

Let's look at Tyler Colvin (CF):

2010: RF/9 2.25 - lgRF/2.59 --> -0.34 (*1)

2011: RF/9 2.55 - lgRF/2.55 --> 0.00 (*2)

2012: RF/9 3.04 - lgRF/2.51 --> +0.53 (*3)

Calculation: ((-0.34)+0+(1,59))/6 = 1,25

See what I mean? And those numbers occur for so many players.

When we compare this to reality, you will see in the Rockies games that Dexter Fowler has a so so arm, but has plus speed and a really big range. And you will see that Tyler Colvin's arm is okay, but his range isn't the best.

It is like day and night.

You mentioned adding speed into the equation, and while I don't really love the idea for infielders, it's probably the best method you can go with. In the OF, speed and range are obviously very related, but the range of an infielder is much more about quickness than speed, and as I'm sure you know those two traits are not the same thing at all.

Regardless, I could see tweaking the formula by determining an appropriate percentage of the player's speed rating and introducing that into the equation in a way which seems to work without throwing things out of proportion.

Yap, you are right. In the OF it seems valid. The range stats sometimes do not reflect what we can see in reality. So I think the idea of somehow calculating the OF range by using speed and anticipation may be ok.

Regarding IF I am undecided. Maybe I could use fangraphs "Range Runs" (http://www.fangraphs...ion=SS#fielding)? It seems that this stat reflects reality a little bit more. When I compare Brandon Crawford to Marco Scutaro at SS I see a rather real number:

(Year/Crawford/Scutaro)

2010: --- / -1.5

2011: 2.6 / 0.9

2012: 4.2 / -0.3

As in reality, Crawford would get a way better range rating than Scutaro at SS or IF in general.

I was figuring out a formula for pitchers. still working on some things but will post kind a what I got as of now. This is in regards to pitchers.

Pitch Speed (velocity)> Fangraphs list all the pitchers pitch velocity average. So add +2 to each pitch average to get your max mph velocity.

Pitch Tendency>Fangraphs has the % number of each pitch thrown. If a pitcher has anything below 1% disregard that pitch.

Base on balls allowed> Multiply BB% by 5 then subtract by 100. example from fangraphs they have 4% base on balls listed as excellent so you would say a pitcher has a 4% base on ball rate on Fangraphs> 4 x 5=20, then 100 - 20=80 so that particular pitcher would have a 80 base on ball rating. Another example 10% is listed as awful base on balls on Fangraphs. 5 x 10=50, 100-50=50 so that pitcher would have a base on ball rating of 50.

Strikeout Tendency>Multiply K% by 3.3 Example: Pitcher has a K% of 25% you would multiply 25 x 3.3=82.5 round that off to 83, so that pitcher would have a 83 Strikeout Tendency. Another example 10% listed as awful on Fangraphs. 10 x 3.3=33 so that picture would have a 33 Strikeout Tendency.

Groundball Tendency> Go by GB% so if a picture has a 40% Groundball rate that would be his groundball tendency. 40 Groundball Tendency.

Contact LHB/RHB>Going by fangraphs pitching stats against LHB/RHB Subtract batting average by 100. Example .200 LHB Average 100 - 20=80 so that pitcher would have a 80 contact rating against LHB's.

Durability>Average games per year on DL last 3 years. Subtract from 100.

Stamina>Average pitches Per Game.

Still working on Composure, Clutch, etc.

I will post my opinion and then come up with what I did,

Pitch Speed - I first used fangraphs average and tried to add +1 or +2 to the number. Won't work properly. Look at Justin Verlander.http://www.fangraphs...ion=P#pitchtype He averages 94.6 mph on his fastball, which would be a 96.6 (=97) but he is able to throw 99ers or 100 mph fastballs in his games. When you look here you see that Verlander is able to reach nearly 102 mph: http://www.fangraphs...tion=P&pitch=FA. So I used this "max-Vel", substracted -1 and used that as the speed for a pitch.

Pitch Tendency - I do the same like you.

Base on Ball tendency - I think you misunderstood the ingame rating. The lower the BoB tendency a pitcher has ingame the lower his walks will be. The higher his BoB tendency the more walks he will issue. I for myself used a somehow more complicated formula, which is also a little bit different for relievers and starters.

First I calculated the weighed (you know 3 times, 2 times, 1 time) average for league BB for last 3 seasons. This would be 3,147.

Then I took the weighed 3 year average for the pitcher. Let's do it with Matt Cain (http://www.fangraphs...4732&position=P). He got (2010) 2,46 BB/9, (2011) 2,56 BB/9 and a projected (2012) 1,89 BB/9. That makes up for a 2,208 BB/9 weighed average.

Now I simply used the formula: 70 + 25 * (weighed average of the pitcher - weighed average of the league).

In Matt Cain's case we would come up with: 70 + 25 * (2,208-3,147) = 46,53.

So we would have a BB/9 tendency of 47 for Matt Cain.

Do not wonder about "70" and "25", those numbers are quite random but guarantee accurate results. Do not know why, but it works.

For relievers the "70" is exchanged with "50". Sometimes relievers with low BB/9 numbers would get BB tendencies in the 10s or so, but in those cases I feel free to add +10 to this number. Sergio Romo would get a tendency of 13 using this formula, but I add +10 and so he gets 23.

Strikeout tendency - Could it be that you took some strange numbers? I do not remember pitchers having a K% of 25 or something like that. Ingame a high strikeout tendency means that a pitcher normally gets many Ks.

First I calculated the league's weighed 3 year average (7,246 K/9), then I calculated the weighed 3 year average for the pitcher.

Let's take Tim Lincecum: 9,79 K/9 in 2010, 9,12 K/9 in 2011, 9,34 K/9 in 2012. So we would get a weighed 9,34 K/9 for his last 3 seasons.

After that I use the formula: 60 + 15 * (weighed average of the pitcher - weighed average of the league). Again: Do not care about 60 and 15.

So Lincecum would get: 60 + 15 * (9,342-7,246) = 91,34.

So Lincecum gets a strikeout tendency of 91.

Since relievers tend to strike out players more often than starters I changed the formula into:

65 + 5 * (weighed average of the pitcher - weighed average of the league).

Sergio Romo has a weighed average of 11,25 K/9 for the last 3 seasons, he would get a 85 strikeout tendency.

Groundball tendency - same like you. Easiest number for pitchers.

Contact - I do not like to use avg. for this for some reason, I use H/9 instead. Problem with your approach is that .200 is a very damn good number for SPs, while many relievers (e.g. left-handed specialists like Randy Choate, finess bullpen guys like Romo or boombastic closers like Kimbrel) have a much lower number than 200. To put it in a nutshell I find that a starting pitcher who has a .200 BAA should have a much higher contact rating than 80, while it could be that a reliever with .200 BAA could have a rating in the 80s.

Again I first calculated the 3 year weighed average for H/9 for the whole league (8,781).

Then I calculatd the 3 year weighed average for the respective pitcher.

For starters I use the formula: 80 + 10 * (weighed league average H/9 - weighed pitcher average H/9).

Relievers: 75 + 10 * (weighed league average H/9 - weighed pitcher average H/9).

Some examples for starters.

• Tim Lincecum (8,217 H/9): 80 + 10 * (8,781 - 8,217) = 85,63 --> 86 BASE contact rating.
  
• Let's take Joe Blanton (10,502 H/9) of the Dodgers for a comparison: 80 + 10 * (8,781 - 10,502) = 62,782 --> 63 BASE contact rating.
  
• Or Josh Johnson (7,492 H/9) of the Marlins: 80 + 10 * (8,781 - 7,492) = 9,288 --> 93 BASE contact rating.
  

You get it: Josh Johnson doesn't allow many H/9 so he gets a very good number for it. Tim Lincecum struggled this year, so his rating is smaller than Johnson's. Blanton really is a joke of a starting pitcher, allowing many hits, so he gets the worst number.

Some relievers:

• Jeremy Affeldt (8,151 H/9): 75 + 10 * (8,781 - 8,151) = 81,29 --> 81 BASE contact rating.
  
• JJ Putz (7,175 H/9): 75 + 10 * (8,781 - 7,175) = 91,06 --> 91 BASE contact rating.
  
• Kameron Loe (8,573 H/9): 75 + 10 * (8,781 - 8,573) = 75,57 --> 76 BASE contact rating.
  

It may be that some relievers could get a calculated number above 100. For those I feel free to substract -5 or -10. Sometimes relievers would get numbers about 50. For those I feel free to add +5 or +10. I was not able to find a better formula. For starters it works nearly perfect, but for relievers it is very difficult. The stats for relievers are very widespread. But I can live with that.

Now let's talk about what I mean with BASE contact rating. When we look at the split stats of the pitchers we will see that some (example) right handed starters may have one bad year against right handed batters, while in the same year they shine against left handed batters (e.g. Matt Cain last season as far as I remember). But when we look at their career split it may be that this starter is overall better against righties than he is against lefties. Got me?

And not only to talk about starters, but relievers have this problem still more. Since relievers sometimes only have a few at bats versus opposite hitters (e.g. LHP against RHB) over the course of a season it may screw up their stats. So I find it very inaccurate to calculate pitching the contact ratings vs. LHB and RHB independent.

So I think it is a better method to look at the career split of a pitcher and then calculate the contact vs RHB or LHB based on the career split.

Let's look at Tim Lincecum. Career split vs RHB .224 BAA, career split vs LHB .224 BAA. His numbers are indeed identic, so he gets a 86 contact rating against LHB and 86 against RHB.

Joe Blanton: .269 career BAA against LHB, .271 career BAA against RHB. Nearly identical, Blanton gets 63 for both.

Josh Johnson: .249 career BAA against LHB, .224 career BAA against RHB. Finally we got a pitcher that has different success vs LHB and RHB over his career. I found out that it would be a good method to substract the two numbers from each other and multiply the result by *2. After that I use this number and substract it from the BASE contact rating. So let's do it: .249 - .224 = .025. Then let's take 0.25 as a 2.5 and multiply it by 2 and we get 5. So Josh Johnson would get a 93 contact rating against RHB and (93 - 5) 88 contact rating against LHB.

I know it is not the best way but when I would calculate the numbers separately it would be screwed up.

Double - I can not calculate everything. For pitchers doubles rating I make it easy. When a player has a high groundball rate I assume he does not give up many doubles. So I add a certain number (for pitchers with 55 GB% about +5) to his contact rating or I substract a certain number (for pitchers with 35 GB% about -5) of his contact rating.

Triples - It seems that every pitcher ingame had a rating of 25. So I let it be.

Homeruns - I calculate a BASE homerun rating and according to the career BAA split difference I substract this difference from the BASE homerun rating, too. Otherwise there would be this problem where the stats and then the ratings would get awkward.

Formula for both starters and relievers: 99 - 20 * ( weighed league average HR/9 - (weighed league average HR/9 - weighed pitcher average HR/9)

Examples:

• Tim Lincecum (weighed 0,82 HR/9): 99 - 20 * ( 0,974 - ( 0,974 - 0,82) = 82,576 BASE homerun rating
  
• Josh Johnson (weighed 0,462 HR/9): ... 90 BASE homerun rating
  
• JJ Putz (weighed 0,74 HR/9): ... 84 BASE homerun rating
  
• Kevin Gregg (weighed 0,99 HR/9): ... 79 homerun rating
  
• Matt Reynolds (weighed 1,52 HR/9): ... 69 homerun rating
  

Durability - Is there anywhere a statistic which tells us how often a players was on the DL? As far as I (really!) like your idea I do not know of such a statistical overview.

But what I do know is that the durability rating (not only stamina) ingame does affect somehow the regeneration of a player. For that starters with many IP per season should get a higher durability rating than relievers.

For that I use different formulas.

Starters: 70 + (weighed 3 year start-IP * 0,1)

Relievers: 50 + (reliever's stamina rating * 0,3

Stamina - Starters stamina: (weighed 3 year start-IP / weighed 3 year GS) * 15

Relievers stamina: (weighed 3 year relief-IP / (G - GS) ) * 45

Some results:

• Josh Johnson: 94 stamina, 85 durability
  
• Tim Lincecum: 92 stamina, 90 durability
  
• Sergio Romo: 36 stamina, 61 durability
  
• JJ Putz: 43 stamina, 63 durability
  

Composure - Since I think that ingame the differencies regarding some ratings between the good and the bad are way too big, I simply use those formulas:

Starters: 80 + 100 * (weighed 3 year LOB% of the player - weighed 3 year LOB% of the league)

Relievers: 75 + 100 * (weighed 3 year LOB% of the player - weighed 3 year LOB% of the league)

Examples:

• Josh Johnson: 84 composure
  
• Tim Lincecum: 80 composure
  
• Sergio Romo: 89 composure
  
• JJ Putz: 81 composure
  

Since ingame the composure rating stands for nothing else than how fast a players gets rattled when he allows guys to get on base I think LOB% is a valid choice.

Clutch - quite easy once again. Fangraphs offers a clutch rating for every player that determines how a player over- or underperforms in high leverage situations.

Starters and relievers: (5 + 3 year weighed player average * 2,5) * 10

Two last examples:

• JJ Putz: 44 clutch
  
• Tim Lincecum: 56 clutch
  

I do not know anything more for the moment. If you have any questions feel free to ask me

Edited September 13, 2012 by wudl83

[wudl83]

I played around with some batting formulas. Still not finished. The places that are empty still need tweaking. The rest looks quite good. I will calculate the Contact, 2B, HR and Eye via the BASE ratings and career splits I mentioned while talking about the pitchers.

Here are some players:

Name: Jose Bautista

Clutch: 44

Durabil

Contact 74

2B 73

3B 56

HR 100

Eye 88

Bunt

DragBunt

Speed 78

Aggr 55

Awar 70

Accel 79

Bu Tend

K Tend 59

GB Tend 41

BB Tend 94

SB Tend 42

Glove 76

ArAcc 90

ArStr 93

OF Rng

[wudl83]

I played around with some batting formulas. Still not finished. The places that are empty still need tweaking. The rest looks quite good. I will calculate the Contact, 2B, HR and Eye via the BASE ratings and career splits I mentioned while talking about the pitchers.

Here are some players:

• Name: Jose Bautista
  
• Clutch - 44
  
• Durabil -
  
• Contact - 74
  
• 2B - 73
  
• 3B - 56
  
• HR - 100
  
• Eye - 88
  
• Bunt -
  
• DragBunt -
  
• Speed - 78
  
• Aggr - 55
  
• Awar - 70
  
• Accel - 79
  
• Bu Tend -
  
• K Tend - 59
  
• GB Tend - 41
  
• BB Tend - 94
  
• SB Tend - 42
  
• Glove - 76
  
• ArAcc - 90
  
• ArStr - 93
  
• OF Rng - 78
  
• IF Rng - 65
  
• Antic - 79
  
• C -
  
• 1B -
  
• 2B -
  
• 3B -
  
• SS -
  
• LF -
  
• CF -
  
• RF -
  
• Bl Plate -
  
• Bl Pitch -
  
• Game Call -
  

• Name: Hunter Pence
  
• Clutch - 65
  
• Durabil -
  
• Contact - 79
  
• 2B - 79
  
• 3B - 57
  
• HR - 71
  
• Eye - 77
  
• Bunt -
  
• DragBunt -
  
• Speed - 83
  
• Aggr - 60
  
• Awar - 71
  
• Accel - 81
  
• Bu Tend -
  
• K Tend - 67
  
• GB Tend - 56
  
• BB Tend - 57
  
• SB Tend - 38
  
• Glove - 78
  
• ArAcc - 84
  
• ArStr - 87
  
• OF Rng - 84
  
• IF Rng -
  
• Antic - 78
  
• C -
  
• 1B -
  
• 2B -
  
• 3B -
  
• SS -
  
• LF -
  
• CF -
  
• RF -
  
• Bl Plate -
  
• Bl Pitch -
  
• Game Call -
  

• Name: Derek Jeter
  
• Clutch - 47
  
• Durabil -
  
• Contact - 84
  
• 2B - 83
  
• 3B - 35
  
• HR - 56
  
• Eye - 80
  
• Bunt -
  
• DragBunt -
  
• Speed - 75
  
• Aggr - 64
  
• Awar - 84
  
• Accel - 65
  
• Bu Tend -
  
• K Tend - 48
  
• GB Tend - 69
  
• BB Tend - 52
  
• SB Tend - 43
  
• Glove - 80
  
• ArAcc - 82
  
• ArStr - 76
  
• OF Rng - 48
  
• IF Rng - 68
  
• Antic - 71
  
• C -
  
• 1B -
  
• 2B -
  
• 3B -
  
• SS -
  
• LF -
  
• CF -
  
• RF -
  
• Bl Plate -
  
• Bl Pitch -
  
• Game Call -
  

This was calculated via fangraphs and baseball reference. I will try to keep it simpler so that I only have to use one page. Fangraphs is my preference because with this page I can extract (hopefully) all necessary stats while I had to make it manually for every player with baseball reference.

Edited September 13, 2012 by wudl83

[MissionMaximus]

The base on ball tendency I listed should be base on balls allowed, my mistake. I feel going by K/9 for strikeouts would inflate some relievers K% also the same in regards to BB/9. Going by K% & BB% would eliminate that because some pitchers face more batters than others.,What I was referring to inregards to K% and BB% was this scale on FanGraphs in regards to k/9, k% and bb/9, bb%http://www.fangraphs...<br /><br />

In regards to Relief Pitchers I find this statistic of great value when evaluating Relief pitchers.

http://www.fangraphs...pitching/sd-md/

Edited September 13, 2012 by MissionMaximus