In the original set, there were three starters who were graded A&C: Bob Gibson (plus an XZ), 31-game winner Denny McLain (XZ), and AL ERA champ (1.60) Luis Tiant (XYZ). In the older formulas, McLain was an “automatic” A&C since he had 30+ wins. At this point, the A&C as a starting grade had only been given once before — two years earlier to Sandy Koufax — although it had been given in the past to OFAS pitchers. To somewhat counter the high number of A and A&C pitchers, cards were given an extra 7.

It would be pretty impossible to get a 1.12 ERA from an A&C, even in the deflated year of 1968. A baseline was set with Tiant getting the A&C for 1.60, and you would think there would need to be some sort of improvement to get to 1.12. The one basic game replay I could find had Gibson’s ERA at 1.55.

In the reissued set, Gibson got an upgrade to A&B (MG 25), but he lost the Z. McLain also got a bump to 25, kept his Z but gained an L (he did give up 31 dingers to lead the AL). Tiant stayed at A&C (MG 20) but lost his Z as well.

In the modern card-making formulas, the loss of Gibson’s Z is not unexpected. The older formulas used BB per 9 innings, and the line was somewhere around 3.0. Gibson, at 1.8, had no problem getting a Z. However, modern formulas use unintentional walks per batter faced. The line for the 1968 NL set was around 0.046. Ron Reed (0.045) got a Z, while Ray Washburn (0.047) did not. Gibson was at 0.048. Since Gibson faced so few batters, by using BB/9 there was a somewhat inflated sense of having above-average control.

Of course, with BBW and League Manager, we can test to see how well Gibson’s new grade and now average control worked. I replayed the 1968 season 10 times, using Duke Robinson as the manager, adding all players to a roster somewhere, and limiting AIM to game injuries only. Here are the overall league stats (both NL and AL) of the 10 tests:

I present the overall stats to see if there is anything is out of whack compared to the real life. The first thing that stands out is that the replays produced on average 0.5 more runs per game despite the averages pretty much on target. This is mostly an effect of not enough double plays, as the average replay had 1.45 DP per game as opposed to 1.74 in real life. Not trying to go too deep into the weeds, but I would assume that the formulas to determine 24’s was not adjusted to a lower amount of base runners to trigger double plays.

So routing back to Gibson, it looks like the changes to the card are right on the money:

Since it is impossible for him to get the proper mix of singles vs. extra base hits, fewer singles are produced to compensate for the hard-to-stop doubles and triples. The G also helped somewhat, as only once was he above average in HR. The Micromanager did pull him early, so he did not get as many innings as he should have. His BB/PA was 0.053, obviously higher than his real life 0.048, but not chart busting. Also, with Gibson rarely having runners on to begin with, the Z would have less effect than it would for the average pitcher.

However, there is someone whose card needs adjustment. And that’s Gibson’s AL counterpart: Denny McLain. The pitching between the two leagues was pretty indistinguishable, so you would expect that McLain would have a lower grade. However, as mentioned earlier, it is also a 25 and with a Z, but also with an L. Those subratings are going pretty much going to cancel each other out, and the only difference between him and Gibson is the G. However, that’s not enough of a difference to replicate McLain’s somewhat higher ERA (1.96):

The biggest problem is his home run total. The only time with a grade that high an L matters is with either runners on 1st-2nd or runners on 2nd-3rd. And with McLain not allowing a lot of baserunners, it really has no impact. He likely needs to be an M. His BB/PA, even with the Z, is higher (0.62) than Gibson’s (0.053), and I really don’t have an explanation for that. The Micromanager consistently gave him more starts, which I think would be to him being pulled more often than he should be (he is a Q4 for some bizarre reason) which allowed him to pretty much be on a 4-day rotation.

A lot of this is going to depend what you want to consider the benchmark for a successful recreation of his season. Trying to hit 31 wins in 41 starts is going to difficult, as his replay stats shows he only wins 70% of his starts, and 70% of 41 is only going to be 29 wins. The better thing is to get past the issue that he had 31 wins to begin with. The Tigers as an offense scored 4.15 runs per 9 innings, and McLain gave up 2.30 when accounting for unearned runs. Those numbers would normally mean a 25-8 season, but he got lucky in the fact he only had 4 no-decisions, he had 5.2 run support, and didn’t have a save blown by a reliever.

I did try one simulation with McLain at 22MXZ, and it was a little closer (even though he got 50 starts again): 30-9, 1.70 ERA, 30 HR and “only” 367 strikeouts. This is probably a better grade to use, although 30 wins is going to be highly unlikely.

And just for kicks, here are the leaders for each of the 10 simulations. The Tigers not once were lower than their win total. The AL had an under .300 batting champ in one season and an under 100-RBI champ in another. Also looks like Ferguson Jenkins strikeout benefits a little bit from being just on the X side of the X/Y line.

]]>- Did APBA assign more 26s? Nope, the average card in 2014 had 1.28, and the average card in 2015 had 1.29.
- Can players get a third 26? Nope, not a single card in either 2014 or 2015 had a third 26.

So as seen below, the second 26 had merely swapped positions, leaving 62 for 21 (the *x*-axis representing the play result and the *y*-axis representing the percentage of plate appearances with that particular play result):

However, on a swap, you would expect the now much lower counts of 30 and 32 to take the place the relocated 26. However, although there are some, it’s not complete. The sum of the 2014 fly balls on 21 is .51 vs the 2015 sum of .24, but the sum of the 2015 fly balls on 62 is .07, only a .04 increase from 2014’s .03.

Do we have less 30s and 32s? A touch less, but not much. The 30 saw a decrease of 1.51 to 1.44, and the 32 1.52 to 1.49. However, that’s only a total of .10. So where did they go?

It looks like 24 and 26 became more popular destinations, although their increased presence is more than overshadowed by the 13:

Since the number of 13s remain largely unchanged (5.93 in 2014, 5.94 in 2015), the ones displaced from 21, 24, and 26 scattered to other numbers, especially 42 and 56.

Before wrapping up, I just thought I would mention one other curiosity, that even though the play numbers of 24–32 on the cards in the 2015 set aren’t that different from 2014, the ratios of them seem to be somewhat different. Here’s a list of the top 10 sets from 2014 and 2015:

2014 Out Numbers | Percentage |
---|---|

25-26-27-28-29-30-31-32 | .024 |

24-25-26-27-28-29-30-31-32 | .023 |

25-26-27-28-29-30-30-31-32-32 | .020 |

24-25-26-27-28-29-30-31-31-32 | .018 |

25-26-27-28-28-29-30-31-31-32 | .018 |

25-26-27-28-28-29-29-30-30-31-31-32 | .018 |

25-26-27-28-28-29-30-30-31-31-32 | .016 |

25-26-27-27-28-28-29-29-30-31-31-32-32 | .016 |

24-24-25-26-27-29-30-31-32 | .016 |

24-25-26-26-27-27-28-29-29-30-30-31-31-32 | .016 |

2015 Out Numbers | Percentage |
---|---|

25-26-27-28-29-30-31-32 | .029 |

24-25-26-27-28-29-30-31-32 | .027 |

24-24-25-26-27-28-29-30-31-31-32 | .021 |

25-26-27-28-28-29-30-31-31-32 | .021 |

24-24-24-25-26-27-28-29-30-31-31-32 | .020 |

24-25-26-27-28-29-30-31-31-32-32 | .017 |

25-26-27-28-28-29-29-30-31-31-32 | .016 |

25-26-27-28-28-29-29-30-30-31-31-32 | .016 |

24-25-26-26-27-28-29-30-31-31-32-32 | .015 |

25-26-27-28-28-29-29-30-30-31-31-32-32 | .015 |

The #8 combination in 2014 disappears entirely, and #3 is barely used in 2015. I don’t know if this is something that happens every year, or just something unique to the 2015 set. Something to explore for another day.

]]>

The concept of the goalie card is simple. The card is called upon to redirect a saved puck or say a goal has scored. About 80% of the time, it will always be a redirection. About 20% of the time it will determine whether a goal has scored, and even a 0.2% difference is different from one card to the next. Each card contains the following sets of numbers:

- 1–6: the always save numbers
- 7–18: the sometimes a save, sometimes a goal numbers – a card will at most have one of these
- 19–45: the always a goal numbers

The stats you will need to calculate the card are not much and are available for many of the seasons:

- Games Played
- Minutes Played
- Save Percentage
- Assists Scored
- Penalties Caused
- Team Shots Allowed
- League Average Save Percentage

First, we’ll handle the top of the card.

- All players will get a Minutes rating of 60.
- For the Assist rating, figure out the assists per game, multiply by 56 and round down. Convert this number to an 11-66 base.
- For the Min/Maj/Mis ratings, unfortunately there aren’t too many players with the ratings to figure out a good pattern. This is my best guess, but could be way off:
- Min: figure out the minors per 60 minutes, multiply by 10, subtract 1 and round down. Convert this number to an 11-66 base. Players who are negative receive a 0.
- Maj: figure out the majors per 60 minutes, multiply by 50 and round up. Convert this number to an 11-66 base.
- Mis: figure out the miscounducts per 60 minutes, multiply by 20 and round up. Convert this number to an 11-66 base.

- For the J rating, use games played. 64+=J-0, 45-63=J-1, 34-44=J-2, 19-33=J-3, 1-18=J-4.
- All other ratings are given a 0.

And now for the play numbers:

- Subtract the league average save percentage from the goalie’s save percentage, multiply by 200, add 18.33, and round to the nearest third. This gives you the number of ‘save numbers’ the goalies need.
- Use the ‘save numbers’ with the chart below, the column that the number appears in is the play result that dice roll gets. For the ‘7-12‘ and ‘13-18‘ columns, the red number indicates what is assigned, as a goalie will not have more than one of these per card.
- If the ‘save number’ does not match any of the ranges, they receive the number in the ‘Other’ column.

Dice | 1 | 2 | 3 | 4 | 5 | 6 | 7-12 | 13-18 | Other |

11 | 12+ | 2 | |||||||

12 | 25 | ||||||||

13 | 17+ | 15-16⅔ | 14⅔=9 | 14⅓=15 | 37 | ||||

14 | 30 | ||||||||

15 | 23+ | 16-22⅔ | 12-15⅔ | 7-11⅔ | 6⅔=7 | 6⅓=13 | 25 | ||

16 | 28 | ||||||||

21 | 20+ | 19⅔=8 | 19⅓=14 | 23 | |||||

22 | 12+ | 3 | |||||||

23 | 21+ | 17-20⅔ | 16⅔=11 | 16⅓=17 | 20 | ||||

24 | 23+ | 19 | |||||||

25 | 25+ | 18-24⅔ | 12-17⅔ | 8-11⅔ | 7⅔=8 | 7⅓=14 | 26 | ||

26 | 25+ | 24⅔=7 | 24⅓=13 | note (a) | |||||

31 | 21+ | 16-20⅔ | 12-15⅔ | 10-11⅔ | 9⅔=10 | 9⅓=16 | 28 | ||

32 | 26 | ||||||||

33 | 18+ | 12-17⅔ | 3 | ||||||

34 | 31 | ||||||||

35 | 17+ | 15-16⅔ | 11-14⅔ | 10⅔=11 | 10⅓=17 | 29 | |||

36 | 24+ | 23⅔=11 | 23⅓=17 | 33 | |||||

41 | 24+ | 19-23⅔ | 18⅔=7 | 18⅓=13 | 24 | ||||

42 | 25+ | 18-24⅔ | 15⅔-17⅔ | 12-15⅓ | 11⅔=12 | 11⅓=18 | 30 | ||

43 | 29 | ||||||||

44 | 15+ | 12-14⅔ | 4 | ||||||

45 | 18+ | 17⅔=12 | 17⅓=18 | 21 | |||||

46 | note (b) | ||||||||

51 | 18+ | 15-17⅔ | 12-14⅔ | 9-11⅔ | 8⅔=9 | 8⅓=15 | 27 | ||

52 | 27 | ||||||||

53 | 25+ | 20-24⅔ | 16-19⅔ | 15⅔=10 | 15⅓=16 | 36 | |||

54 | 32 | ||||||||

55 | 18+ | 12-17⅔ | 6-11⅔ | 5⅔=6 | 5⅓=12 | 24 | |||

56 | 22+ | 21⅔=10 | 21⅓=16 | 34 | |||||

61 | 21+ | 20⅔=7 | 20⅓=13 | 22 | |||||

62 | 21+ | 16⅓-20⅔ | 14-16 | 13⅔=8 | 13⅓=14 | 32 | |||

63 | 38 | ||||||||

64 | 25+ | 21-24⅔ | 16-20⅔ | 13-15⅔ | 12⅔=7 | 12⅓=13 | 31 | ||

65 | 35 | ||||||||

66 | 1 |

- This number is assigned by team as a method of giving the better defensive teams a better chance of not allowing rebounds. If the goalie’s team is in the top quartile of fewest shots allowed per game, award a 20. If the goalie’s team is in the bottom quartile, award a 30. Otherwise award a 24.
- If the player has 58 or more games played, award a 41, 46-57 games played is a 42, 36-45 games played is a 43, 20-35 games played is a 44 and all others receive a 45. If the season has something different than 82 GP, adjust accordingly.

One key element of a goalie card is how the saves distribute the puck. Very few goalies will have the numbers 1–6 evenly distributed, so something to take note when placing the skaters in their positions.

39s and 40s, which could give a goalie a minor penalty, as far as I know have never been awarded. Even Ron Hextall and his 19 minor, 113 penalty minute 1988-89 season could not give him one.

So this should give you all the ammunition you need, if you dare, to make your own hockey cards. Good luck!

]]>Before we can go placing numbers on the card, we need to calculate a few things. All of the stats you will need were listed in the previous post.

- You will need to figure out the penalty numbers. How to do that is listed in this post.
- Next is the number of 9’s (power play-only shots) needed. This is calculated by the percentage of goals that are power play goals: 1-25%=one 9, 26-34%=two 9s, 35-49%=three 9s, 50-59%=four 9s, 60-74%=five 9s, 75-80%=six 9s, 81-89%=eight 9s, 90-99%=nine 9s, 100%=ten 9s.
- Determine the number of 30s desired to award penalty killers. These are based on reputation, a good method would be to sprinkle anywhere between 0 and 4 extra 30s among the team depending on the overall effectiveness of the team’s penalty kill unit. Forwards mostly get the extra 30s, defensemen getting them are rare.
- Determine the shots on goal taken per 60 minutes, otherwise known as S/60M.
- You will also need the number of games played and the Assist rating.

To assign the player numbers, go through each roll one row at a time, stopping once you have a number to assign:

- First, check the Pen# column. In the number of penalties needed is equal to or greater than the number needed, assign a penalty number. The lowest penalty number goes on dice roll 36, then 56, and so on in the order of Pen#.
- Second, check the %PP column. If the number of 9s needed is equal to or greater than the number needed, assign a 9 here.
- Then check the PK Rep column. If the number of 30s needed is equal to or greater than the number needed, assign a 30 here.
- Check the S/60M columns. If the S/60M value is greater than the number for the row, assign that play number. Example: If a player’s S/60M is 2.35, you would assign a 3 for dice roll 11, as it is greater than the number indicated in the 3 column but less than the 2 column.
- Check the Assist Rating columns. If a players Assist Rating is greater than the number for the row, assign that play number. Example: If a player’s Assist Rating was 42 and you were checking for Dice Roll 26, you would assign a 14, as it is greater than the number for the 14 column but less than the number than the 12 column.
- If a play number is still not assigned, use the play number in the Other column.
- Some results have a letter in parentheses [e.g. ‘(a)’]. This indicates a note at the bottom of chart may be applied here.

Dice # | Pen# | %PP | PK Rep | S/60M | Assist Rating | Other | |||||||||||||

9 | 30 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 18 | |||

11 | 3.35 | 2.75 | 2.22 | 1.64 | 1.35 | 0.84 | 7 | ||||||||||||

12 | 6 | 13.62 | 11.85 | 10.64 | 25 | ||||||||||||||

13 | 10.85 | 9.59 | 8.46 | 7.34 | 6.27 | 44 (a) | 22 | ||||||||||||

14 | 5 | 13.62 | 12.25 | 10.85 | 30 | ||||||||||||||

15 | 8 | 7.34 | 6.27 | 5.26 | 4.50 | 3.75 | 2.22 | 11 | |||||||||||

16 | 4 | 13.07 | 11.39 | 9.96 | 64 | 28 | |||||||||||||

21 | 11 | 11.39 | 9.96 | 9.00 | 62 (a) | 54 (a) | 43 (a) | 31 | |||||||||||

22 | 4.80 | 3.86 | 3.33 | 2.60 | 2.10 | 1.50 | 1.00 | 12 | |||||||||||

23 | 14 | 10.68 | 9.19 | 7.34 | 6.43 | 45 (a) | 22 | ||||||||||||

24 | 1 | 41 | 27 | ||||||||||||||||

25 | 9 | 7.71 | 6.61 | 6.00 | 5.27 | 4.50 | 2.90 | (d) | 15 | ||||||||||

26 | 13 | 4 | 61 | 46 | 41 | 32 | 26 | 24 | |||||||||||

31 | 9.59 | 8.47 | 7.51 | 6.61 | 5.70 | 4.80 | 3.64 | 32 | 16 | ||||||||||

32 | 10 | 2 | 12.71 | 11.06 | 26 | ||||||||||||||

33 | 3.50 | 2.95 | 2.39 | 1.84 | 1.00 | 0.01 | 8 | ||||||||||||

34 | 7 | 13.07 | 11.63 | 10.30 | 31 | ||||||||||||||

35 | 8.46 | 7.34 | 6.27 | 5.16 | 4.17 | 44 | 32 | 20 | |||||||||||

36 | 1 | 33 | |||||||||||||||||

41 | 5 | 10.55 | 9.58 | 8.90 | 7.92 | 52 | 44 | 35 | 24 | ||||||||||

42 | 9.15 | 8.09 | 6.91 | 5.90 | 4.78 | 43 | 34 | 21 | |||||||||||

43 | 8 | 13.07 | 11.63 | 63 | 62 | 29 | |||||||||||||

44 | 5.40 | 4.74 | 4.04 | 3.34 | 2.64 | 1.91 | 1.00 | 13 | |||||||||||

45 | 12.71 | 8.47 | 52 | 46 | 44 | 35 | 26 | ||||||||||||

46 | See Note (b) | 36 | 33 | 31 | 24 | ||||||||||||||

51 | 9.15 | 8.02 | 6.83 | 5.94 | 5.11 | 4.00 | 3.17 | 13 | 14 | ||||||||||

52 | 9 | 3 | 12.25 | 10.55 | 27 | ||||||||||||||

53 | 2 | 10.26 | 9.11 | 8.39 | 7.60 | 43 | 35 | 19 | |||||||||||

54 | 8 | 13.62 | 12.25 | 10.63 | 32 | ||||||||||||||

55 | 10 | 5.95 | 5.26 | 4.57 | 3.86 | 3.17 | 2.49 | 14 | |||||||||||

56 | 2 (c) | 9.68 | 8.16 | 7.29 | 33 | 26 | |||||||||||||

61 | 15 | 10.87 | 9.16 | 52 | 44 | 28 | |||||||||||||

62 | 3 | 8 (e) | |||||||||||||||||

63 | 12 | 6 | 10.87 | 9.16 | 45 | 42 | 35 | 25 | |||||||||||

64 | 9.16 | 8.09 | 8.01 | 5.90 | 4.80 | 44 | 23 | ||||||||||||

65 | 3 | 13.62 | 12.71 | 29 | |||||||||||||||

66 | 2.01 | 1.84 | 1.50 | 1.00 | 0.51 | 7 |

- Check the assists before the shots
- 1-40 Games Played=45, 41-53 GP=44, 54-65 GP=43, 66-75 GP=42, 76-81 GP=41 (adjust accordingly for seasons other than 82 games)
- If 2nd penalty number is 36 or greater, skip the penalty step, shift the 2nd number to 65 (and all subsequent ones as well one slot) and assign 2, 3, 4, 13 or 26
- Assign an 11 if the S/60M is between 1.00 and 2.89
- Assign an 29 if the S/60M=0

A few things I noticed while trying to compile the grid:

- Some of the lower and upper ranges may not be accurate as there are not a lot of examples of high and low shot ratings or high and low assist ratings.
- The number 17 is not assigned.
- There is some consideration to cards with many penalties getting their shot ratings downgraded. Since they were so rare I didn’t explore further. Also, no player may get more than 15 penalty numbers. If they deserve more than 15, more of the 36-39s are converted to 40s.
- There is no consideration that cards with higher shot frequencies have more non-shot rolls because of passes.

This should give you all of the information you need to make the skater card. In the final article of the series, we move onto the goalie card. Don’t fret, it’s easier. Trust me.

]]>Well, initially there was no pattern. Once I finally broke down and decided to figure out the whole card, a pattern then emerged. So I figured if I was going this deep, I might as well try to write an article on how to calculate a card. Or, in this case, three articles. The first article will go over the top and bottom of the card, i.e. all of the ratings. The second article will go over how to calculate the actual play numbers for the skaters, while the third article will go over the goalies.

For this, I had used a data file I had made a few years ago of the Western Conference players of the 2010-11 set. Using this data, I was able to figure out most of the patterns and calculations. I had to rely on a few other cards from the other seasons I had to figure out some of the outliers.

Please note that this will not be 100% accurate. There were some things I could just not figure out, or there were things that may have been mistakes on the card maker’s part. However, this will get you most of the way there.

The stats you will need:

- Games played
- Minutes played: Most stat sites will have this by ice time per game, so just simply multiply that by games played to get the minutes for the season
- Shots
- Goals
- Assists
- Power Play Goals
- Minor Penalties
- Major Penalties
- Misconduct + Game Misconduct Penalties
- Faceoffs Won
- Faceoffs Lost
- Blocks

And here we go:

**Position Ratings:** This is based on reputation, with some correlation to ice time. Higher ice time players tend to get the higher ratings. Somewhat based on the shots allowed per team, the higher ranked teams (those with the fewest) will have their players average around 3.5 while the lowest (those with the most) will average around 2.5. This is also somewhat tempered by the block ratings for the defensemen, which are calculated purely by stats. Teams with a high amount of blocks should have a slightly lower average than those with a low amount of blocks. For example, 2010-11 Detroit had a very low block rate, but had a higher than expected average defense to compensate.

**Forecheck:** There is no rhyme or reason to what players get, this is strictly on reputation. Only forwards get this rating. A handful will get a 5, about four players per team will get a 4, eight players will get a 3, about one per team will get a 1 and the remainder will get a 2.

**Shot Frequency:** Take the number of shots, multiply by 163.8, divide by the number of minutes played, and subtract one (unless it is already zero).

**Shot Rating:** Surprisingly, even though this is based off shot percentage, it is not entirely linear. Use the shot percentage (Goals / Shots) and assign accordingly:

Shot % | Rating | Shot % | Rating | Shot % | Rating | Shot % | Rating |
---|---|---|---|---|---|---|---|

0.1-3.1 | 11 | 9.6-11.0 | 21 | 18.1-19.5 | 31 | 26.6-28.0 | 41 |

3.2-4.1 | 12 | 11.1-12.5 | 22 | 19.6-21.0 | 32 | 28.1-29.6 | 42 |

4.2-5.5 | 13 | 12.6-14.0 | 23 | 21.1-22.0 | 33 | 29.7-31.0 | 43 |

5.6-7.0 | 14 | 14.1-15.5 | 24 | 22.1-23.5 | 34 | 31.1-32.5 | 44 |

7.1-8.5 | 15 | 15.6-16.5 | 25 | 23.6-25.0 | 35 | 32.6-34.0 | 45 |

8.6-9.5 | 16 | 16.6-18.0 | 26 | 25.1-26.5 | 36 | 34.1-35.5 | 46 |

**Minutes:** Ice Time per game rounded to the nearest minute.

**Faceoff:** Somewhat based on reputation and somewhat based on stats:

- All 5’s took at least 48 faceoffs per 60 minutes and had a success rate of 52.8% However, not everybody who crossed this line got a 5
- Most of the 4’s took at least 40 faceoffs per 60 minutes and had a success rate of 48%
- 3’s are either infrequent faceoff takers with good percentages or frequent faceoff takers with mediocre percentages
- 2’s are either infrequent faceoff takers with below average percentages or frequent takers under 45%
- 1’s are everybody else who took at least a few faceoffs a game.
- 0’s those who took less than a few faceoffs a game.

However, there are ratings that make no sense. Here are a few:

Player | Rating | FO % | FO/60 Min |
---|---|---|---|

Koivu (MON) | 5 | 52.9 | 56 |

Fiddler (PHO) | 4 | 53.9 | 67 |

Kopitar (LAK) | 4 | 49.9 | 43 |

H. Sedin (VAN) | 3 | 52.0 | 53 |

Backes (SLB) | 3 | 44.4 | 44 |

Backlund (CAL) | 2 | 48.0 | 45 |

**Min/Maj/Mas:** I detailed how to calculate the penalty ratings in this article.

**Clearing:** Reputation, some correlation to defensive abilities, no correlation to team shots allowed. Only defensemen get a clearing rating.

**Intimidation:** Reputation, no correlation to defensive abilities or penalty minutes.

**J:** Based off games played: 80+ for J-0, 71-79 for J-1, 60-70 for J-2, 42-59 for J-3, 1-41 for J-4.

**C: **This is mentioned in the Version 2 Rule Book, but if you don’t have it:

- Left of Slash (Offensive) is the higher of (Forecheck or Clearing) plus the A Rating (described later in article) plus Intimidation Rating
- Right of Slash (Defensive) is the highest Defense Rating plus the higher of (Forecheck or Clearing) plus Intimidation Rating

**I:** This is also mentioned in the Version 2 Rule Book, but if you don’t have it:

- Left of Slash (Offensive) is the S Rating (described later in article) plus Intimidation Rating
- Right of Slash (Defensive) is the highest Defense Rating plus the Intimidation Rating

**Blk:** This is derived from Blocks per game. 2.10 or higher gets you a 3, 1.26-2.09 gets you a 2, .76-1.25 gets you a 1, everything else is a 0. Only defensemen get this rating, even if a forward warrants one.

**A/S:** Although this appears in the Version 2 Rule Book, the cards I looked at appear to be a little different. Here is what I saw:

A/S | Assist Rating (for A) | Shot Rating (for S) |
---|---|---|

1 | 11-16 | 0-5 |

2 | 21-23 | 6-12 |

3 | 24-33 | 13-19 |

4 | 34-42 | 20-27 |

5 | 43-53 | 28-35 |

6 | 54-66 | 36+ |

This is a lot to chew off in one article, so we’ll go over how to make the actual card for the skater in Part 2.

]]>

**The Current System**

The play numbers 33 to 40 are used to dole out the trips to the sin bin, and essentially go in order from least severe (33) to most severe (40). Assuming that the average player spends 80% in Forecheck 2, 5% in Forecheck 1, 5% in Forecheck 3, 5% in Power Play and 5% in Shorthanded, each number should produce the following assuming you use none of the enhancements:

- A 33 is a coincidental penalty in Forecheck 2 72% of the time and a minor in Forecheck 3 and power play, so this will produce a coincidental penalty 58% of the time and a minor 10% of the time.
- A 34 is a minor penalty in Forecheck 2, a coincidental penalty in Forecheck 1 and shorthanded, and a coincidental penalty 72% of the time in Forecheck 3 and power play. This is a minor penalty 80% of the time and a coincidental penalty 17% of the time.
- A 35 is always a minor
- A 36 is a flight 72% of the time in Forecheck 2 and always a fight in Forecheck 3 and power play, producing a fight 78% of the time.
- A 37 is a fight in Forecheck 2, 3 and power play, or 90% of the time.
- A 38 is a fight in Forecheck 2, 3, power play and a fight 72% of the time in Forecheck 1 and shorthanded, or 97% of the time.
- A 39 is a fight
- A 40 is a major 42% of the time in Forecheck 3, 28% of the time in Forecheck 2 and 8% in Forecheck 1, otherwise it is a fight. This returns a fight 73% of the time and a major 27% of the time.

Players also receive a minor, major and misconduct rating. These are the typical ranges seen in all APBA Games (11-66 or 0).

Any fights or coincidental penalties involve checking the correlated penalty rating (minor or major) of all of the players on the other time and finding an opponent. For minors, after one check through if no opponent can be found it converts to a minor penalty. For fighting, after one check for a major opponent you then convert the penalty to coincidental minors. Major penalties do not have an opponent.

If you play via this method, it becomes quickly apparent that the players with the fight numbers receive too many penalty minutes. So in the second and third versions of the game, there are modifications offered to get the numbers better in line. This involves canceling a 33 for low-minor rated players or adding minors on some fight numbers. The third version streamlined some of the rules of the second version, which had results checking for penalties on non-penalty numbers.

**Figuring out the Ratings**

So what determines what penalty numbers players get? To do this, you need to find out the number of minors, majors and misconducts (both 10-minute and game). These numbers are available on the NHL site for recent seasons. You will also need to know the total ice time for the player, which is the game ice time times game. For this example, I’m go to explore the 2010-11 card of Justin Abdelakder of the Detroit Red Wings.

- Abdelkader played 12 minutes a game for 74 games, totaling to 888 minutes. Dividing 888 by 60 gives us 14.8. We’ll refer to this as “per 60 minutes” in later calculations.
- He had 61 minutes in penalties from 21 minors and 3 majors. Note that the math here (21 * 2) + (3 * 5) only produces 57 minutes. The NHL considers a double minor as one minor, when for these calculations should be counted as two minors. So we will use 23 minors and 3 majors.
- Take the number of minors and divide by the “per 60 minutes” figure: 23 / 14.8 = 1.55.
- To determine the Minor rating, take the figure above, multiply by 8.2, remove the decimal and subtract 2: (1.55 * 8.2) = 12.71, subtract 2 and round which leaves us 11. In the base-6 APBA-world we change that number to 25, as it is the 11th possible dice roll. A player can not be below 0 or above 66.
- To determine the play numbers, we associate the 1.55 figure with this chart:
- 0.00-0.51: 33
- 0.52-0.73: 33-33
- 0.74-1.10: 33-34-35
- 1.11-1.47: 34-34-35
- 1.48-1.75: 34-35-35
- 1.76-2.25: 35-35-35
- Add additional 35s for each .50 above 2.25

- So at the moment, Abdelkader has 34-35-35.

- Take the number of majors and divide by the “per 60 minutes” figure: 3 / 14.8 = .20
- To determine the Major rating, multiply the number above by 26 and convert to base 6: .2 * 26 = 5.2 or 15.
- To determine the play numbers, we use the Major rating and assign the following numbers:
- 0-14: nothing
- 15-23: 36
- 24-35: 36-36
- 36-42: 36-36-36
- 43-56: 36-36-37
- 61-64: 36-37-39
- 65-66: 36-37-37-37-37
- For numbers for players who would have received something beyond 66, 36-37-37-39-39 is used as a base and then additional 37 through 40 numbers are added. At this point there are not enough cases to figure out a set pattern, but basically it’s an extra number per additional .50. In the set I looked at, the largest number was 4.17 for Aaron Voros (Anaheim). He received 3 37s, 6 38s, a 39 and a 40.
- If the player received a 35 in the minor step, the first fight number
**replaces**the 35. All of the other fight numbers are then added. Since Abdelkader did receive a 35, the last one is replaced with a 36 (since he has a 15 Major rating).

- For the Misconduct rating, use the same method as the Major rating (misconducts divided by the “per 60 minutes” and multiply by 26, and then converting to base 6).

**Making it Better**

While in the current version the majors work pretty well, there are still way too many coincidental minors. Also, misconducts are improperly rendered, as the rating should be the number of misconducts into majors. Here’s what I do:

- If the play result is 33, reroll the dice and combine:
- 11-26: Coincidental minors (X)
- 31-36: Minor (X)
- 41-46: Check for misconduct, otherwise NZFO (X)
- 51-66: NZFO (X)

- If the play result is 34: Minor (X)
- If the play result is 35, reroll the dice and combine:
- 11-62: Minor (X)
- 63-64: Double Minor (X)
- 65-66: Double Minor plus coincidental minor (X)

- If the play result is 36, reroll the dice and combine:
- 11-45: Fight (X)
- 46-66: Coincidental minors (X)

- If the play result is 37, 38 or 39, reroll the dice and combine:
- 11-52: Fight (X)
- 53-66: Coincidental minors (X)

- If the play result is 40, reroll the dice and combine:
- 11-56: Fight (X)
- 61-66: Coincidental minors (X)

- Only check for misconducts on fights. Do not check on coincidental minors.
- On Forecheck 1, One Man Up or One Man Down, subtract one from the play number (e.g.: 34 becomes a 33).
- On Two Men Up or Two Men Down, subtract two from the play number (e.g.: 38 becomes a 36).
- If the designated fight opponent has a Major rating of 11-14, check for a possible conversion. Reroll the dice and combine:
- 11-21: Opponent Major 11 converts both to coincidental minors
- 22-32: Opponent Major 11 or 12 converts both to coincidental minors
- 33-43: Opponent Major 11, 12 or 13 converts both to coincidental minors
- 44-54: Opponent Major 11, 12, 13 or 14 converts both to coincidental minors
- 55-66: Keep as fight

**If I Really Had My Way**

Baseball cards sometimes have two columns. Football has three, bowling has two. Why can’t hockey have two? I would envision a chart that would be something like:

- 1-25: various minors (i.e.: 1-hooking, 2-tripping, 3-interference, 4-holding, etc.)
- 26-32: various coincidental minors or double minors
- 33-39: various fights
- 40: no penalty

The number of times a card is played through varies depending on the player’s Shot Frequency. For the purposes of this exercise, I’ll just pull a number out of my hat and say Abdelkader’s card would be hit 4.5 times a game assuming he plays his 12 minutes. Of his 23 minor penalties, 2 were double minors and 3 were paired as a coincidental, leaving 16 true minors. The other penalties he needs are the 2 double minors, half of the 3 coincidentals and half of the 3 majors, or 22 penalty events. He played 74 games and at 4.5 passes per game he would receive a dice roll 333 times, or 9.25 card passes. He has three penalty numbers, so those numbers would be hit 27.75 times.

We would create a second column that would be triggered off any of the penalty numbers. We would then see a card like this:

It would be fun to play. It would also be a nightmare to calculate.

Let’s first take a look to see who has it. In the 2014 set, discounting pitchers and anyone else with less than 100 PA, there are 17 players with a 0 on 66 and a 6 either on 11 or 33 or both. Let’s take a look at the “Special 17”:

- Two players have 6-6-0: Conor Gillaspie and Starlin Castro
- Two players have 6-0-0: Randal Grichuk and Chris Heisey
- The other 13 have 6-0: Corey Hart, Dee Gordon, Stefen Romero, Grant Green, Paul Konerko, Elian Herrera, Jason Kipnis, Nate Schierholtz, Allen Craig, Emilio Bonifacio, Gerald Laird, Joaquin Arias and Clint Barmes.

So what do they have in common?

- None of the players have enough HR to qualify for a first column 1.
- All of them have enough doubles to qualify for at least one first column 6.

And strangely that’s it. Some thoughts I had:

Hypothesis #1: These players all hit a homer with just a runner on 3rd, as a 6 is a HR in that situation: This does not hold water, as four of the players (Herrera, Laird, Arias and Barmes) did not hit a homer in 2014.

Hypothesis #2: These players were all out stretching a double into a triple, and a single column 6 is used as a way to have that happen in the computer game. 49 players in 2014 were out stretching a double into a triple. Only two of the “Special 17” were in that group (Castro and Green), so it’s not that.

Hypothesis #3: It has something to do with batting handicaps. Nope, all of the ratings except 2 are represented and one PL is included.

Hypothesis #4: It has something to do with position or bat-handedness. Nope, every position is represented, every batting hand is represented. Oddly, they all throw right, but that is likely a coincidence.

Hypothesis #5: They have no second-column singles. That actually works for a lot of the players, but it does not work for three of them (Laird, Arias and Barmes).

Hypothesis #6: Something weird happens if your only extra base hits are doubles. This actually has merit. Three of the group (yet again Laird, Arias and Barmes) did not hit any triples or homers. Additionally, all other players with at least 100 PA got a single column 6 or 6s except one, Jose Molina. Jose Molina didn’t really hit enough doubles to warrant even a single column 6, so he was given a 0 with 10 double column 6s). So we have figured out that if all of your extra base hits are doubles, you will get a single column 6 as long as you can warrant it. So we’ll remove Laird, Arias and Barmes from the equation, and move on to the “Frustrating 14”.

Hypothesis #7: Let’s revisit #5 since the three removed were the outliers in Hypothesis #5. So we have 14 players who have no double column singles for those players who have a double column. There are 8 other players who also would qualify in this situation, yet do not have the first column 6. So it’s something more than this?

Data Set Removal #1: One of the 8 players without the single column 6, Andrew Romine, does not qualify for a single column 6 since he only qualifies for one extra-base hit number and a quarter of that number should be a homer. So he is given a single 0 and we’re not going to worry about him in the later hypothesis, so we’ll call this other group the “Silly 7”.

Hypothesis #8: Is it something to do with the double rate? No, as the players seem to mix. The highest ranked doubles hitter in the combination of the two groups is in the Frustrating 14 (Gillaspie) but the second is in the Silly 7 (Jayson Werth).

Hypothesis #9: Is it something to do with the triple rate? Each group has a player with no triples, and the other players mix again, so it isn’t this.

Hypothesis #10: Is it something to do with the home run rate? There is a weird thing where the lowest number in the Silly 7 is 0.65, but there are players in the Frustrating 14 with a higher rate.

Since I can’t think of anything else that seems to work, I’m going to go with this to determine whether a player gets a 6-0, 6-6-0 or 6-0-0 combination:

- Player must have at least one double for each 34.25 PA.
- Player must have fewer than one homer for each 34.25 PA.
- If the player has hit no triples or homers, the player will have at least one first column 6 and then the requisite number of 0s.
- If the player will not have any single column singles, the player
have at least one first column 6, but they may not.**may**

Do you have a hypothesis of your own? Then don’t be a stranger, leave a comment…

]]>

So in my first game, Switzerland pushed Russia to a shootout. So I reviewed the rules:

- Get the shot range for the shooter
- Add 10 (a.k.a 6) to the shot range, and use either this number or 43, whichever is lower
- Roll the dice for the shot
- If the shot roll is within range, roll for the goalie. If you roll a 1 for the goalie, it is not a goal, otherwise, it is a goal
- If the shot roll is outside of range, it is not a goal

Rather anti-climatic, despite the Swiss winning the shootout 3-2. And frankly, I’m a bit worried about the realism. Recent stat geekery has shown that shot percentages tend to be very volatile, and not a great reflection of offensive ability. So, me being me, I had to run a few tests. First I had to see what the current out-of-the box rules produce. A while back I had entered in all of the cards from the 2010-11 season for the Western Conference into Excel, so I do have a base to test things. I gave each player that had a card the same number of shootout attempts he had in the real life season, with the goalie picked at random. This goalie picking was weighted by minutes played, so the more active goalies had a better chance of being picked. After running 100 trials, I got these numbers: Note that since this test was with only carded Western Conference players, the shooting percentage and the save percentage do not match up. The simulation also proved to be a little high, and you had to wonder if the goalie difference is right. Just two different versions of goalies (2 1’s or 3 1’s) don’t seem fine enough. So I had to see what differentiates the shootout in real life. You’ll notice at first glance that it doesn’t seem that there’s a great correlation between shooting percentage and shootout percentage. So first let’s see if there’s any correlation: Pretty much no correlation. But only one conference and one season is going to be a small sample size. Because the stats were (relatively) easy to get, I decided to go with active players with at least 25 career shootout attempts. The particular player’s shot percentage reflects only their stats in the shootout era (2005-present). It’s pretty obvious that using shooting percentage is not the way to go, since there seems to be little correlation. Another idea to try would be shots per minutes played. This is the current hot stat *du jour* for the hockey analyst crowd, and since we’re only looking for an offensive measure, we have it. First, the 2011 Western Conference: Oh my, a **negative** correlation. That isn’t good. Again, maybe just small sample size. Let’s check the careers: Ack, negative correlation again. This isn’t going to be of much help at all, and there really aren’t any other card tricks we can do to help out. So let’s set this aside for a minute and go to the goalies. We saw that the goalies in 2011 all had either two 1’s or three 1’s on their card. And in other seasons, only the fringiest of the fringe goalies will have less than two or more than four. So we pretty much have two classes of goalies: good and bad. First take a look at 2011: Hey, looky there, a bit of a trend line. And likely just enough to show a bit of variation between the bad goalies and the good goalies. Let’s just make sure this isn’t a one time thing:

Even more correlation when we go career, so even though it’s not a large correlation, it is something. In this group, the goalie straddling the 25% percentile was at a .911 save percentage and the 75% percentile was at .918, a .007 difference. The difference between high (.928) and low (.900) was .028. The goalie card comes into play about half the time, so you should be aiming for a 1/72 difference between two and three 1’s. That translates to .014. That’s a little large for the percentile difference, but too small for the high/low difference. So we’ll call it good enough — the 1’s rule stays.

However, there was one unintended quirk I found that can be used to facilitate shootouts: a home-road split. The NHL often publishes some of the team stats in home-road splits, and shootouts is one of them. Surprisingly, the **road** team has the advantage – .338 vs .319 in shot percentage and .524 vs. .476 in winning percentage – since the inception of the shootout. A .019 difference, it’s enough to give the road team a one point advantage when checking the shot range.

So, if you’ve made it this far, congratulations. After thinking it over, I decided I would need some sort of differentiation between the skaters, and it looks like the lesser of two evils is the shot percentage. One thing I did however was to lessen the ceiling used for the shot range, so the players who had a lucky season shooting won’t be overused. Here are the rules I now use for the shootout now, with changes **in bold**:

- Get the shot range for the shooter
- Add
**5**if the home team or**10 (aka 6)**if the road team to the shot range, and use either this number or**33**, whichever is lower (if this is a “neutral” site game, use 10/6) - Roll the dice for the shot
- If the shot roll is within range, roll for the goalie. If you roll a 1 for the goalie, it is not a goal, otherwise, it is a goal
- If the shot roll is outside of range, it is not a goal

The numbers by this method were much more realistic:

The road split was .335 vs .318.

So all of this work just to make a minor change. And it’s still going to be pretty anti-climatic. But at least I now know that there really isn’t a better way to do it without the cards having actual shootout stats.

]]>First off let’s take a look at the most important stat when comparing the fielding ratings: run differential. I had previously established a baseline of 4.04 runs scored per 36 PA for the average 2012 replay. For all of the different fielding ratings, here are the variances away from that 4.04:

Usage | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 |

P | .03 | -.01 | ||||||||

C | .07 | .05 | .00 | -.04 | -.04 | |||||

1B | .10 | .02 | -.01 | -.06 | ||||||

2B | .14 | .08 | .04 | -.02 | -.04 | |||||

3B | .08 | .06 | .03 | -.07 | -.06 | |||||

SS | .15 | .10 | .01 | -.06 | -.09 | |||||

OF | .05 | .03 | -.01 |

One of the things that sticks out is that the “average” rating (e.g.: 2B-7, OF-2) are slightly worse than average run-wise. This is compensated somewhat (with the exception of 3B) by having the average rating of the position over the course of a replay be a little above average. For the outfielders, the sum of the OF divided by 3 is presented, as I had no way to isolate the simulations to just use a certain rating in one of the three fields:

Usage | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | Avg |

P | .37 | .63 | 1.63 | ||||||||

C | .03 | .07 | .46 | .37 | .06 | 7.35 | |||||

1B | .15 | .35 | .28 | .22 | 3.57 | ||||||

2B | .00 | .11 | .42 | .33 | .14 | 7.50 | |||||

3B | .00 | .43 | .37 | .19 | .00 | 3.76 | |||||

SS | .05 | .12 | .59 | .20 | .03 | 8.05 | |||||

LF | .28 | .53 | .19 | 1.91 | |||||||

CF | .01 | .31 | .68 | 2.67 | |||||||

RF | .21 | .48 | .31 | 2.11 |

There are only two positions where the replay average is below the theoretical average: third base (by a lot) and leftfield (by a little). In total, the team defense averages 38.54: 31.86 in the infield and 6.69 in the outfield. With the shift over the last 20 years of 3B producing the second most errors on a team (rather than 2B), upping the 2B ratings and lowering the 3B ratings is one way to do it.

Looking at the main purpose of the fielding rating (errors), there is little surprise in the correlation between the errors per 36 PA for a rating and the errors per 36 PA for an overall replay.

EDiff | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | Avg |

P | .02 | -.02 | .06 | ||||||||

C | .11 | .08 | .01 | -.06 | -.07 | .10 | |||||

1B | .06 | .01 | -.02 | -.04 | .06 | ||||||

2B | .12 | .09 | .01 | -.02 | -.04 | .08 | |||||

3B | .02 | .02 | -.01 | -.05 | -.07 | .10 | |||||

SS | .08 | .05 | .00 | -.08 | -.10 | .15 | |||||

LF | .03 | -.01 | -.03 | .05 | |||||||

CF | .06 | .02 | -.01 | .03 | |||||||

RF | .04 | .00 | -.02 | .04 |

A little surprising is that the cliff between those positions where -2 away from the norm (e.g.: C-5 and C-7), the margin between -2 and -1 away is greater than +1 and +2 away. LF-2 is a negative, and CF appears to be the least important position to have an OF-3 in, while it is the most deadly to have an OF-1. A LF-3 can pay dividends, especially by neutralizing that pesky fielding 2 LF error with the bases empty.

A look at hits is not very surprising either, with not much difference between the bad fielders and the good ones:

HDiff | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 |

P | .01 | .00 | ||||||||

C | .03 | .03 | .02 | .02 | .02 | |||||

1B | .06 | .00 | -.01 | -.02 | ||||||

2B | .05 | .01 | .02 | -.01 | -.01 | |||||

3B | .03 | .03 | .03 | -.02 | .02 | |||||

SS | .04 | .03 | .01 | .00 | -.02 | |||||

OF | .03 | .01 | -.01 |

A bit of an anomaly at catcher, since the hits were up no matter what the rating. Makes me wonder if there is a little bit of an issue with the baseline I am using where the hits may have been a little low. I also have a feeling that the 3B-6 simulation was a little bit of an outlier on the offense side, as both the run and hit differentials were worse than 3B-5.

Let’s move on to the catcher’s throwing arm in respect to stopping (or enabling) the running game:

C | -4 | -3 | -2 | -1 | -0 | +1 | +2 | +3 | +4 | +5 | +6 | Avg |

Rdiff | .05 | .04 | .02 | .03 | .00 | .01 | .00 | .00 | -.01 | -.01 | .00 | |

SB% | .85 | .82 | .80 | .78 | .75 | .73 | .71 | .69 | .69 | .69 | .67 | .80 |

SBA/36 | .66 | .65 | .65 | .65 | .65 | .63 | .57 | .50 | .45 | .37 | .33 | .63 |

Usage | .18 | .31 | .11 | .12 | .11 | .05 | .00 | .02 | .04 | .03 | .03 | -1.21 |

This was probably the largest surprise of this simulation. First off, that the average Throw rating was -1.21 (and a median of -3), contributing somewhat to a higher than expected 80% steal success rate. The real-life rate for 2012 was only 74%. I have a feeling that APBA is still basing the ratings off a strict percentage and not curving them. The rate when the MG first came out was in the mid-60s, and a difference between what used to be an average of +1 in older sets going down to -1.21 roughly correlates to a 6% higher success rate.

Also noted that the success rate does go down as the throw rating gets better until about +3, at which it levels off. Taking over is a go/no-go call of sending the runner a lot less. I don’t have a Micromanager utility to check on Duke Robinson to be sure, but that would be my gut feeling. Duke also tended to steal more when then pitcher was higher rated, and laying off when lower rated.

And on the other end of the surprise, it turns out that the OF arm ratings actually have more influence than the defense ratings. Again, I could not split the run differences, but I could check on assists and the playing time of the individual ratings. To keep this from being too ridiculous, I only checked Arm ratings 20, 25, 30, 35 and 40. The usage figures below are based on 20-22, 23-27, 28-32, 33-37 and 38-40.

OF | 20 | 25 | 30 | 35 | 40 |

Rdiff | .08 | .07 | .03 | -.01 | -.05 |

LF A/36 | .07 | .05 | .06 | .07 | .08 |

CF A/36 | .06 | .06 | .06 | .05 | .06 |

RF A/36 | .07 | .07 | .09 | .10 | .12 |

Usage LF | .00 | .03 | .75 | .22 | .00 |

Usage CF | .00 | .01 | .38 | .59 | .01 |

Usage RF | .00 | .00 | .30 | .67 | .03 |

An arm of 40 has more benefits than an OF-3. And an arm of 20 has a higher penalty than an OF-1. Interestingly, in the 2012 set the arm ratings only range from 25-38. When it comes to assists (like errors, runs and hits), the centerfielder has the least influence of the three spots, with the rightfielder having the most. Just a little quirk of APBA to keep in mind when putting together your lineup.

When determining the whole value of a card by including the hit factors from the hitting articles, note that these are not on the same scale. The numbers in this article are based on every play on the field, whether or not the rating was used. The batting numbers only refer to the time that batter is hitting, so a factor must be either multiplied to the hitting numbers or divided from the fielding numbers depending on the expected batting position of the player (a batter batting 1st bats more often than a player batting 9th). If batting 1st-3rd, multiply/divide (whichever you chose) by .12, 4th-6th use .11 and 7th-9th use .10.

Another way to look at it: the difference between a SS-9 and a SS-7 (a dilemma often seen when trying to determine who to play) is roughly equivalent to an extra 1 for a SS-9:

- The difference between run differentials for a SS-7 (.10) and a SS-9 (-.06) is .16.
- Just assuming the player bats in the middle of the lineup, you would divide .16 / .11 to get 1.45, which is roughly equivalent to the value of a 1 (1.50).

The other positions on the difference between the standard Fielding 3 and 1 are a little wacky compared to the defensive spectrum we’re used to, basically, don’t worry about 2B and OF so hard and the corners are pretty important:

- Third Base (difference between 3B-5 and 3B-3): .13 (equivalent to two extra 7s)
- First Baseman (difference between 1B-4 and 1B-2): .11 (equivalent to an extra 3)
- Second Baseman (difference between 2B-8 and 2B-6): .10 (equivalent to an extra 6)
- Catcher (difference between C-8 and C-6): .09 (equivalent to an extra 8 and a 9)
- Outfielder (by arm, difference between 35 and 25): .08 (equivalent to an extra 9 and a 10)
- Outfielder (difference between OF-3 and OF-1): .06 (equivalent to an extra 7) (RF is likely most important, a 1 should stay out of CF)
- Catcher (by arm, difference between +3 and -3): .04 (equivalent to an extra 9)

The C and OF differences are cumulative. For example a C-8/Th+3 would be worth .13 better than a C-6/Th-3 (similar to the 3B-5/3 split) and an OF-3/35 would be .14 better than an OF-1/25 (equivalent to an extra 6 and 9).

Next up will be the other things that can be a factor: batting handicaps, speed ratings and steal ratings .

]]>When the Master Game came into play, the number of grades increased from 6 to 30, with each grade divided into 5 parts. There were no additional control or strikeout subratings, but additional subratings were added for homers, wild pitches, balks, hit by pitch and holding runners. The only ratings that really made a material difference where the homer ratings: 5 additional ratings and approximately 20 effective grades jumped the classes to 1,500. In a quirk of the computer vs. Master Game, the latter now has an additional control rating and 4 additional strikeout ratings.

A while back, I ran a simulation of the various hit numbers. I then did the same thing, this time concentrating on the pitching grades. Instead of drowning you in numbers, I’m going to go the pretty route and drown you with charts. First up I simply ran a 5-season simulation on every grade from 1 through 30 with all of the ‘neutral’ subratings: no control, no strikeout and no homerun. Here is a comparison of a few of the stats for each simulation:

Looking at that most important stat of runs, we notice that as the grade gets higher, the amount of difference lessens a little bit, and then craters, and actually goes up a little bit at the end. We see the same effects with hits and doubles, although hits stops its descent later, but has the same crater and uptick pattern. Home runs stay relatively constant until a slight descent in the 20s, strikeouts stay relatively constant until the 9 into a 13 kicks in at the A level (grades 18 and above). Double plays gently decline as there are fewer runners to double off in the higher grades.

So the basic formula I came up with to determine the R/36 PA variance away from the mean is listed below. I’ve included a translated version to ERA (if you’re more comfortable with that) *in italics*:

- Grade 1: 1.5 R/36 PA greater or
*2.12 ERA greater*. - Grade 2: 1.15 or
*1.63*. - Grades 3 through 7: .85 – .15(Grade – 3) or
*1.2 – .22(Grade – 3)* - Grades 8 through 12: -.15(Grade – 8) or –
*.19(Grade – 8)* - Grades 13 through 17: -.7 – .1(Grade -13) or
*-.87 – .13(Grade – 13)* - Grades 18 through 24: -1.3 – .09(Grade – 18) or
*-1.59 – .09(Grade – 18)* - Grades 25 through 27: -1.9 or
*-2.16* - Grade 28: -1.8 or
*-2.07* - Grade 29: -1.7 or
*-1.95* - Grade 30: -1.75 or
*-2.02*

So basically an 8 is average. And it looks like the long held belief that an A&C pitcher (grades 23-27) is a little more advantageous than an A&B (28-30), with the ability to knock out the 9 with a runner on 1st more important than knocking an 8 down with a runner on 3rd or runners on 2nd and 3rd.

In the early days of the subratings, it was noted that they were likely determined by BB/9 IP and SO/9 IP. At some recent point, APBA has changed that philosophy and started using plate appearance as the denominator. This was a smart move, as the BB/9 and SO/9 varies by the number of outs you create, but stays constant for the batters you face. As an example, I took the data from above, and plotted it on both an per 9 IP and per PA/36 basis:

Notice how the BB/36 PA stays relatively constant: it does go up a slight bit as the opportunities where the Z comes into play goes down a bit since there’s less runners on with better pitchers, while the BB/9 IP slowly declines. With strikeouts it’s a little more pronounced, but the same thing happens that around Grade 15 they pass each other. Since the per 36 PA seems to give a better indication of what is going on and is likely used by APBA, I will be using those when describing the subratings. Also, to keep this from becoming impossible to manage with dozens of tests for every grade, I only tested grades 3, 8, 13, 18, 23 and 28.

When dealing with home run ratings, it became obvious that the non-HR rated pitcher will give up a little less homers (.90 per 36 PA) than an average replay with the ratings distributed (.94). This difference is made up with the bad HR ratings (L and M) causing more home runs than the good HR ratings (G and H) take away.

Notice that the line for M is furthest away from the baseline, the H is far away but not the distance of M. G barely leaves the path while L does go a bit away, and the effects of the letters are more pronounced with the poorer grades than the better ones. A similar patterns occurs with the actual HR/36 PA. However, that patterns seems a bit odd, since the home run letters only affect the ‘1’ play result and other factors affect home runs from the ‘2’ through ‘6’ numbers. Makes you wonder if something else is at play.

Since it’s pretty difficult to give a set formula due to the vagaries of what grade the pitcher is for both the run and home run differences, I’ll just present a chart:

Per 36PA | Runs | Home Runs | ||||||

Grade | H | G | L | M | H | G | L | M |

3 | -.21 | -.05 | +.18 | +.43 | -.30 | -.05 | +.29 | +.72 |

8 | -.17 | -.09 | +.10 | +.33 | -.22 | -.09 | +.21 | +.60 |

13 | -.17 | -.08 | +.09 | +.35 | -.23 | -.13 | +.14 | +.55 |

18 | -.17 | -.11 | +.02 | +.35 | -.21 | -.13 | +.08 | +.35 |

23 | -.16 | -.03 | +.04 | +.17 | -.19 | -.06 | +.06 | +.27 |

28 | -.32 | -.18 | -.03 | +.02 | -.38 | -.22 | +.02 | +.09 |

With walks, the computer game does not have capacity for the ZZ, although it is unknown if the 5.75 update has added them (there are no 2012 pitchers rated ZZ to check). Similar to the home run rating, the bad rating (W) has more impact than the good rating (Z). However, the unrated pitcher will have a higher BB/36 PA (2.87) than the the baseline (2.80), since there are many more Z pitched innings than W. First the charts:

The effects are a little more linear, but there are impacts of the grade on the final answer. So instead of a formula, I’ll throw in another chart:

Per 36PA | Runs | Walks | ||

Grade | Z | W | Z | W |

3 | -.29 | +.59 | -.82 | +1.04 |

8 | -.15 | +.50 | -.69 | +1.23 |

13 | -.09 | +.45 | -.63 | +1.19 |

18 | -.06 | +.36 | -.55 | +1.16 |

23 | -.08 | +.35 | -.50 | +1.15 |

28 | -.09 | +.26 | -.53 | +1.01 |

One of the things we can do with this data is to find out what a ‘Z’ is worth. Although it is grade-dependent, it basically is a bonus bump on your grade, while a W is probably worth up to a penalty of 4 points in the mid range and 2 at the extremes. An H is worth a point or two depending on the initial grade. A G or L is either a half or whole point depending on the initial grade, while an M is a 2 or 3 point penalty. The strikeout letters proved to be of little change to the results.

On a final note, here are the percentages for each grade in the 2012 set and their subrating breakdowns. Notice that the bad ratings tend to be with the lower grades, and the better ones tend to be with the higher grades. Not going to find too many 19-Ms out there:

One thing to note here is that even though the average grade is an 8 as mentioned earlier, the midpoint of the grades is pretty much between 8 and 9, as the weight of the bad numbers (especially the W) drive the relative grade (assigned grade plus subrating effects) does bring it down a bit.

Next up, the fielders.

]]>