The fundamental problem with the PtP WP/PB system was that it based the pitcher ratings purely on WP and the catcher ratings purely on PB. Furthermore, there was no adjustment made for the pitcher/catcher combination.  From a game mechanics standpoint there is no fundamental difference [...]]]> This is an explanation of how I re-worked the wild-pitch/passed-ball system.

The fundamental problem with the PtP WP/PB system was that it based the pitcher ratings purely on WP and the catcher ratings purely on PB. Furthermore, there was no adjustment made for the pitcher/catcher combination.  From a game mechanics standpoint there is no fundamental difference between a WP and PB, it is really just a scoring decision. Therefore I decided to treat WP and PB as the same event and design a system which would attempt to get a more accurate distribution of the combined event.

The first step is to calculate the overall WP+PB rate for each MLB team.  The next step is to calculate an adjusted average rate for the entire league, “adjusted” meaning that we account for the fact that the WP/PB range on the pitcher card is actually too small to account for the full range of variance.  There are pitchers who throw WP so often that even if all WP/PB events were WP it would not be enough (these are the guys who get WILD-WP!), so the “adjusted” average just means it accounts for the fact that not all WP come from the WP/PB chart.

Once we have the team rates and the adjusted average, the next step is to assign responsibility for 50% of the variance from average to the catchers and 50% to pitchers, and then assign each team’s pool of catchers and pitchers a separate overall rating.  After that is done individual ratings for pitchers and catchers are assigned by referencing their WP+PB rate against the overall rating for the team’s opposite end of the battery.

Instead of a simple hard capping system (e.g. less than X [...]]]> Defensive ratings are subject to adjustment based upon playing time.  Generally these adjustments are simple hard caps based on innings played, but the system for range ratings is significantly more complicated.  Since I frequently get questions about the range adjustments I’ve posted a summary here.

Instead of a simple hard capping system (e.g. less than X innings no higher than Y range), low playing time (PT) players are mapped into a smaller scale than the normal 0-10 (K thru A range).  This means that lower PT players have to exceed the performance of their peers with more PT in order to achieve the same range rating.  It also allows the lower PT guys who are “off the scale” (raw range ratings better than “A”) to exceed the upper end of the reduced scale and achieve higher ratings than what would be allowed under a hard capping system.

The playing time value we use is a calculation I’ll refer to as “adjusted innings”, which breaks down to:

(games played at position) + (innings played at position)
+ (bonus for innings logged at similar positions)

pos     similar
---     --------
1b      2b,3b,ss
2b      3b,ss
3b      2b,ss
ss      2b,3b
lf      cf,rf
cf      lf,rf
rf      lf,cf

The bonus cannot be greater than the number of games + innings at the primary position.

Upper bound for the scale is determined by a step function, starting at 6.0 for 216 “adjusted innings” and increasing 0.125 for every 27 additional “adjusted innings”.  Here’s an abbreviated chart with some of the key values:

         upper
  PT     bound
----    ------
 216     6.000
 324     6.500
 432     7.000
 540     7.500
 648     8.000
 756     8.500
 864     9.000
 972     9.500
1080    10.000

There is one additional hard cap for players under 216, set at the fractional value below 216 (e.g. a guy who was 80% to 216 would be hard capped at an 8).

Finally, there is a bonus for playing good defense at similar positions.  For any player who has their range rating reduced by PT adjustments, if they have a higher raw range rating in more innings at similar position(s) then his adjusted range rating is increased by one.  For CF/SS their defense at the other position(s) has to be 3 grades higher.

The only way the legacy PtP system can generate accurate results is if the “user” [...]]]> In 2005 I implemented a regression algorithm for the LH/RH splits on the cards, replacing the legacy PtP method of basing each side of the card on the raw split data.  This is a brief explanation of the rationale and methodology.

The only way the legacy PtP system can generate accurate results is if the “user” ensures that each card receives the proper PA distribution of LH/RH.  What years of experience have shown is that our “users” do not do this, nor do they want to do this.  The game isn’t as fun when player usage is rigidly enforced.  Therefore, the PA distributions are not allocated properly and the overall player performance is skewed.  If we strike a balance between replicating the platoon split and replicating the overall performance of the player we will improve the overall accuracy of the system.

Now, one one end of the spectrum we have games which ignore platoon splits or enforce a standard split for all players, and on the other end we have the PtP-inspired system which treats vLH/vRH data separately. I decided that a hybrid system, one which regresses platoon splits toward the norm, retains the original intent of the PtP system which we all enjoy while making the game more accurate in reproducing overall performance.  This rationale is not based on ability, but rather on a tacit acknowledgement of how our “users” interact with the product.

Implementation

Once the raw target totals have been calculated for a player card, they are passed into the regression algorithm. The regression algorithm takes the existing LH/RH ratio, and regresses it on a sliding scale toward the standard ratio (LL/LR for LH, RL/RR for RH).

Regression amount is a simple linear function starting at 75% for 0 PA and scaling down to 0% at 900 PA.  This divides neatly, with 300 PA being the 50% point and 600 PA being the 25% point.  For each event, we determine which side generated more offense and use that PA sample as the value to plug into the regression function.  We do this because the offense has final decision on the pitcher/batter machup.

When adjusting the events to match the new target ratio you have to use an “exchange rate” based on the PA split.  So, if you’re moving events from the LH side to the RH side you have to do so at the ratio of RH/LH PA.