The value of Runs Created as a statistic.

[Scorekeeper]

I would like to see someone start a thread explaining and discussing value of rc stat. Would like to know how it should be used for offensive strategy and lineup purposes.

OK Westwind, here’s your thread. The easiest way I know to explain the value of RC is to show how it relates to other metrics. Here’s a lecture from a Sabermetrics class I took this past spring that explains it using simple graphs to compare it to BA, HR, OBP, SLGP, and OPS from the perspective of which stat best explains runs scored, which is the object of all offensive strategy. Of course if you believe there’s some other reason to employ an offensive strategy, you will have difficulty seeing any value in any offensive stat.

https://www.youtube.com/watch?v=F_nF_WbBlSk

Although the stat as discussed in the lecture is aimed at TEAM numbers, choosing players that give the best RC numbers individually will of course produce the best RC numbers for the team, and that’s where it’s value in an offensive stat lies. Its actual use would be the same as any other stat, and that would depend on your philosophy of seeing a lineup.

 

[Westwind]

Excellent. I went through some team stats and compared AVG, OPS and RC individually and made sample lineups from each stat individually. ( I don't believe in using single stats exclusively without considering all the stats in their totality but wanted to see how it affected the lineup.)

It was an interesting exercise. The #1 was same for all three and #2 remained pretty consistent. AVG and OPS had the same player in the #3 spot and RC dropped her to #6. She had half the RC of the #2 hitter.

The #9 in AVG and OPS clearly separated herself from the 9, 10,11,12 of the RC lineup.

For this team, it really delineated 4 subgroups of hitters and showed the most effective 9 of 12 which wasn't as apparent in the other stats.

This really highlighted the failings of AVG as commonly used a lineup tool.

My stats came from iScore which uses the technical version of RC. (Probably not the best version for softball! IMHO.)

Runs created - Wikipedia, the free encyclopedia

 

[BuckeyeGuy]

Seems to me that as RC is somewhat dependent on lineup position it'd be a feedback loop trying to use it to set a lineup?

 

[Westwind]

I wondered that too, Buckeyeguy. Isn't it corrected by dividing by AB plus BB etc?

I am not saying that's correct, I am trying to understand better.

 

[Greenmonsters]

Seems to me that as RC is somewhat dependent on lineup position it'd be a feedback loop trying to use it to set a lineup?

RC=((H+BB)*TB)/(AB+BB) Correct? I'm not sure what you base this on. The only way I see it being lineup position dependent as that a middle of the order hitter may not get as many good pitches to hit (which may or may not equate to more BBs) than a top or bottom of the order hitter based on their reputation (regardless of whether the reputation is earned or simply assumed). Best use of RC might be to determine the best 9 and then look at other stats and qualities to determine the batting order? i.e., I normally would rather have a fast slapper in front of rather than behind a plodding bomber

 

[Scorekeeper]

Seems to me that as RC is somewhat dependent on lineup position it'd be a feedback loop trying to use it to set a lineup?

Every stat that incorporates PAs or ABs is SOMEWHAT dependent on lineup position, but do you think compared to the big 4 offensive stats(BA, OBP, SlgP, OPS) RC is more or less lineup dependent?

 

[Scorekeeper]

RC=((H+BB)*TB)/(AB+BB) Correct? I'm not sure what you base this on. The only way I see it being lineup position dependent as that a middle of the order hitter may not get as many good pitches to hit (which may or may not equate to more BBs) than a top or bottom of the order hitter based on their reputation (regardless of whether the reputation is earned or simply assumed). Best use of RC might be to determine the best 9 and then look at other stats and qualities to determine the batting order? i.e., I normally would rather have a fast slapper in front of rather than behind a plodding bomber

That’s the most basic form of RC, and likely the best form to use for SB, or any venue where there are no leadoffs. I use the more updated version RC=((H+BB-CS)*(TB+(.55*SB))/(AB+BB) because stolen bases are much more prevalent in baseball, and the team I scored for was super aggressive on the bases. To ignore SBs would be ignoring a very valuable factor in the thinking about a lineup.

I’ve played around with lineups all kinds of different ways over the years and have come to some conclusions of my own. The 1st is, no one stat should ever be used to determine a lineup. By that I mean, let’s say you want to BA to set the lineup. Simply taking the players with the top 9 Bas and sticking them in the same lineup positions as their BA rank would be extremely ignorant if not stupid for too many reasons to list here, and the same can be said for any combination of stats.

Instead, if you want to use stats to set the lineup, you’ve 1st got to determine the importance of each batting position, then find the player best suited to that importance. Then, continue doing that until each position has been filled. So let’s say you feel the most important position in the lineup is the #3 position, and you want to fill it with your best hitter. 1st you’ve to define what determines the best hitter, and you’ve got to define it in a way that considers every player. IOW, if you believe the player with the best BA is the best hitter, is a player whose BA is .500 but only has 2 PAs “better” than a hitter that’s 40-100?

That takes a lot of care because it just might be that the player with the fewer PAs really is the better hitter. So how about this? Let the numbers determine the top 3 qualified players in the pool, show them to the person creating the lineup, then let them choose the player to put into the position. Once that position is filled, the player filling it is removed from the pool, then the next most important position can be filled.

Assume the #1 position is determined to be the next most important position. Now you’ve got to define what makes a good leadoff hitter. FI, you want him to get on base a lot, be a very patient hitter, and have great foot speed. The getting on base part is pretty simple. Look at OBP. determining foot speed is bit more difficult. To me the easiest way to do that would be to get a base foot speed, like time in the 40yd and 90’ sprint time, but that’s only raw speed. To factor in skill to some degree, I’d also factor in SBPct. Patience can be as simple or complex as you’d want, but for the sake of example let’s say you use pitches per PA as the criteria.

Getting all those different things is the easy part though. Now you’ve got to put them together to get some number you can rank the players by. Just to make it easy, let’s rank every player in each of those different criteria, then add all their different ranks together to get 1 number. FI, if player “A” was the 3rd best in OPB, the 4th fastest in the 40, 5th fastest in the 90, and had highest PPA of 18 players, her score would be 3+4+5+1 or 13. Player “B” might be the 2nd best in OPB, 9th fastest in the 40 and 11th in the 90, and was the least patient of all the player for a score of 2+9+11+18 or 40. Do that for all of the players, then present the top to the person creating the lineup. Then continue to the next most important position using only the players who weren’t already chosen.

I know that sounds like a boatload of trouble, and it is. But, it only has to be done once! Once the criteria have all been determined, the computer would be able to spit out the 3 best names for each position faster than you can choose one of the names. And if you think about it, isn’t that basically how a coach makes lineups without a computer? The only difference is, rather than using “gut” feelings, real numbers would be used.

 

[BuckeyeGuy]

RC=((H+BB)*TB)/(AB+BB) Correct? I'm not sure what you base this on. The only way I see it being lineup position dependent as that a middle of the order hitter may not get as many good pitches to hit (which may or may not equate to more BBs) than a top or bottom of the order hitter based on their reputation (regardless of whether the reputation is earned or simply assumed). Best use of RC might be to determine the best 9 and then look at other stats and qualities to determine the batting order?

Runs created - BR Bullpen

"James came up with an all-new version of the formula in the 1990's, which takes into account position in the lineup and other factors. He uses this formula to develop Win Shares."

Lineup position is huge, and for exactly the reasons you stated. Your last sentence is exactly what I was thinking. I use a modified OPS calc and then use the

"#1, #4, #2, #5, #3, #6, #7, #8, #9
So, you want your best three hitters to hit in the #1, #4, and #2 spots. Distribute them so OBP is higher in the order and SLG is lower." Optimizing Your Lineup By The Book - Beyond the Box Score

approach.