Total Batters Faced - % of PAs going to L and R hitters

Hello!

I've been working on valuation metrics for the entire player pool. So overall NERP for a player. Defense, clutch, speed etc.

For pitchers it's been a tad straight-forward with just NERP, END, BP #s, Hold, WP/BK, Defense, 4-man/5man status, Ks and a little bit of an imbalance tax. All these have a NERP score that pools into a pitchers' overall NERP. Diving back into this, the imbalance "tax" was something I wanted to refind as guys who dominate R like Scherzer, Millwood, etc or good ~4R or more extreme were terrible "values" according to the trendline of the whole player pool, whereas guys who kill L like G. Peters were seen as too good of values. But this starting assumption was far more problematic. Will change how or if I use an imbalance tax:

So I was using 50%/50% split of vs L NERP and Vs R NERP ---- basically assuming league breakdowns of PAs from Lefty hitters and Righty hitters was equal (for hitters I'm using 72% against R pitchers.) It shouldn't be 50/50?

I just crunched the numbers from 12 different $80mil DH ATG8 leagues. Comes from Team's L/R splits pages. Don't think the 365 game has anything like Report Writer on the PC game...

PA from L hitters = 0.443903715% (so 44.5%?)
PA from R hitters = 0.556096285%


**Strat only provides ABs for hitter vs L/R splits, so imperfect**
PA vs L pitchers = 0.357916107% (so 35.8% ? Far higher than my 28% but not I'm really picking at this)
PA vs R pitchers = 0.642083893%

--------------------------------

Does anyone have numbers or estimates for % of PAs going to L and R hitters? I'm thinking 44.5% vs L hitters....

Honestly, it is league specific. If you get a league where half the teams go hard with lefty stadiums, I've seen over 50% of the leagues pitchers be LH.

I think the numbers you proposed:

44.5% vs. L for pitchers
55.5% vs. R for pitchers

And 35% vs. L for hitters, with 65% vs. R.

When I do my rough estimates, I tend to go 50/50 for pitchers and 35/65 for hitters. My hitter analysis tend to be more accurate than my pitcher ones.

It's also different for right handed pitchers and left handed pitchers.

For RHP I assume 55%-60% against left handed batters.
For LHP I assume 25%-30% against left handed batters.

It's also different for right handed pitchers and left handed pitchers.

For RHP I assume 55%-60% against left handed batters.
For LHP I assume 25%-30% against left handed batters.

Exactly and that's why I think there is a [proper] premium for reverse LHP compared to reverse RHP--it tends to be easier to get them into favorable environments

Thanks for the feedback!

Hamilton17, 50/50 split for PAs vs L or R hitters doesn't seem appropriate. Lefty-killers have a huge advantage that is now clearer to me (For two extreme examples: 7L M. Hampton $6.8 vs 4R K. Milllwood $8.08 --- assuming a 50%/50% split in NERP values between vsL and vsR and all other NERP on the card Hampton is 25.3% underpriced and Millwood is 19.7% overpriced

This pattern largely continues SP. At 44% vs L and 56% vs R, Hampton's iregular and previously formula-breaking card goes to 16.1% underpriced vs Millwood's 15.4% overpriced. Better, but still indicates a 44% rate vs L is too high. (?) *Many R-dominant SP are "good" values and L-dominant SP are "bad" values too, so it's not woefully broken.

Barrmorris, I havn't wrapped my mind around different %s vs L/R for Lefty and Righty pitchers. My comment above about 44% seems to indicate it should be a shifting (and less advantageous for the 5Ls of the world) % in the algorithm. I'd like to make this relatively simple or conservative and just speak to SOM's pricing formula (not so much to success of platoons, although that is/should be factored into pricing a bit.... for what it's worth, my league is a career league without as much access to idealized platoons like M. Duncan, G. Brown or even lesser guys, but I still wanna price like ATG9)

Considering the pool of 930 SP*, I have 28 SP* with a z-Score of +/- 2.5, meaning they are way outliers, likely to be excluded in the pool when getting pricing recomendations (as they are problematic for the data/trendline). (8 pitchers as too good of values, 20 as too bad) I don't think that's an extreme amount considering my naiviety about this vs L/R split for Pitchers.

A). Should the algorithm have different %s vs L/R for Lefty and Righty Pitchers?
B.1). If so, what should the %s be?
B.2). Can it be the same %s for all price points? (pleeeease
C). Would RP use the same %s?
D). Thoughts on "one" algorithm for this pricing? This has been 2+ years in the making for hitters and pitchers and pitchers is on shakier ground as I try to lay out here. Still, many "curves," charts and cohort analysis (SP* vs SP, # prone pitchers, imbalanced) Imbalanced SP have been the most problematic (ie they account for a signficant amount of the scores below the efficiency curve. ----> again, this implies that a 50%/50% split or 44%/56% split isn't enough?

--------------------------------

Excel's trend function is leaving a lot to be desired from in accounting for diminishing returns in terms of incrememntal NERP gains for each additional $mil spent Ex: rates cheaper SP as "great" bargains a bit too much. Want to includes sub-$2mil pitchers in the data, but they're too warping. Similar issue with hitters, but more tolerable. NERP is a bit trickier for pitchers as they go negative and positive.....so I go -21 to all pitcher NERP so all pitchers are in the same range (not having to value +/- 0) **I still get "expected" prices for sub-$2mil players, but it's calculated from a player pool that doesn't include sub-$2mil players. Whatever is on the lower end of the "pool" will always be considered a bit too "good" of bargains. So the human element in pricing consides this and may price a cheaper player 10-15% below "expected" E-NERP values, as this is increasingly normal for cheaper cards.

It's also different for right handed pitchers and left handed pitchers.

For RHP I assume 55%-60% against left handed batters.
For LHP I assume 25%-30% against left handed batters.

Exactly and that's why I think there is a [proper] premium for reverse LHP compared to reverse RHP--it tends to be easier to get them into favorable environments

To this point, 2003 Steve Traschel is a Righty a 9L. My expected price formula has him as the most overpriced at 47.4% overpriced. In fact, 18 of those 20 "outlier" overpriced pitchers are reverse guys. 17 of those 18 are extreme.

$3.96, Trachsel, Steve (R, 9L) = 47.4% (-2.82 NERP vs L, 20.24 NERP vs R averages to a bad 10.09 NERP at 44%/56% splits)
$2.94, Scott, Mike (R, 6L) = 40.5%
$3.72, Buehrle, Mark (L, 7R) = 31.1%
$6.54, Santana, Johan (L, 6R) = 30.5%
$3.44, Wood, Wilbur (L, 5R) = 30.2%
$4.01, Morris, Jack (R, 5L) = 28.8%
$5.33, Zito, Barry (L, 6R) = 28.8%
$2.71, Hurst, Bruce (L, 6R) = 28.7%
$7.06, Darwin, Danny (R, 4R) = 27.5%
$2.98, Cuellar, Mike (L, 3R) = 25.6%
$3.56, Candelaria, John (L, 1R) = 25.1%
$2.69, Garza, Matt (R, 7L) = 25.2%
$6.42, Antonelli, Johnny (L, 5R) = 25.1%
$4.12, Tapani, Kevin (R, 4L) = 23.8%
$5.24, Niekro, Phil (R, 5L) = 23.6%
$5.28, Abbott, Jim (L, 7R) = 23.4%
$3.17, Whitson, Ed (R, 2R) = 22.7%
$5.57, Glavine, Tom (L, 8R) = 22.7%
$4.44, Hurst, Bruce (L, 7R) = 22.6%
$7.22, Mussina, Mike (R, 5L) = 22.3%

~12 of the next 20 "worst" values are also inverted uneven (2L/R or more BAL with reverse handedness)

Different %s for L and R seems appropriate ??? Granted, alot of these pitchers are the kinds I don't like. Cards can obviously just be bad bargains too but a pattern is clear.

15 of the 20 "worst" values for non* SP are reverse imbalances too. Immortal 6L Righty Klipstein is 26th "worst" value among non* SP. Childsmwc's formulas must account for this reverse imbalance thing! Is it a consistent % at all price points?

Great stuff nels. I agree in general, though I think extremes are more likely to have an overpricing bias compared to non- because of the assumptions that have to be made to normalize everything (for instance, you have to assume a L/R % environment to price the card).

But the rub is that's what makes these guys awesome. Put them where you mitigate or avoid their weakness, and you may have a monster. Not for the feint of heart of course...but if a guy like Trachsel faces 1,2,3,4--how many righties before he is NOT worth it. I say anything more than 3, but I'd pitch him against 3 or fewer, especially if the park favors him.

here's an example--I've used Mike Scott's 6L card--even in the 2021 Barnstormers finals league.
https://365.strat-o-matic.com/team/1672405

He was complimented by a hard righty in the rotation so I just had him avoid teams where it didn't make sense. Though that hard righty had a few very rough outings so who knows since Scott was 11-4 and Hensley was 3-13--but the $ investment was minor which was the point--to get to almost .500 with a few mil spent on two roster slots, which freed cash for other things. Like Speaker And Hensley pitched a little relief and also did well where it favored him but no so great in decisions.
NOTE: I wanted a reverse lefty as a compliment--back to the original point--but they were too expensive for what I needed. A reverse lefty would have done very well--better than Hensley for sure. So I think the pricing is spot-on qualitatively. It will never be quantitatively perfect because there is no perfect single price at all caps for all leagues.

And that was a 2.94M guy in a 100M league with the best managers in ATG that year...so it was not a fluke. Not one of those managers lets weaknesses go unexploited.

So guys like Scott et al, can be a value if you use their strength and mitigate the weakness. I think that is true for many of these guys (though not all--some of them are not reverse types I would play).

So for pricing, it is almost a crap shoot because an average/normal distribution (like average L/R %) is never going to accurately represent the value curve of a card that is in the tail of the model distribution. I mean--that's by definition--that type card is not an average card.

Great data nels!

For RHP I assume 55%-60% against left handed batters.
For LHP I assume 25%-30% against left handed batters.

The idea of it being different for LHP and RHP is what I was most unready for. Interesting, but likely true. Just a big can of worms for the algorithm. I'm going to try on the conservative end of this first. So...

LHP = 70% vs R (?)
RHP = 45% vs R (?)

I really wish there was some data about this on a team/league level. Any further insight or thoughts is soooo appreciated. Seems silly to not include the total for each handedness on a Team's L/R Splits. Not to mention League Stats about it. I messaged childsmwc (who helped create pricing model) who has been a huge help before. Clearly a different % is baked into the pricing----as evidenced by the reverse BAL guys I detailed above.

FrankieT, definitely settting guys up to have surplus values by minimizing their "weaknesses" is what makes cards really valuable. And the game fun. Not in denial of that, just need to make one unifying formula to price ALL cards.

If anyone wants graphs showing the pricing curve with different cohorts within it (speed/arms/BAL/positions/etc), send me a PM and I can email whatever. I'd appreciate the conversation/collaboration as I do here. To bad you can't link images here.

LHP = 70% vs R (?)
RHP = 45% vs R (?)

This is probably not far off, maybe even more extreme for lefties.

My experience is that leagues can vary significantly but a lot of managers try to "game it" (quite reasonably so) based on what they see on others rosters. Because of the sheer size the player set there will never be a "natural" distribution that corresponds to real life or is even effectively limited (all 12 teams could have all lefty pitchers and there would still be hundreds left over to pick from). Once managers percieive value in a particular strategy, the distribution is immediately upset-for example in the price efficiency of platoons. Before this player set. platoons were a rarity.

Right handed pitchers seem to face a lot of lefties, which is why I often pick Righties with L ratings. The reverse is true for lefties to an even larger extreme (hence the value of R rated lefties). In a reverse manner, it seems at times that right handed hitters that kill righty pitching can be effective (same is true for lefty swinging lefty killers, but also to a lesser degree).

Undoubtedly leagues vary. No league will fit a "representative" average. But pricing does, or has to

The application of my pricing formula is relevant. Goffchile, you get me thinking about the perhaps conflicting nature of it. I've been trying to closely mirror if not duplicate ATG9s pricing formula, but the player pool I'm using it for, along with access to said player pool is very different.

1. My pricing formula is for a ~90s-today career league. So a certain card is only available for one particular "season" the teams are living through. Way smaller available talent pool. Randomized career starts. So the talent pool will be nowhere as deep as in ATG----but will be priced similarly. The effectively infinite supply of most skills in ATG is extremely relevant to a lot of what we've been talking about. Gotta wonder how/if that's factored into pricing.... IE somebody shouldn't be nearly as willing to overpay for Robin Young or Ernie Banks as they were in ATG2.

2. This is fine. I wanna normalize over/under priced cards a bit as teams will "bid" extra amounts on guys for contracts, on top of their existing salary. So ~$60mil in "talent" in a $80mil league as alot of the budget is those "bids".

The extreme manipulation of the player pool to make ANY kind of team vs the limited supply of my league makes me want to more conservative in these platoon things. Certainly these imbalance opportunities increase as does the size of the player pool.