Looks like there were some tough lefties in this league. Progressive league, relatively high cap. Ted Simmons is a switch hitter, so you'd expect things to be fairly close. They definitely were not.

VS LHP	76	140	----	38	5	0	1	----	8	8	----	----	.271	.311	.329	0.640
VS RHP	149	488	----	153	26	2	8	----	37	40	----	----	.314	.362	.424	0.786

The vs. left is such a small sample size anyway, those probably qualify as fairly close.  I bet if you computed the size of the error bars on that one 140 PA sample they'd reach almost to the numbers he had vs. righties if not beyond and certainly to within the error bars on the vs. right numbers.

I like to look at the pitchers OAV vs. righties and lefties every game.  Then I tweak my lineup. Of course the catchers arm and the ballpark come into play as well. Does anyone else do this?

Its my understanding (from posting the question and getting answers) that the coding of platoon splits here L/R is identical for all players. If so, much of what you see when you look at pitchers OAV vs righties and lefties is random variation - and therefore not a predictor of how the pitcher will perform from here on out. (There would be other factors such as the quality of L hitters and R hitters in the league, parks in use, etc. which could have some effect). But the point is, if the feedback I've gotten from experienced owners is right, tweaking your game to game lineup based upon the differences in L/R splits for the opposing SP is largely tweaking to a mirage. But sometimes I do it anyway, just like you.

Now, I also have seen variation in pitcher split numbers which I have a hard time reconciling to simple random variation. However, with such small sample sizes of a few hundred AB, it is easy to see what you want to see in the stats when in fact the effects are indeed random.

If there is anyone here who believes that platoon splits here are not coded uniformly for all players, I'd be quite interested.

thanks for the feedback jayjelinkek.  I do spend time on fixing my lineup and rotation every game and that is why I's rather have one or two teams at a time. I enjoy it.  It may only make a slight difference in wins.  I may make my first playoff appearance soon. Ive learned drafting really is important along with solid pitching. Almost like in the majors. thanks again!

@ jayjelinek - it is known with absolute certainty that platoon splits are coded uniformly for all players.  It's a 5% advantage to the hitter (or disadvantage to the pitcher, I forget which, but the result is similar) if you end up with an opposite-side matchup, with a 5% hitter disadvantage (or pitcher advantage) for same hand matchups.  For switch hitters no advantage is given.  This has been explicitly stated in multiple ticket responses to multiple owners.  We've also seen it in some presentation slides one of the developers presented at a meeting a few years ago.  This method is necessary because platoon splits aren't necessarily available for some of the older players in the database.  As such...

@chisox278 - honestly, you're right.  Massaging your lineups based on performance will make only a slight difference in wins unless you do anything incredibly stupid and/or kneejerk.  What you probably don't realize is the difference is negative - that is, if you've put together a good lineup to begin with, changing it around is going to cost you a few runs and potentially a few wins.  Performance with respect to expectations is entirely random if the expectations are correct.  If your guy is overperforming it doesn't make him any more or less likely to be successful in his next game than a guy who is underperforming.  The only time you should change your lineup is if you look deeper into the stats and discover a reason why a guy isn't doing as well or is doing better than you had anticipated.

At the risk of hijacking this thread (we can start another one if this takes off), the following statement by dahsdebater got me thinking. (BTW, dahsdebater, thanks much for your reply and insights!)

This method (use of generic platoon splits for all) is necessary because platoon splits aren't necessarily available for some of the older players in the database.

Actually, the lack of platoon split information for older players is not, IMO, a real basis for excluding platoon split information when we have it. CS numbers aren't available for all older players either, but I doubt an argument for thereby excluding CS numbers from consideration for all players in the database would take traction here.

IMO, there is a separate, and good, argument supporting the use of generic platoon splits. The internet mathheads that produce the hitter/pitcher projections at the beginning of each baseball season (i.e. CHONE, ZiPS, etc) have found that the players previous season platoon splits are a very poor predictor of the players platoon splits the next year. Use of generic splits is a better predictor in this case. And this means that any extremes in players platoon splits over a single season are largely the product of random variation rather than a true measure of the players intrinsic ability. So I cannot find fault with the decision to code the players with a uniform split %.

However, I would also not find fault with a WIS decision to code the players to reproduce their actual platoon splits for each year. This would produce extremes, and these extremes would be exploitable by experienced owners, and possibly this consequence of making the gulf between rookie owners and vet owners that much wider would be undesirable.

There is also an underlying philosophical question here which is related, and subtle, but important...

When WIS codes a specific player season, is WIS trying to code that players actual performance for that year, or code to some estimate of that players actual ability for that year? This is a REAL distinction. For example, in 1961, Roger Maris hit 61 home runs, due to a combination of good fortune, good circumstances (batting in front of Mantle), and good ability. But I'll venture that very few would argue that 61 Maris really had more actual home run hitting ability than his teammate 61 Mantle. I'd sure argue Mantle was the greater HR hitting threat, and the fact that Maris was never intentionally walked during the whole 1961 season leads me to believe most MLB managers would agree. But Maris did hit more home runs in his season, due to better circumstances, and better luck.

So...do an experiment. Run 1000 WIS teams, identical in every way except 500 of them have 61 Maris and 500 have 61 Mantle. Which player would you expect to hit more HRs? If you say Maris, its because he hit 61 that year and Mantle only hit 54. Your argument is based upon the assumpion that WIS codes player performance to the exact observable stats for the season. But if you choose Mantle, it is because youre arguing that Mantle in fact had more ability, Maris just happened to be more fortunate, and WIS realized that and coded this way. Another way to say it is that you'd judge 61 Maris as a 45 HR/year guy on average who could max out at 61, while 61 Mantle is a 50 HR/year guy who could have hit 65 that in RL with some luck and better lineup around him, but didn't. Your argument would be based upon perceived ability, not solely actual stats.

So which way should WIS code players? Based upon their observed stats for the year, or based upon an estimate of the players true ability in that year (for which the observed stats represent a fuzzy, noisy set of data points). Should a 61 Maris run 500 times produce a bell curve in which the mean is 61 HRs/season, or should the 61 Maris produce a bell curve with mean much lower, but the 2-3 sigma outlier season is the 61 HRs? I don't have an opinion on this question and wouldn't argue much with either choice. Possibly some of the experienced owners here know more about the WIS thoughts on this than I.

But my point is this. If WIS has chosen to code players based upon observed stats, I would think this decision should filter down to the use of the observed platoon splits also, at least when those splits are known - for design consistency if nothing else. If WIS however has chosen to code players based upon estimates of their true ability; then generic platoon splits can certainly be used because nothing else is a better predictor of the players true performance looking forward.

Interested in what thoughts on this experienced owners have. I'm certainly new around here and just trying to learn.

I'm not an experienced owner but I'd rather have it to bare bone stats.  How could WIS look into every players ability? The bottom line is Maris hit more HR that year and he is more of a HR threat than Mantle at least for that season. I guess they could tweak it a little but boy that would be alot of work! 

Dash- If an opposing pitchers OAV vs. lefties is .214, and his .OAV vs. righties is .306 I am (not always,but sometimes) going to move a guy like 95 Boggs down to 5th or 6th in the lineup and move my righties possibly up one spot in the lineup.  For some reason that is how my mind works, it just makes sense. I understand the sim is just a 5% swing no matter what, I need to look into this more with my teams. For the most part I just do it without really looking too much into the results.