"Best" Stats To Look For?

[Roovin]

Hey guys,

First time posting on this site.

Anyway, I've done some searching and I cannot seem to find what I am looking for. I've been reading the manual and looked at the beginner's guide and am still confused about something.

I have tons and tons of questions but the one thing that I would like to know is what are generally considered the most influential stats to look for when comparing people.

If that doesn't make sense, here is an extremely basic explanation. Lets just say that batting averaging was the most influential statistic to look for. I simply would order by batting average and compare year by year to see who consistently has high batting averages.

Now that is extremely simplistic since BA, to my knowledge at least, really isn't the best way to evaluate a hitter.

Could anyone be willing to give me a general view of what to look for when evaluating a players offense, defense, pitching? Or point me in the right direction?

[Déjà Bru]

I'm a traditionalist. The sabermetric guys will be along soon to make your head spin with modern stats.

So, here's what I look for in players when I look at their stats, not their scout ratings:

• Batting - yes, my eye usually goes to AVG first but I force myself to look at AVG/OBP/SLG together. As you know, OBP includes walks and the more somebody gets on base, the more he's likely to score. SLG is a measure of power and basically more power drives in more runs. One shortcut is to look at OPS which is merely OBP+SLG. That may be the best batting stat of all and I use that plus Runs and RBI's to judge HoF-eligible careers.
• Pitching - yes, my eye usually goes to ERA but the statheads will be all over that one. For pitching, then, the key stat for me is WHIP which is walks and hits per inning. Also good to look at and compare to other pitchers are HR/9, BB/9, and K/9.
• Fielding - I am admittedly weak here. I usually rely on scout ratings alone, and I shouldn't. You cannot look solely at Assists (A) and Errors (E) without taking into account Total Chances (TC). Fielding Percentage (PCT) is supposed to do that for you, but it's boring. These days, unless one has bricks for hands, everybody is above .900 and you are comparing players with, like, PCTs of .982 and .987. Yet, the other stats shown now in OOTP: RF, ZR, and EFF are all fuzzily defined and therefore arbitrary (watch the statheads rip into me for that one). So I guess for fielding I would keep an eye on E, TC, and PCT together.
• Catchers have a couple of extra fielding stats that I would follow: PB, CS% and . . . drumroll, please . . . CERA (Catcher ERA). I actually like that relatively new stat. Take that, sabermetrics!

[endgame]

Nothing wrong with your approach, Bru. There's fact, current theory, personal fiction, and individual favor. Whatever you choose or choose to believe for your use works and if it works for you, you're entitled.

[NoOne]

approach it in an objective way... this is ootp video game, so i am only talking in reference to it, not RL baseball. and, only Batting Average - example should be easy to extrapolate and apply to other things.

What's relevant in the game?

Contact is made up of BABIP, Avoid K's and Power in OotP.

We have access to stats that directly and objectively relate to these ratings


understand statistical environment not changing much is also key: the league environment as defined by a static set of well-callibrated LTM/LT and related Stats and AI settings, plus PCMs that created the players when LTMs are callibrated with auto-calc (-if). in these situations, ratings = probability of results and remain consistently defined.. even if LTMs change a small bit (auto-calc, for example), as long as the totals and other stats and ai settings are unchanged, it will be in the near vicinity, nonetheless.

So,

BABIP stat = BABIP rating
K/AB = Avoid K's
HR/AB = Power

with a large enough sample, these things will tell you fairly accurately what the ratings are. With a little experience and objective observation you can easily translate this into an expected BA for your league's statistical environement. (ignoring that ratings can change constantly, of course... have to consider that, too... if not stats-only, scouting reports help greatly)

because these 3 stats resolve earlier much earlier than 1000AB (a confident BA sample size), you can also deduce whether a small sample BA will increase or decrease over time, assuming consistent ratings/talent. with ratings on, what i said above about easily tanslating is the same thing.

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

I've never cared to deduce the equation for Contact, and each league is a bit different for results... i.e. different league totals will require a different expected BA results for the exact same ratings (not contact, the three that make it). example: higher HR league total means power will have a greater influence on BA, than a lower HR league total.

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

the key anything similar is that it should be a rate and directly or partially tied into the factor at hand -> BA in this case (in ootp this means the individual basic ratings you find in editor, not necessarily the profile - e.g. you don't see overall defense or overall/potential etc, contact is broken up even if displayed added together)

competition obviously plays a role, too.. big part of why you need to consider sample sizes of these values (it's part of what needs to be resolved with time/opportunity)... stop thinking of it as per season... samples of data should be from when they are similarly rated, or you believe them to be if you play stats-only.

physics don't change seaon to season, people do. the slack off, the work harder, whatever... in the game this is slow changes over time based on personality and aging etc etc.. so, applying this is more sophisticated than explaining it... it's all relative though! you accept what the information is - good, bad, ugly and work with what you have.. knowing when it is a more of a guess is useful information, too.

Also, not all resulting stats we have access to so neatly and cleanly relates to the ratings... but, there's always a way to make sense of it. and, relative to this stuff there is empirically a "best" way to do it... simply about playing odds and nothing else.

[Déjà Bru]

Originally Posted by endgame

Nothing wrong with your approach, Bru. There's fact, current theory, personal fiction, and individual favor. Whatever you choose or choose to believe for your use works and if it works for you, you're entitled.

Thanks, eg. That approach worked for generations and feels more substantial then the current scientific alternatives. However, thanks to OOTP, I am slowly coming around to some new stuff, if only in the context of OOTP, not baseball itself. For example, I find myself following WAR in OOTP because, basically, I trust Markus to do something like that right even though I will never know quite what he does in that regard.

[NoOne]

you definitely shouldn't take stats as gospel.. just need to frame them right... somethings are very clear, others not so much.

the concepts remain relative, though... e.g. if we could exclude data due to playing with various types of injuries, or eventually account for their effects on results etc etc... it's just one more layer of sophistication to account for.

do that over and over for each possible force/factor and you can do some cool things for sure. i'd love to see what's going on in some of the more sophisticated analytics departments around the MLB.. the ones that are further along the learning curve. they will always been ~10-15 years ahead if they try.. as time goes that's less and less of a difference, however.

plus, a bit of art involved... although not of the artistic kind .... rather, educated speculation.

[lightgrenade07]

One thing to look for in pitchers is ground ball percentage. IIRC, 55% is considered average. Generally, you want this number to be as high as possible, but it's a situational thing.

If you have a half-way decent infield, you want as many ground balls as possible: fewer chances of extra base hits, let alone HRs, and your infielders will be in range to make some stops.

On the other hand, if your best fielders are in the outfield and you have a park that's not friendly to HR hitters, then you might want to consider pitchers with lower ground ball percentages. If you have outfielders with plenty of range, then a lot of the resulting fly balls will turn into easy outs.

[NoOne]

i think ~average is best maybe bit below or above depending... many factors involved in results, so hard to distinguish.

alot of this depends on how things are defined:

flyballs are the easiest outs by their slash. (somehwere like Fenway may not apply)

line drives have the highest slugging/best slash (probably similar to a "hard hit ball rate" you hear about, some of these may get some altitude but are not 'flyballs')

groundballs have a higher average than flyballs, but fewer XBH likely.

Then, how does this all translate to OotP, specifically -- even more grey added to the picture

defense is going to be key for groundball pitchers... not as important for a fly-ball pitcher (if you read that as not important, not my fault)... etc etc... no 1 answer. it's all relative to the forces at play at that time.

so, what might be best for regular season may not be best for the playoffs... what's best for a leadoff hitter might not be what's best for a power hitter... since we can't swap them out in an unlimited way, you gotta weigh things and make a good choice.

[chazzycat]

If you must choose one stat, WAR is probably the best IMO. That accounts for all sorts of things like defense, baserunning, and perhaps most importantly: position scarcity. For example, if you have two players with identical batting lines, but one is a shortstop and the other is a first basemen (with similar defense ratings)...the shortstop is worth a lot more, and WAR will show that. It is the only all-purpose statistic that lets you compare players across positions like that.

[lightgrenade07]

Originally Posted by NoOne

i think ~average is best maybe bit below or above depending... many factors involved in results, so hard to distinguish.

alot of this depends on how things are defined:

flyballs are the easiest outs by their slash. (somehwere like Fenway may not apply)

line drives have the highest slugging/best slash (probably similar to a "hard hit ball rate" you hear about, some of these may get some altitude but are not 'flyballs')

groundballs have a higher average than flyballs, but fewer XBH likely.

Then, how does this all translate to OotP, specifically -- even more grey added to the picture

defense is going to be key for groundball pitchers... not as important for a fly-ball pitcher (if you read that as not important, not my fault)... etc etc... no 1 answer. it's all relative to the forces at play at that time.

so, what might be best for regular season may not be best for the playoffs... what's best for a leadoff hitter might not be what's best for a power hitter... since we can't swap them out in an unlimited way, you gotta weigh things and make a good choice.

Exactly. When it comes to groundball vs flyball, everything is contextual. The numbers are useless if you aren't taking your park or the rest of your team into account. On the other hand, get to know your park factors & your team's defensive strengths, and you increase your odds of finding pitchers who are best fit to thrive on your team.

Most recently, I was playing as GM for a team in a fictional version of the PCL. I inherited a silver slugger CF who was average defensively, but I was able to put together a solid infield with a golden glove catcher and short stop. Combined with my home field's HR-friendly ways and ~.900 range double/triple ratings, and that was pretty much ideal for groundball pitchers. A couple years in, I traded a four-star flyball SP for another team's three-star groundball SP, and I found that both pitchers improved their ERAs after the trade.

Now I'm in charge of a fictional team in the 1880s when nobody's hitting HRs at all. My strategy this time around is all about finding the least crappy OFers and pairing them up with flyball pitchers for those easy outs.

It's all contextual.

[Roovin]

Thanks guys for all the help.

Going to take me a long time to be able to apply all this but knowing the basics will help out a lot

[NoOne]

just take one rating and try to see that on its own before you incorporate more. work out a little system of your own, after that it's just more wrinkles but same conceptually repeated over and over.

avoid k's is an easy one to start wiht, and i don't htink it takes a huge sample size to normalize. after that it's just understanding league baseline to compare to deduce their true rating (whether visible or not).

[Big T]

OPS for me, for batters and pitchers when developing lineups and reliever usage. WAR for roster planning (promotions and demotions) and evaluating off season.

[NoOne]

if you focus on one stat, more power to you... just in case that's what you think i was intimating, not so much.

the more fundamental and basic the stat is, the better off you are at knowing preciseley what it means. so, you'll want numerous stats that define in some cases exclusive or somewhat exclusive things.

relative to ootp, since ratings control everythign, you want the stats that most directly relate to each rating -- in some cases many for just 1.

e.g. i often relate ops+ or WAR to the comments most agree on when referring to Overall and potential. essentially the same logic, but for some reason not viewed in the same light by many, i don't think. although ops+/war is probably better info than ovr/pot any day, any league.

sophisticated stats are not worthless by any means. they most definitely help make the data relateable. especialy if you ahve league evolution or change baseline statistical environments etc etc. they provide an undeniably important feedback in these instances... if very familar with your world and baselines don't change, it's second nature eventually.

[cooperthegiantscat]

I use the following...And note I've always been a stats guy and always will be a stats guy. Just because someone is 80 overall and has a 77eye doesn't make him a good hitter. My experience has shown that basic stats don't tell the whole story and need to give context. So if you can use metrics such as those below you will see quickly where you are lacking. My Detroit Tigers file shows the whole team was a strikeout machine so I focused on guys with lower SO% and higher OBS for example.

Pitchers:
IP
FIP
WAR

Batters
PA
SO%
BB%
OBP
OPS
ISO
BABIP
WAR

I try to minimize greedy players and low IQ players when I can. Nick Hundley has me on myast nerve right now...
WAR

[sonicqaz]

Originally Posted by Déjà Bru

Thanks, eg. That approach worked for generations and feels more substantial then the current scientific alternatives. However, thanks to OOTP, I am slowly coming around to some new stuff, if only in the context of OOTP, not baseball itself. For example, I find myself following WAR in OOTP because, basically, I trust Markus to do something like that right even though I will never know quite what he does in that regard.

If you gave me 1 night and an open mind, I could convince you that sabermetrics is going in the right direction (in regards to real baseball), some of the advanced stats are the best we have to evaluate performance, and that it's still not perfect. There is still no single stat that lets you evaluate players head to head alone, which is probably a good thing, because then the game would be boring. Knowing which stats to use, knowing which stats to ignore (even when they're typically reliable), and evaluating based on what you are looking for exactly (what do you mean by best player, Pitcher A might be better than Pitcher B but Pitcher B is a better fit for Team C) is the coolest part about sabermetrics.

[Roovin]

Thanks for all of the tips guys!

I've slowly been able to implement a lot of these.

[84Cubs]

I like wOBA