Hit/pitch splits are commonly associated with fantasy baseball auction drafts auction drafts – e.g., spending \$180 on hitting, \$80 on pitching or 70/30 hit/pitch split. The discussions around hitting vs. pitching in snake drafts tend to be less mathematical in nature – often limited to debates on which round to draft one’s first starting pitcher and relief pitcher.

This post will demonstrate how to calculate hit/pitch splits for snake drafts as well as analyze 2013 Razzball Commenter League data to see what the most optimal hit/pitch splits are.

Part 1: Converting Draft Picks Into Dollars And Percent of a Team’s Budget

I estimated the value of each pick by tying it to the value seen in our Player Rater for that rank (so pick #1 = top ranked player). I ended up subtracting 1 from each spot so that the team totals neared \$260 (came out at \$255). The value of each fantasy baseball snake draft pick can be found here. I recommend ignoring the dollar estimates and just sum up the ‘% of Team Draft Value’ column.

Note: for those that like to create their own top rankings prior to a draft, this chart can help you understand whether your rankings have the intended hitter/pitcher skew. I showed the hit/pitch splits within various published rankings sources in my Fantasy Baseball Rankings Review for 2013.

Part 2: Analyzing 2013 Razzball Commenter League Data

Two caveats before you read further:

• This data is based solely on 2013. Every year is a bit different in terms of how players do versus pre-season expectations. Sometimes there are years where seemingly every reliever loses their job – other years they all do great. etc. While I do not think 2013 is particularly anomalous, that does not mean 2014 will play out exactly the same way.
• As an upcoming analysis will prove, active RCLers perform better than average (and vice versa for inactive ones). So some practices such as ‘punting SP’ may appear more successful based on the below analysis because they are disproportionately deployed by active players. The inverse may hold true for inactive players (including auto-drafters). I could/should have neutralized this potential bias but I realized it too late in the process.

Based on my values per draft pick, the max/min for hit/pitch split in a 13 hitter/9 pitcher/3 bench format like RCL is about 88% hitting/12% pitching (if you drafted hitting 1-17 and pitching at spots 18-25) and 25% hitting/75% pitching (if you drafted pitching 1-12 and hitting 13-25).

The below chart shows the results of all 768 RCL teams with the ‘x’ axis representing their % of draft invested in hitting and the ‘y’ axis representing their final standings points. The red data point is me. The yellow data point is Grey. The average hit/pitch split across all teams was 66/34 with a median of 66.5/33.5.

The trendline in the graph represents the ‘best fit’ if one was trying to predict a team’s standings points by the percentage of the draft they invested in hitters. The trendline here has a slightly upward slope indicating a slightly positive relationship – i.e., the more a team spent on hitting, the better they did in the standings.

If we focus on the teams that were within one standard deviation away from the median (60-71.8% hitting – roughly 3/4 of teams), the trendline is almost perfectly flat (see below).

The 92 teams that invested less than 60% in hitting did fare worse (56 standings points) than the average team (65.5 points). The 85 teams that invested 72+% in hitting fared about average (65.8 points). Thus, it appears that based on 2013 data, there is very little impact in a team’s final standings based on drafted hit/pitch split as long as they clear 60% for hitting.

Before I show how this impacts hitting and pitching standings points alone, I think it is worth highlighting what I think is the biggest misconception involving snake draft hitting vs. pitching investments. While early rounds are disproportionately more valuable than later rounds, I think most drafters overrate the value of early rounds.

A telling example is comparing Grey’s draft vs mine in the RCL Experts league. I ended up drafting SPs in the 4th/5th round (Lee/F-Her) and my first RP in the 7th round (Chapman). Grey drafted his first SPs in the 6th/8th rounds (Greinke/Gallardo) and first RP in the 10th round (Holland). Given these differences, I think the common perception would be that I significantly over-invested in pitching compared to Grey. If we focus only on SP1/SP2/RP1, Grey had invested 15.5% of his overall budget while I invested 19.2%.

But once I grabbed F-Her and Chapman because I thought they were great values, I purposely pushed back subsequent pitching picks to solidify my offense. Grey ended up investing an extra top-11 pick on a pitcher (Niese). Thus, by the 11th round, Grey had actually spent more on pitching (20.4%) than myself (19.2%). I then spent my next 4 picks on pitchers (3 RP, 1 SP) while Grey only spent two of those picks on pitchers (1 SP, 1 RP) and that difference of about 6.3% of budget is what eventually led to my relatively small higher pitcher investment (63.7/36.3) than Grey (65.5/34.5).

 Grey’s Draft Rudy’s Draft Round H/P Pick Value % H/P Pick Value % Hitting 14 65.5 14 63.7 Pitching 11 34.5 11 36.3 1 H 14.3 H 14.4 2 H 9.0 H 8.9 3 H 8.3 H 8.4 4 H 7.2 P 7.1 5 H 6.7 P 7.0 6 P 5.9 H 5.8 7 H 5.6 P 5.6 8 P 5.2 H 5.2 9 P 4.9 H 4.9 10 P 4.3 H 4.2 11 H 4.0 H 4.1 12 P 3.5 P 3.5 13 H 3.3 P 3.4 14 P 3.0 P 3.0 15 H 2.8 P 2.9 16 P 2.2 H 2.1 17 P 1.9 H 2.0 18 H 1.7 H 1.7 19 P 1.6 P 1.6 20 H 1.3 P 1.2 21 P 1.1 H 1.1 22 H 0.8 H 0.8 23 P 0.7 P 0.7 24 H 0.3 P 0.3 25 H 0.0 H 0.1

My point here is focus less about hard rules on when to draft your first SP/RP and more on making sure you maintain your desired hit/pitch split. If you go with Clayton Kershaw as your 1st or 2nd round pick, you will need to adjust your subsequent picks unless you are okay with a high overall pitcher investment. For instance, if I were to draft Kershaw, I would probably cap my SP investment at 4 and might not take my 3rd/4th SPs until past the 16th round.

The below scatter graphs show the impact of the percentage invested in hitting against teams’ final hitting standings points and the same for pitching. Where there is a positive correlation for hitting ( R^2 of 8.65%=29% correlation), there is virtually no correlation when it comes to pitching (R^2 of 0.22% = 5% correlation). If I limited this to just those teams within 1 standard deviation of the median, the correlation %s would still show hitting to be greater (22% vs. 6%). This seems to confirm the common perception that hitting is ‘more reliable’ than pitching…..but I am not so sure.

In the 2013 Fantasy Baseball Rankings test, the pitcher pre-season rankings actually correlated a little better with the end of season results than with hitters. This seemingly negates the ‘hitters are safer’ argument. So why does pitching make for a poorer investment?

My hypothesis on seeing the data was that this is caused by the ‘streaming effect’ – i.e., there are a higher percentage of pitching innings thrown by FA/waiver pickups (aka streams) than with hitter ABs. This ‘waters down’ the impact of drafted pitchers in relation to hitters. The fact that our Razzball Commenter Leagues use a GS cap vs an innings cap amplifies the ‘streaming effect’ because savvy RCLers pad their counting stats (and potentially help ERA/WHIP) with middle relievers.

Given that this phenomenon would impact SP value more than RP value, I ran the same analysis splitting out SP investment from RP investment. Wow! Starting Pitching investment is nearly the inverse of hitting – a negative 27% correlation (R^2 of -7.82%) with a team’s final standings points. When I narrowed the analysis down to those within 1 STDEV (18-30%) of the median (24%), the correlation only decreased to -18% (R^2 of -3.3%). The performance of the 50 teams with the heaviest SP investment was awful (average of 52 standings points, 13.5 below average) while the performance of the 50 teams with the lowest SP investment was strong (73.6 standings points, 8 above average).*

* REPEAT OF EARLIER CAVEAT (reads like that should rhyme) – I would bet that the low SP investors are much more active players than the high SP investors. As the red dot on the scatter graph shows, I veered slightly to the left of consensus in starting pitcher investment (by drafting only 4 SPs) but I am not a fan of relying as heavily on streaming pitchers as those to the left of me.

I

The story is the opposite for relievers. While they do not return the same level of investment (measured by standings points) as hitters (R^2 of 3.97% vs. 8.65%), they were a much better investment than starting pitchers. This strongly contradicts the ‘don’t pay for saves’ argument that I have disagreed with for years (here’s a post I wrote in 2008(!!!) on it). When narrowed down to those within 1 STDEV (5-14%) of the median (9%), the correlation decreased from 20.3% to 12.3% (R^2 of 3.97%). The teams who were the 50 heaviest investors in relief pitchers (me – in red – being one of them) averaged 70.7 points (5 above average) while the 50 lightest investors averaged 53.6 points (12 points below average). Those top 50 investors averaged 5.7 relievers where the league average was 3.7 relievers so it is as much about reliever volume as it is about quality (fwiw, i drafted 7 – Aroldis, Reed, Street, Jansen, K-Rob, Marmol, Al-Al)

Final Conclusions/Notes

• Converting draft picks into \$/% of budget increases transparency as to how much you are investing in hitting vs pitching during snake drafts.
• When putting together rankings for RCL or leagues with similar formats, I suggest building your rankings with a 66/34 split so that it is easier to identify draft bargains. I will be moving to this in my pre-season rankings that I’ll be posting on this site (last year, I used 62/38 while Grey used 69/31).
• The common ‘sweet spot’ among RCLers is 60-72% hitting / 28-40% pitching. Within that range, there is very little difference in how teams performed. If you are going outside this range, it is better to be skewed toward hitting (72+%) vs pitching (>40%).
• The common ‘sweet spot’ among RCLers is 18-30% SP / 5-14% RP. Based on the findings above, it would be wiser for RCLers to err on the light side for SPs and heavy side for RPs.
• Given these correlations are not overwhelmingly massive, if you have a particular strength in terms of drafting (hitters vs SPs vs. RPs), I would follow it and only caution against investing too much in starting pitching.

## 65 Responses (Jump straight to the comment form)

1. Q says:

A little off topic:

Thought the link converting round to \$/% was interesting. I play in an auction league, but it’s a points league, so the points shares haven’t been as useful to me.

I know the chart is based on a 12 team league, w/ 276 players. Any advice on converting the % numbers to a league with just 240 players picked?

• You can just use my 10-team end of season player Rater – razzball.com/playerrater then click 10-team for ESPN. That is based on 220 rosters players but the last 20 are all around \$1 anyway

• Spammer Jay says:

@Q: the % would stay the same right? dollar values would drop by 13% no?

• Spammer Jay says:

@Spammer Jay: or increase by 13% i mean.

• I don’t think it is quite that easy but the % aren’t that far off. Can use the end of season player raters to calc this for various other formats

2. Paul H. says:

Famtastic stuff Rudy. Pretty interesting to see the impact of streaming on the RCL leagues. Makes me realize how important it is to adjust your draft strategy to the league you play in rather than just draft by rankings or projections – even if they are Grey’s. I play in a relatively deep league – 25 active players and 10 reserves that limits FA acquisitions to 10 all season long. It basically means the streaming strategy is out the window for me in terms of pickups and drops. But I can employ a similar strategy by picking streamable pitchers in the reserve rounds of the draft. Keep up the good work!

• Thanks. Yes, limited transaction leagues change things a bit. SPs on the bench in that format is great for weekly matchups. In NFBC, my goal is to have the most projected SP IP post-draft.

3. Fireball Feliz says:

Interesting read, good stuff. My 6×6 (QS and OPS) roto league also has a GS limit which makes closers and middle relief guys more valuable.

• The addition of QS increases the value of good SPs from average SPs though. I’ll have preseason \$ for 6×6 QS. Could look at end of 2013 5×5 vs 6×6 QS to see how that impacted SP vs RP values…

4. Goose says:

Great stuff, Rudy. Thanks.

This logic seems to apply very well to my 14-team 6×6 league, with holds as a cat. In order to succeed in our league, it seems clear that investing early in RP is necessary.

We have a 5P-2SP-2RP roster format, so really it means 7 RPs/team with a rotation in and out of those two SP slots. With a GS cap and roster size limitations, there is almost always easy value to be found on the wire from streaming SP – thus no need to invest heavily, percentage-wise, in SP in the draft. If you have an ace and fairly solid 2/3, you can stream all year with ease. Also, with most every 7-8-9th inning guy on someone’s team, you have to invest in RPs in the draft in order to ultimately succeed.

This skews pitching values (SP vs RP) in a crazy way, which makes it somewhat hard to research as compared to most leagues – but the % of hitting value vs. % of pitching value is consistent, I believe, with your breakdown here. So, again, thanks!

• Thanks. Yes 6×6 holds leagues would decrease SP value vs RP. I do a 6×6 w/holds pre-season player Rater for 12-team that u could adjust for 14 teams…

• Goose says:

@Rudy Gamble: Estupendous! Will look fwd to that.

5. salachor says:

awesome post. great job.

6. Dave Uitti says:

Terrific. Thank you. Keep these types of articles coming.

• There will be 1-2 a week for the next month or so…

7. Sky says:

• We are the champions my friend…

8. the rock says:

are you investing in any NFBC DC leagues this year?

• @the rock: yeah, 1-2 at the \$150 level. going to have a post up later this month seeing if any other commenters want in on a ‘razzball writer and commenter NFBC Draft Champions league”

• the rock says:

will you approach the draft differently then last year in the NFBC and what is your biggest challenge in this type of league? Thanks

• @the rock: i liked my strategy from last year but can see sticking w/ younger pitchers for my first four (last year’s C-Lee, Sabathia, Beckett, Vogelsong went 1-for-4) and aiming for better overall AVG.

• mrrr says:

@Rudy Gamble: Was planning on giving NFBC a try this year. I’d be interested in a razzball NFBC league. What timeframe were you thinking of starting?

9. Robby D says:

I once loved a girl who was into statistics. Let’s just say her deviations were anything but standard.

10. SteveNZ says:

My H2H league (either 12 or 14 teams this year) is a 6×6 with OBP and Holds being the extra categories. Under normal circumstances, this is tailor-made for the strategy of really loading up on offense, waiting on starting pitching (then drafting 3 solid ones or so), grabbing as many closers and MRs as possible and then streaming the hell out of it.

BUT, this league has a rule that we can only use 10 ‘adds’ per week, which is probably less than you need to stream effectively, especially if you want to grab hot hitters, maximise playing time etc.

Last year I made the final by going relievers only, looking to win Holds, Saves, ERA and WHIP every week, while having a strong enough offense to win most of those categories.

Obviously this means I’m pretty much punting Wins and Ks. So my question is, can you see a way to stream effectively under my league settings, or am I on the right track with the relievers-only strategy?

I do feel that the Razzball tools would give me an advantage streaming, though there is not a lot of starting p[itching on the wire, with most guys loading up on it at the draft.

Jeez, I hope all of that makes some sort of sense…

• @SteveNZ: yup, makes sense. going 6×6 with Holds devalues SPs vs. hitters. I can see drafting a couple of SPs (vs punting) but if the league drafts them heavily, I can see focusing on hitters and Closers/MRs. I’d focus less about ‘adding’ new players and more about optimizing your bench. Having MRs to rotate in /out is nice. I’d also look for players that have good matchups (Colorado hitters, lefty hitters, maybe a 1B/OF or C,1B type to rotate in for player off days). Just some thoughts…

• SteveNZ says:

@Rudy Gamble: Thanks Rudy – that’s great. Our rosters are pretty shallow (no CI or MI, only 3 OF), so do you think it’s fair to say (all things being equal) that I could afford to fill all my hitting spots before looking at pitching and still draft 2-3 solid SPs?

• SteveNZ says:

@SteveNZ: Given the depth of pitching, I mean…

• @SteveNZ: I don’t like punting SP that much. I would want 2-3 SPs before i draft C and one of 2B/SS (so maybe draft one and punt one).

I feel like you’re the voice of reason with pitching on this site. People take the grey approach way too far on here, often drafting awful staffs in the hope that every grey sleeper pans out. While pitching as a whole is more volatile from year to year, you cannot overstate the advantage of a low-risk ace to anchor your staff.

• Grey and I aren’t that far apart in practice (as shown in this analysis) but he’s less amenable than me to early round SP picks

12. SwaggerJackers says:

I enjoyed the article from 2008. Man, back then I wasn’t even playing fantasy baseball. Now I look at this shizz daily.

• I just get more addicted every year…

• SwaggerJackers says:

@Rudy Gamble: Have you found a way to make any money from it? Your ad revenue should be decent from the visitor volume and reader engagement from comments but I have no idea if that means you make \$2 a day or \$200?

• SwaggerJackers says:

@Rudy Gamble: Maybe you can sell some of the fantasy tools you’re developing to a paid site like Baseball Prospectus.

OR

You could make the stream o nator a paid App.

We have to come up with some way to get you and Grey on Shark Tank to pitch a fantasy baseball related business venture!

13. AJA says:

Wow! This is a great analysis. Thanks for putting it together.

Here’s my unsolicited 2 cents. You may be introducing a bias by assigning draft position values using end of year rankings. This ignores the undrafted players who come up and outperform guys who were drafted. This effect disproportionately impacts picks in the later rounds. For instance, in expectation the 1st pick overall will give value equal to that of the best player at the end of the year. By comparison, the 100th pick might in expectation give you the value of the 110th rated player at the end of the year because you would expect 10 undrafted players to come and outperform the 100th rated player.

Another way of assigning values to draft picks (for future anlyses like this) would be to fit a curve to the relationship between average draft position and realized dollar value. If possible it would be best to use multiple years of data to minimize the impact of outliers. I doubt it would make too much difference here (although it might distinguish between yours and greys pitching investment more than you let on).

OK that was my 2 cents. Sorry to nitpick.

• Good points. The actual realized value per pick will follow the same curve but you’d need a massive multi-year sample to smooth it out so pick 41 doesn’t net more than 43, etc. But I’ll re-run this analysis next year to see if it holds. If I change RCLers behavior, it should reduce the inefficiencies I noted. But hopefully it also means everyone in home leagues kicks ass!

14. Hurleyman says:

What do you see in the optimal SP/RP splits in eagles where you cannot stream (weekly periods)?

• The 60/40 to 72/28 range is all fine with me but I’m using 67/33 as my target this year for all mixed leagues (maybe 69/31 in al/NL).

• Brian England says:

I assume all your leagues are Roto.

• @Brian England: Yeah, Points leagues would vary but I imagine the mix would be closer to 60/40.

15. Obmij76 says:

Hey Rudy, are you planning on posting a projections review this year? I enjoyed your rankings review and loved both posts last year. Thanks again for all your hard work!

• Yeah, sometime this month. Have a whole trove of analyses that should hopefully please.

• Obmij76 says:

@Rudy Gamble: Awesome! Your posts are invaluable

16. Jacks says:

Rudy! Could you save me a ton of time and tell me which reduces overall reliever point contribution more in a roto format, straight Ks or K/9? My league is likely replacing Wins and QS and introducing a fifth K based category. My concern is that we’re really truncating closer contribution with QS instead of Wins and I was thinking we should use K/9s instead of Ks to preserve a bit of weight for relievers.

• Jacks says:

@Jacks: That should be replacing Wins with Quality Starts.

• @Jacks: Took me a while to think through this one….but K/9 helps relievers. If you completely punted starters, you can still do really well in the other 4 categories. When you have W/QS and K’s, it makes a reliever-centric strategy non-advisible as you’d be punting two categories.

I don’t like the reliever-centric strategy so I’d go with K’s.

17. The Unchosen says:

I’ve been messing with the Point Shares/Player Ranker and really like the system. Problem is my league has too many position and category differences to use it for my rankings. Is there a discussion on this site that explains how to build a point share formula for a given category?

I’ve captured what our league average player looks like per position and per category based on last year stats. So I think I have the basis. Just don’t know any easy way to convert that to a formula per category (e.g. multiplier for each OBP point above/below avg, multiplier for each HR, etc).

• I haven’t publicly documented my point shares methodology – mostly to protect state secrets.

I will be using the survey results to help think through ways to better accommodate alternative categories. What format is your league?

• The Unchosen says:

@Rudy Gamble: Where’s the survey? I’ll enter it there. Understand the state secrets comment. It is a great feature over other sites.

I started to reverse engineer the formulas for each category and can likely get there but it’s painful. You’ve got most of our unique stuff covered in your options (6×6, OBP, QS). We just use a lot of “net” vs “gross” on categories and we use 2 catchers.

I’ll save myself some time and use the PreSeason Ranker (once you load the projections) to get an initial ranking and then adjust from there.

18. Charles says:

In my league over the years I’ve found a very strong correlation between the number of transactions a team makes and its overall success. Adding that factor to the analysis may help isolate the value of the SPs in the draft by removing the impact of streaming. The assumption being that high transactions indicate a team that likely streamed.

• I don’t have transaction volume per team. I do have team AB/IP which helps identify teams that streamed RPs.

• Charles says:

@Rudy Gamble: Tragic. It’s typically listed right there on the ESPN league standings next to all the stats.

• @Charles: Our commish didn’t capture that point. But it doesn’t matter much b/c I think AB/IP totals is a better gauge of effective streaming (e.g,. transactions that are not increasing total IP or AB have little impact on results)

I think the reality is that you need to be active to win a daily league. Here’s a sneak peek at some of the data in my upcoming post that quantifies the impact of maxing IP/AB. The difference between a well-drafted hitting team (75th percentile) with 50th AB and 75th percentile AB is 3.4 standings points in a standard RCL league. The impact is higher if they drafted poorly (more to gain).

My guess would be that – adjusting for active usage using AB/IP – that drafting SPs wouldn’t have as much of a negative relationship with standings points but I do think there’s a negative relationship since the replacement level for SPs leveraging streaming is higher than for hitting and RP. I’ve got another analysis working on that one…

• Charles says:

@Rudy Gamble: I follow. AB/IP should be a great measure. I agree that to truly to see the value of draft picks, you need to adjust for activity levels by the owners. You may even have to account for the (possibly strong) correlation between high-activity and hitting-heavy drafters. Love this stuff, btw

19. Tyler says:

Hey Rudy!

Will there be an article this year showing the differences in proposed value vs. Yahoo ADP to find value in the draft? It was very helpful last year.

• Yup. Once I finish with projections. Prob mid-Feb for projections, earlyarxh for yahoo adp post

• Tyler says:

Awesome. Thanks!

20. UncleLarryWalker says: