Draft Speed or Pound the Power?

Introduction

On the latest episode of The Great Fantasy Baseball Invitational Podcast, I posed the following question:

In 2020 fantasy baseball drafts for roto leagues, which will you do early on?

A) Draft Speed Early
B) Pound the Power

Simultaneously, I posed the identical question on Twitter, yielding the following results:

The Twitter responses, as well as the members of the TGFBI Beat the Shift Podcast panel were pretty evenly split on what was more important to focus on. Obviously, fantasy owners need to focus on both; players who can amass a broad base of stats are ideal. But the question is still a valid one – in a vacuum, all things being equal, which player type should you favor in a draft?

My immediate reaction to this question was to side with the “pound the power” camp. Speed players were unquestionably pushed up in drafts this year – higher than their projected values indicated that they should be. If you came into an auction or draft with a value sheet, you would notice that a 4^th round speed talent often went in the 3^rd round, or a $15 speed demon typically went for $20. The cost for stolen bases was on the high side, but was it justified?

To help us answer this question, let’s take a look today at some “hit” and “bust” rates for power vs. speed players. Let’s also take a look at the accumulated value that each player-type has returned.

Definitions

Let’s place all batters into four different groups. We will categorize all batters solely based on their pre-season projections as follows:

• Power – Projected for at least 25 HRs, but not 15 SB.
• Speed – Projected for at least 15 SBs, but not 25 HR.
• Combo – Projected for both 25 HR and 15 SB.
• Neither – Projected for less than 25 HR and less than 15 SB.

For this exercise, we will use the Steamer pre-season projections as of March 28, 2019. We will also ignore all other scoring statistics. Batting average and run production stats, although extremely important to fantasy teams (and is crucial to early round selections), will be overlooked for the time being. A player like Xander Bogaerts who was projected for 19 HR and 9 SB – would fall under the “Neither” category.

Let’s also define a batter’s season success (as of today) to be one of the following – also called a player’s verdict:

• Hit – A player who in 2019 has accumulated either:
  • 20 HR
  • 14 SB
  • 18 HR & 12 SB
• Bust – Everyone else

We will define the success of a player who is on a full-season’s pace for either the “Power” category, the “Speed” category, or a sizeable showing in both categories.

We will also use a player’s accumulated auction value. For this, I ran the FanGraph’s Auction Calculator with 15 team 5×5 NFBC settings. These auction values are prorated for the % of season completed and are not annualized. [As I will only be comparing relativities of player categories, there is no need to annualize the figures.]

One final item to note – in the ensuing analysis, I refer to the “Top N” hitters. For this, I refer to the NFBC ADPs of each player, for the month of March 2019. A player who is in the top 100 for example, means that on average, the player was selected in March 2019 NFBC drafts within the first 100 hitters.

Hit Rates

Let’s take a look at the success rates for each of the four projection categories. To set the stage, let’s first take a look at the total number of players that fall in each group.

The “Neither” cohort takes up 22% of the top 50 hitters, 34% of the top 100 hitters, and 47% of the top 150 hitters. There is nothing strange about that. From the above, we see that the “Combo” players are almost all found within the top 50 – and in fact, other than Rougned Odor, they were all found within the top 15 hitters selected.

For the “Power” vs. “Speed” selections, we see that there were far more power players within the top 50 [roughly 2 to 1 ratio]. Speed was more even with the power players in the next 50, and then the power dominated the 101-150 range once again. Speedsters such as Dee Gordon, Mallex Smith, Ender Inciarte, A.J. Pollock, Jose Peraza and Byron Buxton were examples of some appearing in that middle-range.

Now onto hit rates. Let’s take a look at how successful each category was in 2019:

The combo players were all successful. In fact, all 9 combo players hit at least 20 HRs in 2019 thus far. 6 out of the 9 players also have at least a dozen swipes. Only Mike Trout (10 SB), Javier Baez (10 SB) and Rougned Odor (8 SB) have 10 or fewer steals.

The key takeaway here for combo players is the 100% figure. Of the players projected for 25 HR + 15 SB, they have produced most of the statistics that they were projected for. Amazingly, and likely fluky (or maybe not?), none of these players were injured enough during the season to derail them entirely from achieving their “Hit” status. A combo player in 2019 was an extremely safe investment [looking at power and speed alone, ignoring the other scoring categories].

The power players were the next most successful group. Almost 80% of the top 100 and 150 were hits. Within the top 50 (early in drafts), there were only 3 busts out of 20 power hitters – Aaron Judge, Giancarlo Stanton and Khris Davis. Judge & Stanton’s down seasons were injury related, but Khris Davis – who was one of the safest fantasy players heading into ’19, was a huge healthy disappointment. [Who knows? Perhaps we will later find out that Davis had a lingering injury all season long and played through it.]

Let’s move now onto speed. At the top, the success rate is decently high (73%). Afterwards, the speed hit rate noticeably drops off. For hitters 51-150, only about half of the players were successful.

Looking through some of the speed busts:

• Andrew Benintendi was categorized as a speed player, because of his 18/18 projection. He has only produced a 12/9 season to date.
• Other notable disappointments in the top 50 include Jean Segura and Lorenzo Cain.
• For the following 100 players, some busts include A.J. Pollock, Ender Inciarte, Jake Bauers, Jose Peraza and Wil Myers.
• Even in the “hits”, the stolen base totals fell short of expectations [ex. Jonathan Villar, Dee Gordon, Yoan Moncada, Starling Marte, Trea Turner, Whit Merrifield, Jose Altuve].

Magnitudes

To recap what we learned about hit rates in 2019:

1. Combo players had an almost certain hit rate.
2. Power players had a better hit rate than speed players.

But what about magnitudes? Which projection categories have the best return per success?

Let’s take a look at average accumulated auction values per player for each category-verdict combination. For this analysis, accumulated auction values will be capped at $0 (below replacement level players will count as $0).

For the top 150 players:

• Combo hitters returned an average of $23.
• Power hitter “hits” returned an average of $14.
• Power hitter “busts” were on average below replacement.
• Speed hitters retuned an average of $11 for “hits” and an average of $5 for “busts.”

We already saw that the speedsters have a lower hit rate than the power bats. Not only that, when a speed player did hit, they returned a value less than that of an equivalent power bat. What is interesting though is that the speed busts still returned positive value on average – especially up top. The power “busts” did not return any value.

Combining the hits and busts for each category, produces:

Here, we can see that the power hitters clearly outperform the speedsters. The speed players only return a marginally better value than players in the “Neither” category. Since the “Neither” players cost far less to acquire than the speedsters, the expected profit on a speed player is by far the lowest of any group.

Conclusion

Simply because 2019 played out a certain way – it does not a guarantee that next year will be the same. 2020 may ultimately differ for each player category. However, it is still important to study the past, and we can learn from the 2019 empirical results. Let’s recap what the numbers are telling us.

Looking at the Power vs. Speed players:

• Power players have a better hit rate.
• Power players have a better expected return on the players that hit.
• Power players have a better overall return.

From a return on investment perspective, power hitters are the better investment.

However, in rotisserie leagues, it is hard to completely punt speed. Fantasy owners still need to buy stolen bases to compete. The question is – Where is the most effective place to purchase speed?

From the above analysis, I would conclude the following:

• It is worthwhile to purchase a combo player, even for a slight premium at the top of the draft. It is a safe investment. Additionally, banking that base of stolen bases allows you to tilt towards purchasing more power during the rest of draft, which will give you a better overall ROI.
• If you purchase a speed only player, the best place do so is within the top 50. Whereas the power hit rate remains stable throughout the draft, the speed hit rate declines in the 51-150 range.
• Rather than overpay for pure speed later on in the draft, look to acquire a few players who can steal a moderate number of bases along the way (9-13 projected SBs). It would have been wise in ’19 to target players later on such as Kevin Pillar, Adam Eaton, Willy Adames, Niko Goodrum, Marcus Semien, etc.

Other Assorted Notes (In no particular Order)

• For 2020, it is possible that Mike Trout would drop out of the “Combo” category. ZiPS 2020 projections currently has Trout projected for 36 HR and 20 SB, but I have a feeling that the speed projection may lower as we get closer to next season.
• If my comment above is correct (that Trout will no longer be a combo projected player), then based on this article’s analysis, we can make a strong case to draft Ronald Acuna Jr. ahead of Trout in 2020.
• 25 HR and 15 SB are somewhat arbitrary filters. The exact thresholds of course could be debated. I rounded what I thought was a reasonable figure down to the nearest five.
• This analysis is based off of the Steamer projections. Had we instead used the ATC Projections, the following players would not have been classified as “Neither” players, yet were hits: Anthony Rendon, Gleyber Torres, Eddie Rosario, Max Muncy, Rafael Devers, Paul DeJong, Hunter Renfroe, Trey Mancini and Manuel Margot.
• On the flip side, if we had used the ATC Projections instead of Steamer – a few bust players would not have been classified as “Neither.” They include David Dahl, Mitch Haniger, Nicholas Castellanos, Brian Dozier and Aaron Hicks.

What should the ideal draft strategy be for hitters in 2020? Feel free to weigh in on the comments.