As is the time-honored tradition on this blog, started here a week ago, I will review start/sit recommendations from the previous week’s Matchup Report — specifically, the ones I stated I had the most confidence in. For Week 2, each of my top five recommendations was a start recommendation. Three of them worked out well, though one was for reasons that were entirely different than the ones I argued for. The other two wound up being bad calls.

If there is a theme from this week’s picks, it’s that making roster decisions based on matchup data can be a fraught process this early in the season. This is not about sample size. As those who subscribe to the Matchup Report know, these picks aren’t based merely on this season’s data, and in many cases, they are informed by several seasons’ worth of data.

The picks that went awry — namely the start recommendations for Ryon Healy and Manuel Margot — did so because some of the opponents they faced have not been their usual selves so far this season. Given that we are only just more than two weeks into this season, these apparent changes may be nothing more than small sample artifacts. There is some chance, though, they are signs of actual improvement since last season.

If I look back to early last season, I almost certainly would have picked on hitters who were facing Dan Straily once he was put in the Reds’ rotation. In all likelihood, it wouldn’t have gone well, as he held opponents to a .212 batting average in his three April starts. That improvement held up, as Straily’s batting average allowed at season’s end was .220. How were we to know that Straily would allow a total of five flyball base hits that weren’t home runs all season long (per Baseball-Reference’s hit classifications)? He did a better job of preventing batters from pulling flyballs than he had since his previous best season of 2013, but in the early week of 2016, we didn’t know that was coming.

Frequently, established players sporting a new look or a new skill within the first two weeks revert back to their former selves, but that’s not always the case. Was I faked out on my Week 2 misses by players who changed their games in the offseason? Or were these just cases of some random statistical fluctuation that can (and will) happen in any given week? Only time will tell…

On we go with the recommendations review, starting with my top pick, in terms of confidence level, for the week.

1. START Wily Peralta.

What I wrote: Aside from Kendrys Morales and Joey Votto, Peralta won’t have to face any worrisome hitters batting from the left side. Over his last 13 starts dating back to last June, Peralta has held righties to a .199/.246/.321 slash line. With two starts and these favorable matchups, Peralta is close to must-start. He is also just about universally available in mixed leagues.

What Peralta produced: 2-0, 2 QS, 12 IP, 5 ER, 8 K, 1.25 WHIP.

Takeaways: While Peralta has likely overperformed during this extended period of dominance against righties (.219 BABIP), his 23.3 percent strikeout rate and 60.5 percent ground ball rate against them over a 15-start stretch says to me that we can trust him in many of his starts.

2. START Ryon Healy.

What I wrote: With a series at the Royals and a home series against the Astros, park factors would seem to be working against Healy, but so far in his young career, he has hit for power just about everywhere. He really likes hitting against flyball pitchers (.306/.340/.590 in 141 plate appearances), and he will face one in each of the three games in Kansas City (Ian Kennedy, Jason Hammel and Jason Vargas).

What Healy produced: 4-for-19 (.210) with a double (off flyball pitcher Travis Wood) and eight strikeouts.

Takeaways: Kennedy and Vargas are off to surprisingly good starts, though Kennedy’s looks fluky, given that he is still allowing flyballs (51.1 percent rate) and hard contact (42.5 percent rate) at high rates, yet has allowed an .075 Iso to date. Vargas, on the other hand, has a 50 percent ground ball rate, and he may be able to credit an improved changeup for those results. I will be more reluctant to pick on Vargas for his upcoming starts.

3. START Manuel Margot.

What I wrote: The speedy outfielder will almost certainly benefit from his visit to Atlanta, where Tyler Flowers and Kurt Suzuki share catching duties. In 2016, they threw out a combined total of 15 would-be stealers in 127 attempts.

What Margot produced: 6-for-24 (.250) with one home run, four runs and two RBI, but 0-for-1 in steal attempts.

Takeaways:  In those three games against the Braves, there was only one other stolen base attempt — a successful one by Travis Jankowski. Margot was thrown out by Tyler Flowers, who has thrown out three of seven basestealers so far this season. In 2016, Flowers’ success rate was 3-for-63. This was a case of a bad call resulting from good process. Having stolen base success against Flowers seemed like a sure thing, and both he and Suzuki seem to invite multiple steal attempts, much in the way that Travis d’Arnaud did last year. I’m not sure what to make of Flowers’ early success, but for the time being, I plan on being a little less aggressive in picking on him.

4. START Lucas Duda.

What I wrote: If you are looking for power this week, and you’re a Hosmer or Belt owner, Duda is your home run-swatting alternative. Over the last three seasons, Duda has hit 20 home runs in 375 at-bats against flyball pitchers, and he will have five of them -– Jerad Eickhoff, Vince Velasquez, Wei-Yin Chen, Adam Conley and Dan Straily – – on his schedule. If you’re concerned about Chen and Conley, remember that Duda has had some success in hitting for power against lefties the last couple of seasons (.240 Iso, 16.0 percent HR/FB).

WorldSeries Game 1- Lucas Duda (22885474115) (cropped)

What Duda produced: 6-for-23 (.261) with three home runs. three RBI and three runs.

Takeaways: Neither Hosmer nor Belt hit .200 or had a home run, so this worked out nicely. Duda’s home runs came against Adam Morgan (34 percent ground ball rate), Jeanmar Gomez (career 50 percent ground ball rate, but 32 percent in 2017) and Edison Volquez (career 49 perccent ground ball rate, but 42 percent in 2017). Even in a case where a recommendation was on target, the result was influenced by the unexpected flyball tendencies of Gomez and Volquez. Also, Duda did not homer off any of the five starters who seemed likely probable victims. This early in the season, I still think the best bet is to focus on the tendencies of starters, since that is who the hitters will get the most plate appearances against. However, it is difficult to know (once again) which starters may be improving or declining and what the impact of relievers (e.g., Morgan) might be. This will get easier later in the season.

5. START Francisco Cervelli.

What I wrote: He may not have Rupp’s power, but Cervelli is a superior option this week. He has a .329 batting average against lefties since 2015, and he is scheduled to face four southpaws in Week 2 (Finnegan, Amir Garrett, Eduardo Rodriguez and Jon Lester).

What Cervelli produced: 5-for-22 (.227) with one home run, two doubles, two walks, four RBI and two runs.

Takeaways: Though Cervelli didn’t hit for average, he was still tied for sixth among catchers in CBS fantasy points and better than several more highly-owned alternatives in both points and Roto value. All three extra-base hits came against righties, and he went 1-6  with two walks against the three lefty starters he faced (Cervelli did not start against Rodriguez). His 0-for-3 performance against Garrett really put a damper on his week.

Statistical credits: FanGraphs, Baseball-Reference.

Photo credit: By Arturo Pardavila III from Hoboken, NJ, USA (#WorldSeries Game 1: Lucas Duda) [CC BY 2.0 (http://creativecommons.org/licenses/by/2.0)], via Wikimedia Commons.