Midseason Analysis

[Rockies GM (Dan)]

A Brief Introduction

This is meant to be both a reflection of current strengths and an analysis of algorithmic accuracy.

For each team, I'll post four graphs - those are (in layout order)

#1 = The preseason projected statistical spider graphs posted here.       
#2 = A current spider-graph of accumulated data, along with Z-Score.
#3 = I ran the program using only the stats accumulated to this point, and simulated the first twelve weeks. This is the distribution.
#4 = Future strengths spider-graphs, along with corresponding win% generated by my algorithm (based on historical stat win% and Z-Score correlation)


NL West

Biggest difference between preseason projection and what Arizona has done to date has been in the power category (HRs and OPS). They've had the good luck to benefit from multiple power breakthroughs this year. They added 8 HRs in Fowler after the projection was posted, and Bregman, Chapman, Almora and Pillar all are on pace to set new career-highs in HRs - Bogaerts is ahead of pace as well, though barely.

They also beat projections in the wins category - they've gotten 11 from rookies Mike Soroka and Jordan Yamamoto, along with 8 bullpen wins.

The Future

Projections see the HR outburst of Arizona's stars diminishing, or at least regressing to a point, although still well above what was originally projected. We also see something that will begin a trend - historically, rate stats (ERA, WHIP, AVG, and OPS) are the most unreliable. Even the best teams in these stats often don't win the category more than 80% of the time. While Arizona's elite pitching staff gives them a leg up above the rest of the league in ERA and WHIP, a bad week can easily see them lose those categories. Ks, Saves, and SBs are among the most reliable categories, as we'll see in a few teams.

The Rockies have recently gone through an extremely rough stretch with their young pitching, pushing down their ERA and WHIP. Otherwise, projections have held up to be true. More than computer/algorithm error on preseason projections, I'm willing to chalk this up to simply projections being wrong - most didn't see Foltynewicz's complete breakdown nor Archie Bradley and Wade Davis being shells of last year.

The Future

While future projections see some ERA and WHIP correction, basically, these Rockies are what they are. While the Rockies aren't elite in any one category, they should win nine categories at above a 60% clip, topping with an almost 90% category win rate for strikeouts. Overall, the team isn't as good as the last few years, but the program thinks they're better than a normal 8-4 team.

Offensively, apart from AVG, the projections pretty much were head-on. Again, preseason projections are notoriously bad at gauging rate stats, so I'll take it. The Dodgers also made a number of in-season moves that improved their offense, so it's remarkable that the projections are that accurate.

Pitching wise, the projections were off on K, ERA, and WHIP. Projections failed to see the breakout of Giolito, and a number of Dodgers pitchers have beaten projections the first half of the year (while LA has also had the good sense to sit an unexpectedly flailing Kyle Freeland for some of his worse starts)

The Future

ROS projections think Los Angeles is what we think they are at this point, maybe with a little ERA/WHIP regression, but nothing severe. The biggest flaw with this year's Dodgers team is that they are overly reliant on rate stats - good, but not great in AVG, OPS, ERA, and WHIP. Considering those are the most unreliable stats week in and week out, the Dodgers have not left themselves much wiggle room.

Our first team whose projection spider makes you squint really closely in order to tell if anything matches in what has happened to this point. The makeup of the team was substantially altered with the losses of Salvador Perez, Mark Trumbo, Devon Travis, Wilmer Flores, Taijuan Walker, Keone Kela, and more. Quite simply, this team isn't the same team analyzed in April. And yet, they are often beating offensive projections off the back of productive first halves from Jay Bruce, Trey Mancini and Paul DeJong

The Future

IL returnees will help San Diego in WHIP, ERA, and K. Unfortunately, the Padres face an uphill climb to make the last wildcard spot. Will this help be enough?

Really, the biggest discrepancy between projections and reality is the saves stat. Hader has been a closing machine this year, and maybe the projections didn't take the Ian Kennedy: Closer experiment seriously enough. Meanwhile, the Giants are beating their HR projection on the backs of Gary Sanchez returning to form and a Josh Bell power breakout.

The Future

Apart from some pitching rate stat regression, the computer doesn't see much change for SF. With a team strong in both rate stats and counting stats, the Giants are a formidable foe, guaranteeing at least a win in K and SB almost 90% of the time for each.

[Rockies GM (Dan)]

NL Central

The Cubs have suffered from a power outage, with a number of players underperforming, most notably, Matt Carpenter. Adalberto Mondesi hasn't quite had the breakout that many expected of him. On the pitching side, while they haven't been bad, they don't have any real standouts. Stroman is the closest thing they have to an ace, and, as such, they've come in under preseason projections.

The Future

Chicago may get some power rebound in the second half, according to ROS projections. Beau has smartly built this Chicago team - while rate stats, the more unreliable stats week-to-week are Chicago's weakness, many of the counting stats are its strengths. Chicago will win their share of rate categories - hovering between 40-55% win rates in each category, but high percentage rates in SB, K, and SV make Chicago difficult to beat on a consistent basis.

Daniel Vogelbach finally breaking through, Yandy Diaz already having eleven times his career HR output, Jesse Winker nearly doubling his, Eduardo Escobar nearing career high HR, and an amazing start to Austin Riley's career have completely transformed this Reds offense. No projection could have predicted all of those breakouts. Their pitching, however - apart from holds - has held this team back, underperforming even modest projections.

The Future

The computer doesn't completely buy the power surge in Cincinnati. But, even so, the Reds are mostly projected to be a middle of the road team.

Another team, another whose biggest fluctuations are in rate stats. I came into this exercise half-convinced I'd need to change some of the formulae I use to process the data from projections, and leaving more bolstered that the things I get wrong are the things which should be expected to be incorrect, for the most part. While the Brewers have had a number of players outperform projections, most of them are not of the power variety, hence the improvement in R, RBI and AVG, but not real changes in HR and OPS (Had Judge not been injured, that may have differed). While Yonder Alonso has floundered, Realmuto has upped his R production being in a Phillies lineup vastly better than his old Marlins, Urshela has shown up, Sogard is putting up career-bests and Choi and Fletcher are performing above expectations. Meanwhile, their pitching has been the weakness of the team thus far.

The Future

Again, a team which mitigates its weaknesses by virtue of the unpredictability of rate stats. even being in the bottom 20% of the league in WHIP and ERA should result in the Brewers winning those categories about 40% of the time. The Brewers, despite a weak pitching staff, are still in the wild card hunt.

Moustakas is the biggest breakout here, looking to possibly have a career year at age 30. Ozuna, Maybin, Soler and Flowers have also had first seasons beyond that of what was expected. Cuthbert has also looked highly improved, albeit in a SSS. On the pitching side, Pittsburgh is mostly what we expected, if not a little better.

The Future

This looks much more like the original projection than what has happened to-date. The gains made in the SB category look more like a speed deficit by the rest of the league than any truly big difference in sustainable production, and the R production may fall off a bit as well. Still, the Pirates have a good shot at the NL Central, and have all but wrapped up a wild card berth as well.

A caveat: The second spider graph is taken from stats from the Fantrax "Season Stats" page, which counts ERA and WHIP even if the team hasn't reliably met weekly minimums. Thus, the difference in ERA and WHIP. In reality, the projection graph is more correct. I think Luke Jackson becoming a closer explains the switch in saves and holds seen between the two graphs.

The Future

The program believes that St. Louis has enough pitching between what it has currently and potential graduations to clear the minimum (calculated by projected remaining total innings/remaining weeks), so I'll let it be. St. Louis has some pitchers on their bench accumulating innings, so I can see the algorithm's point. One thing I may look into changing is demonstrated by the program giving St. Louis even a ~30%-40% chance at AVG and OPS. The only problem is that I can't simply assign teams that don't meet the minimums a 0% chance, as that could mess with some of the equations if a 0% team met another 0% team.

[Rockies GM (Dan)]

NL East

Mostly, the Braves match what was expected of them pre-season. However, Gant has a chance to run with the closer role in St. Louis after Hicks' UCL tear. While that isn't factored into the graph below (apart from the 12-week sim, all of the data is from last week...this took time to put together), it will have ramifications, both positive and negative. Despite underperforming expectations with only 5 wins to date, the Braves are still tentative favorites at the division

The Future

The Z-Rates are still similar to what is seen above. The algorithm already liked them to improve their Save ranking, and now with Gant becoming the full-time closer, they should gain even more saves at the expense of holds. The Braves have had some bad luck yet, but they are a dangerous team.

Again, biggest differences are in WHIP and AVG, rate stats. They've gotten good seasons thus far from relievers, led by Adam Ottavino's 18 holds. While they have had a great season thus far, the Marlins have come in undoubtedly on the high end of what could be reasonably expected from a team with these statistics with their four wins..

The Future

Projections see some of their holds turning into saves. The winning steak doesn't look likely to last - while Miami's win percentage looks much better overall than their Z-Scores, their highest single stat results in about a 60% win rate (Holds). Nearly everything else is around 50% or worse.

Regression? What regression? While every single projection system was calling for some regression from Christian Yelich, he's arguably having an even better season this year. The Mets have also benefited from breakouts from Tatis, Kepler, Narvaez and Clint Frazier. Their pitching was predicted to be good, but Buehler, Chirinos and Cole have pushed their WHIP to unexpected heights.

The Future

Largely, the algorithm buys the breakouts we've seen and the results to this point. The Mets have put themselves in excellent position to take advantage of a weak NL East and win the division seemingly out of nowhere.

Not a ton of power here, but Ian Desmond having a good season for once, Tommy La Stella and Tom Murphy breaking out has raised their AVG and OPS a bit. The ERA/WHIP split is intriguing here, and probably due to a good season from the bullpen.

The Future

The projections don't really buy the strides made by the Philly offense. Similarly, it sees the WHIP coming back down to earth. The Phillies have a shot, given the NL East being wide open this year, but they are more of a dark horse.

Mookie Betts was transcendent last year. This year, he's merely good. Guzman was expected to take a step forward in the power department, and he hasn't. Brendan Rodgers has really struggled. Benintendi hasn't shown improvement. Still, even with the lowered power output, this isn't the look of team that should be 4-7.

If you look at the simulation, the Nats are below a 1 percentile outcome at 4-8. So...why? It's tough to say. The Nationals have gotten some terrible, terrible luck this year for sure - hitting opponents at their absolute best. But some of the bad luck may be explained by team makeup. The strengths of the Nationals are largely in rate stats - the most inconsistent week-to-week.

The Future

This can be taken with a grain of salt with the Nats going into full-on firesale. But, again, this isn't the graph of a normal 4-8 team. But, again, they may be above 90% normalized Z-Score in ERA and WHIP, but that only results in ~60% win rate in those stats. They're also throwing away saves every week, so they effectively have to win 6.5 categories out of 11 rather than 12. It doesn't fully explain the tailspin, but perhaps a bit.

[Former Angels GM (Mike C.)]

so I can see the algorithm's point. You talking with the algorithm Dan? (will robinson for the old folk) A old joke but amazingly a ton of information and work. Sure seems like its getting dialed in.



Question on Arizona D-Backs since they are first listed, maybe changed the most in projections since pre season and really no significant change in team since pre season. It just seems like they have gone from 0.01 to 49 percent chances and grow every new listing.

HR are up for many across the league, not really exclusive there, 8 pen wins, it isn't out of sorts but definately 11 rookie wins would be a huge point of difference. Is That really enough to drastically change them or is it just more of a benifit from other NL teams falling?

[Rockies GM (Dan)]

Jun 25, 2019 13:25:04 GMT -8 Former Angels GM (Mike C.) said:

so I can see the algorithm's point. You talking with the algorithm Dan? (will robinson for the old folk) A old joke but amazingly a ton of information and work. Sure seems like its getting dialed in.



Question on Arizona D-Backs since they are first listed, maybe changed the most in projections since pre season and really no significant change in team since pre season. It just seems like they have gone from 0.01 to 49 percent chances and grow every new listing.

HR are up for many across the league, not really exclusive there, 8 pen wins, it isn't out of sorts but definately 11 rookie wins would be a huge point of difference. Is That really enough to drastically change them or is it just more of a benifit from other NL teams falling?

It's a combination of both, I think. San Diego got hobbled early, who were an early favorite for a wild-card, the NL East hasn't been strong this year, etc.

While it's true to say that HRs are up across the board, I don't think projections foresaw Bregman (whose career-high was 31) being on pace for over 40 HRs, or Matt Chapman, who hit 24 last year, being on pace for nearly 40 himself. In addition, NO ONE saw James McCann suddenly hitting like he was in-his-prime Brian McCann. Keston Hiura also hit extremely well in his debut, and Bogaerts is having a career year. A lot of things are falling right so far for the Diamondbacks.

If I'm not mistaken, once I switched over to Z-Score and fixed the problem with the teams not expecting to meet minimums dragging down the top teams, they were one of the biggest beneficiaries. The algorithm always thought they'd have a great pitching staff, at least by ERA and WHIP. They're proving that isn't all they are.

[Rockies GM (Dan)]

AL West

What happened to the Astros offense? While Gleyber Torres has broken out in a big way, almost everyone else (Belt, Kike Hernandez, Hicks, Morales, Bauers, Piscotty) either have regressed significantly or not met the expectations of predictions, and thusly, the Astros offense has cratered. The bullpen, as expected, is the strength of this team - but the team was expected to have more strengths than that coming into the season.

The Future

The projection systems aren't buying strong positive regressions from the 'Stros, and thus the Z-Score projections strongly resemble what they currently are. Good news is that their stat category win% give them a little bit more of a fighting chance. Still, most hover around 40%, and even in a weak AL, that probably won't be enough. They've fallen behind Seattle and Oakland in their own division, all of whom are looking up at the next team on our list.

Happy to say that this is a breakout the algorithm predicted before almost anyone else (other than Mike himself). but the Angels have even outperformed the projections. They've come in on the high end of their win spread, and smart managing of their pitching staff has them outperforming WHIP predictions, even with two starters above a 7.00 ERA. Mets rookie phenom has been outperforming any reasonable expectations, putting up an MVP-caliber season in his rookie year thus far. Muncy is proving last year wasn't a fluke, Kendrick is having an unexpected career year at the ripe old age of 35, and Blackmon is back to his 2017 heights

The Future

The projections see the Angels trading in some saves for holds, and sees some (but not much) offensive regression, but this is still an extremely strong Angels team. They should be in the running for a pennant, if not more, come September, especially if they acquire some speed at the deadline and fix their one real weakness.

Wow, did I get this one wrong. Looking at it, the offense hasn't been BAD per say, but it certainly isn't meeting expectations. Robinson Cano has been a shell of his former self in a Mets uniform and McCutchen got injured early in what looked like a big bounce back season. Apart from that, they don't really have anyone who has been struck terribly by regression or injury, but the projections may have been looking for bounce back years from guys like Abreu and others, but it hasn't happened. Others, like Jose Martinez, have underperformed, and except for recent call-up Yordan Alvarez, they just haven't had offensive breakouts to counteract what they've lost. On the plus side, on the backs of a strong bullpen, and now-suspended PEDs-user Frankie Montas, their pitching staff has been moderately more impactful than anticipated.

The Future

The projections think Oakland may yet have a good second half offensively. But, perhaps because of the Montas suspension, it sees the pitching categories regressing, as well. Ultimately, the A's have put themselves in fighting contention for a Wild Card spot, even if they aren't as strong as anticipated.

And here we encounter the issue that's my biggest to-do in making the program even more accurate. The Mariners have a strong offense, but six of their strongest option are all CI-eligible, with the only multi-position player out of those players being Yoan Moncada. The Mariners can't play more than three of the rest of their five CI on any given day. I haven't yet found a great solution for this, but I have several ideas. That's probably the biggest reason you see the discrepancy in R and HR. They've also swapped saves and holds, which shouldn't come as a great surprise, either (apart from Cody Allen's being ineffectual). Still, while I may have been early in my bold prediction that the Mariners would take the AL West, young  studs like Moncada and Soto have the Mariners in good shape.

The Future

The projections don't believe that the holds will (pardon the pun) hold up. But, whether it's a product of the multiple CIs or not, it expects positive offensive regression. The Mariners still have a good shot at a wild card, helped, Despite the 4th place interleague schedule being stronger than expected with the rise of Pittsburgh and Arizona.

Our first spider graph to come with a caveat - the one on the left is out of 100%. The one on the right only goes up to 25%, so it will look disproportionately large. Despite saves looking like a spike, they've traded away what little of them they've had (2 of their grand total of 3). In fact, they've traded away any semblance they've had of a pitching staff. The only stat they're almost even decent at is steals, and even then, they are still in the bottom 25th percentile.

The Future

The Rangers are strong underdogs in any game they play - this graph only goes up to 40%. Their offense may show some marginal improvements with rookies like Cavan Biggio. And by that, I mean, they may actually meet minimums.

[Rockies GM (Dan)]

AL Central

Mostly the big changes have occurred in the rate stat categories. The offense has vastly improved, buoyed by breakouts from Yasmani Grandal and Ketel Marte (especially the latter, who is two HRs away from doubling his CAREER output). Aside from holds, their pitching is merely good, not great. But if the hitting holds up, that's what will make the difference for Chicago.

The Future

While there may be some regression for the offense, the hitting should remain at least 80th-90th percentile in the league. At the moment, the AL Central is theirs to lose.

Cleveland's offense has taken some steps backwards this year. Goldschmidt hasn't looked like himself this year, with an OPS over .150 below what he hit last year. Zobrist has absolutely cratered, and Maikel Franco hasn't shows that last year's strides were real. Depth like Altherr and Nick Williams have been thoroughly ineffectual. In addition, offseason acquisition Gregory Polanco has been injured most of the season. The offense has been saved by strong fist halves by Willson Contreras and Javier Baez. Similarly, the pitching has been good, although not great, as was expected. A lot of this falls on having four starters over a 4.30 ERA, including Aaron Nola, who hasn't been the ace he was expected to be.

The Future

The Indians may be in line for positive gains, especially in offensive categories, where Goldschmidt et al are projected to have better second halves. It also appears that the projections are betting that some of those Red Sox holds are going to be converted into saves as the situation gets sorted there.

The Tigers have been just a little worse offensively than what was expected - though, as predicted, the strength of the offense lies in the steals generated in their MI tandem of Hamilton and Anderson. The pitching, also has been weaker than expected. No projections foresaw Trevor Rosenthal being unable to record an out - a 22+ ERA and WHIP just short of 4 means he's been unemployed since early in the season, and Arodys Vizcaino has been injured as well.

The Future

The Tigers are what they are. I really don't know what else to say here. They win almost ever category at only a 50%-40% clip. Even their biggest strength, steals, only wins them the category about 60% of the time.

Again, caveat here - the graph on the right only goes up to 60%. And you still see almost everything below the 15th percentile. The Royals are going to be in position for a lottery pick.

The Future

The Royals don't win any category over a 50% clip. Can't win games when your second best category only gets you wins 30% of the time.

And the category for team breakout the projections didn't see coming goes to Minnesota. Wow. Santana has enjoyed moving back to Cleveland, with an OPS over .950 after barely getting over .750 last year. LeMahieu is the odd ex-Rockies player who actually improved away from Coors, although Yankees Stadium is a hitters park in its own right. Jorge Polanco is finally breaking out. Luke Voit is proving that last year's performance in NY was more than a SSS mirage. And, maybe most importantly, Buxton is FINALLY showing why he was such a highly touted prospect. In the pitching staff, Martin Perez has shown up this year after a terrible year in 2018. A surprisingly strong bullpen is keeping their mediocre starting pitching afloat.

The Future

While projections see some fall-off from the lofty offensive output, this Twins team was severely underrated in the preseason (through really no fault of the program - the Twins are just outhitting projections). As such, while they won't win the AL Central, they're firmly entrenched in the wildcard race.

[Rockies GM (Dan)]

AL East

About what was expected, other than the two offensive rate stats and pitching wins. They're getting a good year from catcher Christian Vazquez, who is batting over eighty points better than last season, which is offset by a decline in AVG by Jesus Aguilar (albeit projections preseason did think a regression was coming in that department). The really big differences come from breakouts by Scott Kingery, Austin Meadows and Lourdes Gurriel, all of whom are showing good to great progression in their sophomore seasons. An injury to Matt Moore and Dan Straily's ineffectualness have lowered expectations on wins. Although Lance Lynn is doing his best to make up for their lack, with ten wins by himself.

The Future

Not much great here. The average is expected to keep up, with some power regression. Good news is, behind those three sophomores, watch out for Baltimore in the coming years. 

While the AVG and OPS are a little less than predicted, the Red Sox team is who they were projected to be. Again, important to note that the right spider graph only goes up to 60%. Tim Beckham and Avisail Garcia have cooled off after hot starts. Garrett Cooper is hitting better than expected while Ryan Zimmerman has seemingly fallen off a cliff, although injuries are surely playing a part. Not much to say about the pitching staff. Merrill Kelly is having a decent rookie season after coming back to the states.

The Future

This graph is out of 50%, so regression is expected to be on the way for the offensive categories. Boston should again be well in line for a high draft pick.

Largely what was expected, minus some strikeouts and with rate stats smaller than expected. A lot of that offensive regression probably falls on the shoulders of Buster Posey showing signs of age, Christin Stewart disappointing, and Asdrubal Cabrera regressing. Injuries to Matt Harvey and Dellin Betances help explain the disparity between strikeouts. On the positive side, Michael Chavis has had a promising rookie season, as has Dakota Hudson. Ohtani has continued being a force at the plate, and hopefully will return to being a star on the mound next season.

The Future

Betances's return should result in some extra holds in the second half. Overall, an average team. They're in the wild card race, but would need a lot of things to fall their way to get there.

A little more RBI and better rate stats, but largely the team that was expected to come in second in the division. Now the overwhelming favorites, due to circumstances we'll cover with the next team. Believe it or not, the Rays have actually gotten unlucky this year in terms of results. The offense has been above average, behind excellent seasons from Rizzo and Springer. Resurgent seasons from Ken Giles and Shane Green have sparked their bullpen. While their starters haven't put up the best seasons in ERA and WHIP, they've won a ton of games, headlined by Odorizzi's ten wins.

The Future

Those wins aren't likely to continue at the same rate, as they aren't necessarily supported by the ERA of their starters (Eduardo Rodriguez has 8 wins and a 4.79 ERA; Senzatela is at 7 and 4.83, respectively, with a 1.622 WHIP). Stilll, even with regression, the Rays have put themselves in a good position for a long postseason run.

Throw out the preseason projections. These Blue Jays haven't been the same since blowing things up after a disappointing start to the season. The Jays, already lacking depth, but supposedly high in star power, had an absolute power outage the first month of the season. Altuve has been on a decline for the past two seasons. Before the Blue Jays blew it all up, the algorithm still saw them as slight favorites for the division, and still one of the strongest projected teams in the AL. Instead, the Jays sold. Since being traded to the Dodgers at his lowest value, Corey Seager has put up a .400 average and 1.189 OPS. Jose Ramirez, however, has continued having a perplexing, free-falling year since also being traded to the Dodgers (at least for Ramirez, the Blue Jays got a future cornerstone in Wander Franco). Coming off of two seasons in which Ramirez hit a minimum of .270 with a .939 OPS, he's absolutely cratered into a .214/.634 line this year.

Enough about ex-Blue Jays. Francisco Mejia hasn't shown why he was considered the rare catching prospect to be truly excited about, and Harrison Bader and Ramon Laureano have regressed. Even a good season from Domingo Santana and a strong start to Willie Calhoun's career haven't been enough to save this offense. The pitching, on the other hand, has remained strong behind Berrios and deGrom. Even so, they haven't quite been as good as expected, as Chris Archer and Raisel Iglesias have joined the disappointment bus from Toronto.

AND YET. Even with everything the Jays have gone through, they've gotten unlucky. With the stats they've put up, a team with average luck would have around seven wins and still be in contention for a wild card. Instead, they, like the Nationals are under a 10% outcome.

The Future

Pitching remains strong, hitting remains the weakness. Just like the Golden State Warriors in the NBA, it appears an era is ending in PB's Baseball.