Yordan Alvarez

[raslavens]

Please explain/defend why he's still in the low 70s ratings-wise. Other lesser prospects have soared up the ranks with far lesser performances. I know it's a game and I enjoying playing it as such (with all its limitations and quirks), but the way he's been treated bears no relationship to reality.

[Maxfire5]

Originally Posted by raslavens

Please explain/defend why he's still in the low 70s ratings-wise. Other lesser prospects have soared up the ranks with far lesser performances. I know it's a game and I enjoying playing it as such (with all its limitations and quirks), but the way he's been treated bears no relationship to reality.

ZIPS doesn't favor him as a rookie. He was projected for 1 WAR this season, 1.5 next, and 1.7 in 2021. Averages of his 3 year projection for the start of the season had him at a wOBA in the mid .330s, 28% k rate, 9% BB rate.

For this season alone, he was projected at a .255/.315/.440 slash line with a 27.5% K rate, and 8% walk rate with a .322 wOBA. Also subpar defensively too which hurts ratings.

[raslavens]

The fact that ZIPs so dominates his ratings despite his amazing real-life performance is a serious flaw in the game or the devs' judgment. How did ZIPs rate Treinen or Doolittle? Probably pretty good -- but then their real-life stats cratered and so did their PT ratings. Not a sufficient answer after several months of excellence from Alvarez.

[Maxfire5]

Originally Posted by raslavens

The fact that ZIPs so dominates his ratings despite his amazing real-life performance is a serious flaw in the game or the devs' judgment. How did ZIPs rate Treinen or Doolittle? Probably pretty good -- but then their real-life stats cratered and so did their PT ratings. Not a sufficient answer after several months of excellence from Alvarez.

Blake and Doolittle were rated highly. Relief pitchers have swung way more than any players in the game. Will smith dropped from a 95 to an 85 in 1 week because he gave up 2 hrs.

You're comparing apples to oranges though. Relief pitchers, and pitchers in general are struggling to be projected by ZIPS. We're at all time highs for HRs. Blake has years of data, and coming off a 1.82 FIP season was projected to do well, but not quite as well as 2018. So when he quickly reached a 5.15 FIP his ratings changed accordingly. He was giving up HRs at twice the rate in his career, walking nearly twice the rate in his career, and striking out at a normal clip.

Therefore his movement and control were heavily impacted by the egregious jump in HRs and walks.

Hitters are less likely to make jumps, one way or another, from what I've seen. Maybe it's a function of what the calculations look at. I'm not sure. Just looking at the data and what the devs use for their calcs based on numerous dev responses to other posts about "why isnt player X higher". They mention projections quite heavily. They do indicate they're always looking to improve their calcs, and hope to see some balancing for live players next year. I think the current year stats don't weigh heavily enough on the season but Kris Jardine has indicated the rate of reliance does change as the season goes on.

It's the first full season of PT so it's a work in progress!

[CrazyWR]

the other part of this is that his non-hitting ratings are static, so there is no possibility to gain ground in that regard. However, it was explained that over time the projections play less and less a part until at the end of hte season only in-season stats were relevant for ratings.

[OMGPuppies]

What hurts him too is that he is a DH. Defensive ratings play a big part in overall rating.

[dkgo]

Without seeing the stats on the card, Alvarez doesn't play defense or run which are components in overall rating.

He has an obscenely high .372 babip

There is also the fact that the absurd league numbers make every individual hitter less impressive by comparison. There are going to be over 1,000 more homers hit this year compared to last year. That is ****ing nuts.

I would expect it to be higher than low silver but not by much for a bat only player in a year where everyone has a bat.

[Matt Arnold]

Yeah, even a guy like Nelson Cruz has virtually peaked as a mid-to-high Gold card despite his great year, so Alvarez with half the PA for a full season simply has trouble getting too high in ratings.

We'll certainly take things under advisement for next year. I also don't like how slow he was to get moving in ratings, and would agree that given how much he's been crushing the ball, he should be a lot higher. We'll certainly solicit some more feedback for next year, but expect some things to change in how we handle things going forward.

[steamrollers]

We all know their are current flaws in how the system works with live ratings. Clearly in-season stats don't have enough of an impact in some situations and adjustments need to be made. It was pointed out that hitters are less likely to make jumps, true in some cases but not all. We have Jack Flaherty putting up historic second half numbers only seen a few times in the history of MLB and carrying the second best WHIP in the NL for the entire season behind on Perfect DeGrom yet Flaherty is basically never moving up and isn't even flirting with gold card status. The guy is going to get plenty of Cy Young votes and will be basically a useless mid tier silver card unless you are in the absolute lowest levels of the game.

[HRBaker]

Without getting into mathematical details, the problem with the current rating system is that it favors its own expectation over what is actually happening on the field. However the Devs choose to adjust that, they need to keep in mind that field performance is what's "expected" from LIVE cards - not some mathematical result that degrades performance because of some complex calculation.

[jeffw3000]

Originally Posted by HRBaker

Without getting into mathematical details, the problem with the current rating system is that it favors its own expectation over what is actually happening on the field. However the Devs choose to adjust that, they need to keep in mind that field performance is what's "expected" from LIVE cards - not some mathematical result that degrades performance because of some complex calculation.

At the same time you cannot just throw the predictions out the window, because someone has a great month or 2, especially when they are somewhat based on past performance. The ratings are based on ZIPs for a reason. in 1979 Roy Smalley was leading the AL in hitting the first half of the season. he had an OPS of .959, and was hitting .341. He finished the season with a .271/.353/.441 clip the rest of the way. Pretty close to his typical numbers. he hit .185/.262/.327 the rest of the way. this is an extreme, but the point is that you have to be careful to over react to the numbers during the season. I am not saying Alvarez should not be rated better. just saying that you should not react to deeply to a player's performance during the season, because the reality is, while the cards are live cards they are live cards for the full season which is not completed, and we do not know how accurate or inaccurate ZIPS ratings are until the season is completed or near completed.

[DotDash]

The algorithm used is a big mystery to me. Consider the following three pitchers and their recent performances:

Cal Quantrill, 66 rating before this week.
Last outing (9/21): 5.0 IP, 2H, 1ER, 0BB, 6K
Last 7: 1-5, 9.27 ERA, 1.64 WHIP

Yu Darvish, 71 rating before this week.
Last outing (9/22): 8.1 IP, 7H, 3ER, 0BB, 12K
Last 7: 2-2, 2.70 ERA, 0.84 WHIP

Homer Bailey, 67 rating before this week.
Last outing (9/18): 7.0 IP, 3H, 0ER, 1BB, 11K
Last 7: 4-0, 1.88 ERA, 0.91 WHIP

I expected Quantrill to keep the same rating or decrease slightly due to continued mediocre performance. I expected Darvish to maybe increase by a point or two. For Bailey, I expected an easy increase to the low 70's; he has performed very well for a prolonged period of time, and had a great outing last week.

These were their actual ratings changes:
Quantrill: no change. OK, no argument there.
Darvish: +3. More than I expected, but I don't know, maybe strikeouts factor heavily into the calculations?
Bailey....: -6 Lower rated than Darvish to begin with, with a very similar last outing (except less hits against and zero runs vs. 3 for Darvish), and somehow he suffers the 5th largest ratings decrease among 1411 live players, while Darvish gets an increase...

I'm also in agreement that Alvarez is rising too slowly. If he had enough AB's to qualify he would lead MLB in OPS, just ahead of Yelich and Trout. Due to the lack of defense, a rating just below J.D. Martinez (say 78-81) would seem appropriate to me. He is currently better than Martinez, but has less of a track record. 73 is low.

[chazzycat]

why does everyone keep mentioning WHIP as if it has anything to do with anything?

[Maxfire5]

Originally Posted by chazzycat

why does everyone keep mentioning WHIP as if it has anything to do with anything?

It's the same people to be fair.

Someone mentioned that Yordan would be ahead/near trout/yelich for OPS. Someone else mentioned sample size (half a season). Funnily enough, Hunter Dozier had an OPS over 1.000 as late as game 53. At that point, he was second in OPS in the majors, beyond Trout. He was still a bronze around 63. He has since regressed as ZIPS projected him to do. Alvarez has 82 games played, an incredibly small sample size. Regression is expected.

Notwithstanding, OPS isn't a metric they likely use either in their calculations but I could be wrong.

Further, for Jack Flaherty his FIP is 3.63 with a career 3.83 FIP. His era of 2.96 isn't likely sustainable as indicated by a career low BABIP of .250, and is very low considering league average is normally .300. His xFIP is 3.72 for the season too. This all points to significant regression for Flaherty.


Edit: FIP wise, he's the 17th best pitcher in baseball.

[dancariaz]

The problem starts with using ZiPS projections for ratings in a video game. That's just not going to work out ever. Imo they should use past performance as a baseline (you know the old saying: past production is the best projection) and give minor leaguers some kind of penalty (20% rating decrease?), so cards start with a more realistic ratings.

[Maxfire5]

Originally Posted by dancariaz

The problem starts with using ZiPS projections for ratings in a video game. That's just not going to work out ever. Imo they should use past performance as a baseline (you know the old saying: past production is the best projection) and give minor leaguers some kind of penalty (20% rating decrease?), so cards start with a more realistic ratings.

Isn't ZiPS based on past production? Weighted for the last 4 years of production relative to an age curve? That helps drive future projections then.

[chazzycat]

Originally Posted by dancariaz

The problem starts with using ZiPS projections for ratings in a video game. That's just not going to work out ever..

Disagree. Using projections is the ONLY way to do it fairly.

Now there may be some tweaks needed on the edges, with regards to rookies in particular. Alvarez being worse rated than Guerrero Jr. at this point in time is clearly incorrect. But in general, it's a good system.

[dancariaz]

Originally Posted by Maxfire5

Isn't ZiPS based on past production? Weighted for the last 4 years of production relative to an age curve? That helps drive future projections then.

they're using past production for their projections but looking at them you wouldn't believe it. If you look at the top players for example ... ZiPS projects everybody to have a down year at the beginning of the season every single year
And I won't even get into ZiPS regularly swinging and missing when it comes to projecting prospects.

Also, I may be wrong about this but I'm pretty sure I read somewhere on Fangraphs last year that the projections done by fans are the most accurate projections out of all the systems they use on their site.

now to my point though:
imo ratings in a video game should be based on recent performance, ideally the last 600-700 PAs from for batters for example. No need to include stats from 2-4 years ago. Ideally the starting point for a 2020 live card is their 2019 performance, right? Why do I need projections? Live updates will (or at least should) make sure that the player will be rated properly at all times, no need to project anything.

[Dogberry99]

When the season is over, will Live cards be updated to reflect actual performance for the non-changing ratings?

Asking for a friend who totally isn't Mike Minor...

[Matt Arnold]

Originally Posted by dancariaz

they're using past production for their projections but looking at them you wouldn't believe it. If you look at the top players for example ... ZiPS projects everybody to have a down year at the beginning of the season every single year
And I won't even get into ZiPS regularly swinging and missing when it comes to projecting prospects.

Also, I may be wrong about this but I'm pretty sure I read somewhere on Fangraphs last year that the projections done by fans are the most accurate projections out of all the systems they use on their site.

now to my point though:
imo ratings in a video game should be based on recent performance, ideally the last 600-700 PAs from for batters for example. No need to include stats from 2-4 years ago. Ideally the starting point for a 2020 live card is their 2019 performance, right? Why do I need projections? Live updates will (or at least should) make sure that the player will be rated properly at all times, no need to project anything.

I mean, for everyone, you need to start somewhere. While it may seem like Zips tends to be pessimistic on a lot of players, time and again it's proven to be one of the better projection systems out there. Basing initial ratings next year on 2019 performance is going to cause too many guys like Kimbrel or Treinen or others to start out way too lot, while other guys will start way too high coming off a career year.

And it's not just the extreme cases like those relievers that got bombed. You get guys like Jose Ramirez or Lorenzo Cain who were just down this season, but have a solid chance of bouncing back. Maybe not to the level they were before, but likely higher than they ended up this year. And that's what those projection systems account for.

I mean, don't get me wrong, Zips certainly isn't a perfect system, ideally we'd probably want to take the top 3-5 projection systems and weight them in a way that we approve of, and it would create a better system overall. But whether that actually makes a better system than simply starting with one system and doing some manual adjustments before the season begins, hard to say.

Originally Posted by Dogberry99

When the season is over, will Live cards be updated to reflect actual performance for the non-changing ratings?

Asking for a friend who totally isn't Mike Minor...

No, I don't believe we have a plan to adjust any of the peripheral ratings. It's something we may consider down the road for future live update versions, if there's other data that we get that we could use to do smaller tweaks to stuff like stamina, holds, speed, or defense throughout the season.

And as we've said, we'll take in everything we've learned for this year and figure out what changes are needed for next year. Please do keep *nicely* telling us what stuff you did like or didn't like from this year's version and we'll take it into consideration.