Inaccurate Statcast Data

[jgb3]

Edit: regarding an argument I made below, thanks to ataranto for pointing out real life examples of Backfoot/falling away throws seeing a significant reduction in mph. I'm taking back my argument below about the throw, but I still think the Turner HR and Andrus speed points are valid. The rest of this post is as it originally appeared.

The following thread should be a big deal to the OS community, since I know you guys care about having a realistic baseball game void of glaring issues. Apologies for echoing a post that I made in a different thread, but I genuinely feel that it deserves its own forum for discussion.

With regards to the "MLB The Show Adds MLB Network" video posted today on Gamestop, I fee obligated to create a thread about the statcast data shown in that video and how wildly inaccurate and arbitrary it might be based on a few observations.

I don't think that the general public really has a grasp of what the norm for most of this data is yet. They see Elvis Andrus running 18.46 mph and they think, "Wow that's really fast. Cool." As I posted in a different thread, the statcast numbers you saw in those videos is massively incorrect.

Case in point, Tulo unleased an absolute cannon in the GIF below, which clocked in at 78 mph.



Let's stack this up against OF's with the worst arms in baseball, per a Feb. 11th Mike Petriello piece on OF arm strength. Here are the 5 worst OF arms in baseball, measured by the top 5% of their competitive throws from 2016.

Screen Shot 2017-02-27 at 9.18.33 PM.png

And what about Elvis Andrus' 2nd to home top speed of 18.46 mph? 2015 A-Rod has him beat.

Screen Shot 2017-02-27 at 8.21.50 PM.png
Screen Shot 2017-02-27 at 8.18.55 PM.png

There were also massive issues with the Justin Turner home run data (Exit velo & launch angle of that homer make it probably the longest HR in MLB history, given comparable homers.)

Anyway, I hope that SCEA has time to make this fix before the game is released. The casual fan/gamer may not notice or care... but to someone that knows what the data means, this is the equivalent of being told a wallscraper was hit 480 feet.

For the record, I am a huge fan of SCEA and the crazy hours they must have put into this game... but I am also a huge fan of the small amount of statcast data is available to the public right now.

[dran1984]

Re: Inaccurate Statcast Data

Keep in mind, this footage is at least 2 weeks old.

[SilverBullet1929]

Re: Inaccurate Statcast Data

When I saw the Showtrack thing today I loved the visual appearance of it but yes I did ask myself how accurate the numbers were. My first thought was that they'd just throw random fake numbers in there because the numbers don't really matter and they're just for presentation purposes but then I thought maybe they actually can make fair calculations with this because this isn't reality it's all at its core just numbers, codes and programs so maybe it's easier for them to actually come up with fairly realistic numbers. Don't know what it will be but we shall see soon enough.

With that said, I can't possibly take this that seriously and I won't care if the numbers aren't accurate. It's like home run distances in the game, meaningless.

[jgb3]

Re: Inaccurate Statcast Data

Keep in mind, this footage is at least 2 weeks old.

To my knowledge, SCEA can't fix the problem if they don't know it exists or if they don't know people care.

Don't want to take a leap of faith assumption that they have it covered since they have no history with statcast data.

[jgb3]

Re: Inaccurate Statcast Data

When I saw the Showtrack thing today I loved the visual appearance of it but yes I did ask myself how accurate the numbers were. My first thought was that they'd just throw random fake numbers in there because the numbers don't really matter and they're just for presentation purposes but then I thought maybe they actually can make fair calculations with this because this isn't reality it's all at its core just numbers, codes and programs so maybe it's easier for them to actually come up with fairly realistic numbers. Don't know what it will be but we shall see soon enough.

With that said, I can't possibly take this that seriously and I won't care if the numbers aren't accurate. It's like home run distances in the game, meaningless.

It could be fairly easy to make the data a function of a player's speed rating. Set a range of outcomes for a certain rating, and you're done. Same way they measure home run distances.

[dran1984]

Re: Inaccurate Statcast Data

To my knowledge, SCEA can't fix the problem if they don't know it exists or if they don't know people care.

Don't want to take a leap of faith assumption that they have it covered since they have no history with statcast data.

Very true. At this point with something like this, its probably too late to fix at this stage. But like was mentioned in the other thread changes have been made since this footage was taken. Nonetheless, you have a fair point. Nothing wrong with pointing out errors.

[jgb3]

Re: Inaccurate Statcast Data

Very true. At this point with something like this, its probably too late to fix at this stage. But like was mentioned in the other thread changes have been made since this footage was taken. Nonetheless, you have a fair point. Nothing wrong with pointing out errors.

Hope it is patchable if it's arbitrary data- would be a bummer if there was zero correlation between a player's overall ratings and the data.

[MetsFan16]

Re: Inaccurate Statcast Data

Work in Progress. I can't imagine the devs will release a product that hands out 575ft home runs to Justin Turner.

These videos are from an old build, not the final product. I'd be shocked if this wasn't already addressed by the devs team.


Sent from my iPhone using Operation Sports

[WaitTilNextYear]

Re: Inaccurate Statcast Data

It's still important to compare apples to apples.

Directly comparing a loaded-up OF throw velocity with an off-balance INF throw velocity is specious. Directly comparing a guy busting it for a triple and the same guy scoring easily from 2B is specious.

As for the HR distance, that does seem a bit inflated, but we'll need to see some more numbers before concluding off of exactly ONE home run that the data is off.

I think we need to see more data before making any judgments.

[Syce]

Re: Inaccurate Statcast Data

If say the Andrus run is based off of his speed rating for example then with the guys who make the rosters then adjust values based on real life data and maybe boost speed rating up if his speed in game doesn't match his real life speed. Just a thought

Sent from my SM-G930W8 using Operation Sports mobile app

[ataranto2]

Re: Inaccurate Statcast Data

I found this video of a very similar play made by Ian Desmond in comparison to the Tulo play. At the 2:00 mark in the video, it says his throw was 63.8 mph

https://youtu.be/4Bnnek1oaN4

[jada855]

Re: Inaccurate Statcast Data

Got to agree with the OP. I think it also has to do a lot with the animations not being able to reflect closer to actual speed.

[Mitchrapp]

Re: Inaccurate Statcast Data

It's still important to compare apples to apples.

Directly comparing a loaded-up OF throw velocity with an off-balance INF throw velocity is specious. Directly comparing a guy busting it for a triple and the same guy scoring easily from 2B is specious.

As for the HR distance, that does seem a bit inflated, but we'll need to see some more numbers before concluding off of exactly ONE home run that the data is off.

I think we need to see more data before making any judgments.

Exactly, an off balanced throw from a SS is not the same as an outfield throw with a crow hop.

[soxnut1018]

Re: Inaccurate Statcast Data

<iframe width="560" height="315" src="https://www.youtube.com/embed/4ajc_kRiYz0" frameborder="0" allowfullscreen></iframe>

Here's the only homer hit last year with a similar exit velocity to Turner's. And that had a launch angle 11 degrees lower. In fact, only 7 players hit a ball with an exit velocity over 118 mph all of last season.

[jgb3]

Re: Inaccurate Statcast Data

I found this video of a very similar play made by Ian Desmond in comparison to the Tulo play. At the 2:00 mark in the video, it says his throw was 63.8 mph
https://youtu.be/4Bnnek1oaN4

I stand corrected. Now what about that 550 foot moonshot Turner hit, or Andrus topping out at 18.5?