OOTP 27 New (Advanced) Statistics

[Daniel_09]

I know we're in the road to release phase and that we'll likely have a weekly video explaining various aspects of the game, but something I've been reading since Friday (the trailer's release) that someone could explain without going into too much detail is the new advanced stats and their actual usefulness for any type of OOTP player.

Many people say it's purely cosmetic, which wouldn't have any impact on the day-to-day performance of an OOT player. Others have explained that it can be used as a bridge between ratings and outcomes (allowing for justification of performances without focusing solely on BABIP). Still others have suggested that these new statistics can predict short- or medium-term changes in ratings. For example, a player with elite power (70-80/80) whose barrel percentage and hard hit percentage are below what's expected for those ratings might indicate that their power will decrease in the next scouting report. (Although, as a player with almost a decade of OOTP experience, I know that ratings don't drop drastically, for example, from 70 to 50 in a single scouting report.)

I've also read that the only use for these new statistics would be for a league or user playing with low scouting report accuracy, which would allow... The new statistics are a direct interpretation of the direct rating, if you will. However, this assumes that the user intentionally uses a low scouting report when starting a new game.

Could one of the developers explain this a bit before the "Road to Release" videos? It's possible that, due to the complexity of the engine, it might not be a central topic in those videos.

And on the same topic, I'm wondering if all players with the same gap power and home run power (assuming 100% scouting accuracy) would then have the same values ​​reflected in the new statistics?

[Fabtron7]

I'd also love to know if these "new statistics" imply that the engine is calculating things in a different way or if they're simply just flavour added on to the play after the results have been determined?

I'm assuming it's the latter because the former would mean some serious changed to the game-play-engine and I feel like they would have announced that as a feature.

[Rain King]

Here is what Matt said about this in the Announcement thread.

Originally Posted by Matt Arnold

It's not changing to a physics-based engine, no. But there is deeper integration to make sure that ratings and outcomes match. So you will have players like Stanton or Judge have harder hit balls than Arraez. They might both gets singles on a play, but will lean towards hits that better match their profiles.

That definitely doesn't make is sound "Purely Cosmetic" to me.

I do think they will have more usefulness the less accurate/precise your rating options are (because in the end it is the ratings that have the direct impact on the result)...which is essentially the same as any statistics. But, really, we probably need more information before making any blanket statements...and I'd hope this is something that is going to evolve along with OOTP going forward.

[Matt Arnold]

Not sure if we'll have time to detail much more before the road to release videos start... but we'll have a few more details coming VERY soon.

[Daniel_09]

Originally Posted by Matt Arnold

Not sure if we'll have time to detail much more before the road to release videos start... but we'll have a few more details coming VERY soon.

Thanks, Matt, for your reply, any additional information on this is super important, I think it would improve the pre-release (less or more hype about it) of the game. I know that the road to release videos can be focused on other things like the World Baseball Classic, dynamic weather, and PT.

Any new info will be very well received for the community.

[jcard]

Originally Posted by Rain King

Here is what Matt said about this in the Announcement thread.



That definitely doesn't make is sound "Purely Cosmetic" to me.

I do think they will have more usefulness the less accurate/precise your rating options are (because in the end it is the ratings that have the direct impact on the result)...which is essentially the same as any statistics. But, really, we probably need more information before making any blanket statements...and I'd hope this is something that is going to evolve along with OOTP going forward.

Inconclusive, and at best in need of clarification. Here, for example, for what is the hit velocity a proxy—Gap? Power? Height / Weight? The Arraez / Stanton comparison would seem to rule out Contact.

[jcard]

Originally Posted by Matt Arnold

Not sure if we'll have time to detail much more before the road to release videos start... but we'll have a few more details coming VERY soon.

Really? Not enough time to post a few explanatory paragraphs at most on something being promoted as a major new feature? Gamers should act more like investors in terms of demanding information when being asked to commit their money.

[Matt Arnold]

Originally Posted by jcard

Really? Not enough time to post a few explanatory paragraphs at most on something being promoted as a major new feature? Gamers should act more like investors in terms of demanding information when being asked to commit their money.

It's the first road to release video coming today https://www.youtube.com/watch?v=1MwX6YdsWXA

[fredbeene]

with all these fancy flavorful 'stats', why can't i get grand slams, All Star Appearances counts, DP, TPlays, inside park games injured, player of game/week etc.

[fredbeene]

No Bunt Hits? No teams games played in weather conditions.

Interesting to see how file size increases.
Is there options how much to track? You have a similar option already.
Any performance impacts on the game? Simming season time?

[fredbeene]

it will be interesting to see how this works with historical seasons.

[Daniel_09]

Originally Posted by fredbeene

it will be interesting to see how this works with historical seasons.

More than in the historical season, it depends on whether you have recalibration enabled. If so, you don't need to know anything else; you'll know when players started improving and stopped producing, since the ratings are adjusted to reflect real-world statistics each year.

[Daniel_09]

Matt, I've already seen the first Road to Release video. I appreciate that the first one was about this topic; it's a great detail that the advanced statistics have their own tab.

The video raises two questions for me that I hope can be answered since it's online.

First, I see that for P there's a usage percentage for each pitch. Does this mean the engine now has control or a record of each one?, or has this always existed, but now it's visible to the end user? If so, then the advanced stats for P would be much more accurate, for example, %whiff or the run value of each pitch.

The second question is similar regarding advanced stats, but more focused on the meaning of player evaluation.

If I have player A with 80 power and player B with 70 power, both with the same scouting accuracy (which I know for this version each rating has an accuracy, but it hasn't been explained whether it takes longer to reach "very high" or if, with the new system, a rating will never be "very high"), and player A has 81% barrel and 81% hard hit, and player B has 99% barrel and 99% hard hit, which I imagine are two advanced stats that should be related to the power rating.

If this happens, what can be deduced? Will player A's power rating decrease? Does player B have more power than the scout indicates, or is it simply randomness of the engine?

[Daniel_09]

The same applies to expected stats like xera, xba, xwoba, etc. Can these be used as indicators in player evaluation, or is it a direct outcome of the engine that a player X will always have xba greater than ba, meaning they will always have bad luck because the engine decided that?

[Matt Arnold]

Just like real life, sample sizes and luck will matter. Just because a guy has a week of not hitting the ball hard doesn't mean that he's about to crash and burn.

As for the pitching stats, we do track every pitch. For stuff like the run values and x-stats, we do track and count them for each pitch, and add them up. There are some simplifying assumptions with them, sure, but our goal is to map things out.

And as I mentioned, we did spend some time to also make sure pitchers are throwing the right pitches in the right counts, and so on. There was also some of that before, but it was a simpler method.

[OutS|der]

nm I was thinking they were new ratings

[Daniel_09]

Originally Posted by Matt Arnold

Just like real life, sample sizes and luck will matter. Just because a guy has a week of not hitting the ball hard doesn't mean that he's about to crash and burn.

As for the pitching stats, we do track every pitch. For stuff like the run values and x-stats, we do track and count them for each pitch, and add them up. There are some simplifying assumptions with them, sure, but our goal is to map things out.

And as I mentioned, we did spend some time to also make sure pitchers are throwing the right pitches in the right counts, and so on. There was also some of that before, but it was a simpler method.

Excellent, Matt, thanks for your response. With your answers and today's video, I think we have a clearer idea of ​​the usefulness of advanced stats. If the engine already operates at a certain capacity per pitch, adding a strategy slider for pitch usage for the pitcher would force the engine to follow a specific path.

I think it's something worth considering for future versions.

[mars001]

Super excited to see the graphs. I've wanted this for so long. It would be great if you could change the date range/periods. YoY, MoM, YTD for current season, etc...

[BPS]

Say you had two people, A and B, playing OOTP. And they are playing an identical game save at the start of a season. Both are identically skilled OOTP players.

Both repeat this same season 1000 times. (So, "small sample size" isn't an issue).

If A closely scrutinizes the new stats and takes action base on these stats while B ignored all the new stats, will A do better on average for the 1000 seasons (in winning games and developing players) than B?

[luckymann]

One more question, how have historical players been treated in the years before these data were available? Have these ratings been "reverse engineered" for them?