I spend far too much time looking at baseball statistics on the internet. In doing so I meet so many fascinating statistics that are outside the realms of baseball reference.
One though has stuck with me, and despite being created over half a year ago, remains in the forefront of my mind. It was created by the incredible Kyle Bland at PitcherList, a personal inspiration of mine. Plus was beautifully explained by Nat Schwartz.
What I want to do today is not only talk about its brilliance but focus on a particular strength it has as a statistic: its ‘stickiness’—or, in easier terms, its continuity year-on-year. See, baseball has become extremely good at analyzing performances. I was even able to show what statistics best contribute to winning, mostly due to the brilliant research and information already out there. Much harder though, is predicting future outcomes.
How do we even try to project?
There are many amazing projection systems available. FanGraphs show you as many as eight projections when going onto a player page. The issue with predicting stems from the issue that you have no data. The statistics we know and love are all based on player performance, even our expected statistics just use data, like exit velocity and launch angles. Next year’s data has not happened yet.
Luckily, the stats we use are the same every year and, in most cases, so are the players. Surely then, last year’s stats are a good indication of next year? Well, no. I have previously spoken about year-on-year correlations, particularly involving BABIP. Ultimately, showing the major limitations statistics have to project itself next year. In Nick’s article, he shows how Process+ has a year-on-year correlation of 0.45, an impressive number when compared to wRC+’s year-on-year correlation of just 0.23. Remember, wRC+ was the most complete offensive stat we found. That article though DID NOT include Process+.
It’s no surprise the success Process+ has in this regard when you learn what it is built on. Another point I made in my BABIP article is how our expected stats are much better predictors than our basic stats. For example, when looking at 2023 to predict 2024 BABIP.
| 2023 stat | Correlation to 2024 BABIP | Correlation to 2024 xBABIP |
| BABIP | 0.37 | .312 |
| xBABIP | .484 | .514 |
Notice that not only is it a better predictor, but it is far ‘stickier’. The reason I mention this is that Process+ is a “per-pitch xwRC+”. A name that beautifully describes its use of underlying numbers, but perhaps understates the impressive amount included.
What does Process+ include to help predictions?
As I’m sure you know Process+ is built by three individual statistics;
- Contact Value
- Power Value
- Decision Value
What is important to note though, is that these statistics were designed to have no interactions. Of course, it is unlikely they’re independent, but it can be very helpful in making sure all aspects of a player’s hitting are included. What exactly goes into the three statistics isn’t stated explicitly. We can, though, discuss how they correlate year-on-year.
Contact and Power fall into the same bracket. Both are using long lists of statcast metrics. Batspeed, for example, features as part of Power. This means these are more like our expected statistics, as they use data at the plate rather than the ultimate result of the pitch. We know this has far better predictive qualities than those statistics that don’t. Hence why Process+ is far stickier than wRC+.
Decision Value is my personal favourite, particularly given how different it is from most other statistics available. It is broken down into oDV and zDV, for outside and inside the zone respectively. Many would expect this to be the most consistent statistic year on year. Established hitters for the most part remain consistent. Most hitters have areas where they will make good decisions and areas where they make bad decisions. The exceptions would be rookies or ageing hitters whose approaches would improve/deteriorate respectively. Knowing this though means projection systems can have built-in aging calculations that adjust accordingly.
So why does any of this matter?
That is, of course, a hard question to answer. The purpose of Process+ was to help show a hitter’s underlying approach, something that “every analyst and fantasy player wants to know”. A statement I find brilliantly ironic. This new statistic doesn’t even care about results, simply knowing what a player wants to do at the plate, and whether he is successfully executing such plans. On paper the opposite of what analysts want to know, It leaves us with an idea regarding projecting:
To know what a hitter is going to do in the future has less to do with past success. It is more so how he went about trying to achieve that success.
Extra innings
That was where this article was supposed to end – an emphatic statement.
As time went by though, it began to bug me. Is this true? In any part of life, that is. Is it realistic to base our decisions on someone’s processes rather than prior results? It bugged me so much I began asking everyone I could what they thought, making it a real-world question anyone could answer. I posed them a simple question. Which of these two candidates would you employ?
- Person One has had 10 successful years at a business, risen up the ranks and based on their resume is the best candidate available, even working a level higher than this job. He’s a good worker, but you assess him to perhaps be fortunate to make it to the level he has.
- Person Two was at another business for 10 years. His resume is enough to get him an interview, but they have never worked at the level of positioning you’re offering. At the interview though, he excels and you are confused as to why he hasn’t been given a chance to do a job similar to what you are offering.
Essentially trying to make them pick between one person who has had a lot of success on paper and another person who they feel has a better approach but no real on-paper success.
Perhaps to my surprise, a clear majority I spoke to chose Person Two. This radical new approach to baseball appears to have been standard practice in the workplace for years. My mum though, made an even more important point. She asked: who’s already employed there? Having a mix is crucial to any baseball team. Having experienced professionals is a really important part of baseball. The signing of Alex Bregman on a purely statistical basis looked questionable. However, you can’t put numbers on his influence in a dressing room full of young, impressive talent.
As a community, we still have a long way to go in projecting the future. Process+ is first a great evaluation tool, but also a step in the right direction for projections.
Finally, thank you to all those who work tirelessly to bring us publicly available projection systems. If this has taught me (and hopefully you reading) anything, projection systems are a near-impossible task brought to life by some incredibly talented individuals.
Article by Alex Williamson. Look out for more analytics-based content from Alex during 2025 and follow him on Bluesky @alexwilliamsonmlb.
Featured image – NYT/Getty Images
