(Photo by Rich von Biberstein/Icon Sportswire)
It’s Friday, and you’re locked into a tough matchup. You notice that while you have a small lead in ERA and WHIP, you are just behind your opponent in wins and strikeouts. You look ahead to the weekend, and you notice that you and your opponent each have just one starter going this weekend. You have to make up that difference, but you don’t want to trash your ERA and WHIP. “I just gotta find a decent streamer,” you think to yourself as you click on the free agents list, “Surely I can find one right?” You look through the probable starters and you narrow it down to three names: Kyle Freeland, Jake Odorizzi, and Julio Teheran. Taking a quick glance at their numbers, you notice that they’ve all got about 3.40 ERA, about a 1.20 WHIP, and their matchups are roughly the same. Okay, kind of hard to separate them, but that’s fine, you’ve learned to look past the basic numbers by now. But wait, they all have basically the same strikeout and walk rates too. It’s probably just going to come down to how you feel about these guys.
What you really need to know, without really knowing it yet, is which pitcher has the least likely chance of blowing up your ratios. It wouldn’t do you any good to catch your opponent in wins and strikeouts, but give up the ERA and WHIP ones. You’re right back to where you started if that happens. You need to find out how consistent these starters have been. Imagine two different starting pitchers both with 3.50 ERAs. One pitcher consistently produces an ERA between 3 and 4 every start, while the other throws shutouts one week and gets blown up the next. Being able to distinguish between the two could provide you quite the advantage if you knew what you needed. Quantifying this consistency isn’t brand new, we’ve had the Quality Start for a long time now, but in fantasy baseball, there’s a huge difference between a pitcher going 6 innings and giving up 3 earned runs, and a pitcher going 8 innings and only allowing one run. Not only that, but QS is a binary stat, either the pitcher met the conditions, or he did not, and we need more outcomes than that when it comes to fantasy baseball. That’s where the Volatility Performance Ratio (VPR) comes in.
What is VPR?
Nick Pollack, founder of Pitcher List and the man behind the weekly Pitcher List among many other things, was the one who came up with VPR back in 2015. He explained it as such:
“The first step was to take QS and change it from a binary situation (QS or no QS) to three outcomes: Excellent Starts (ES), Neutral Starts (NS), and Poor Starts (PS). Here are their definitions:
Excellent Start: 0.00 – 3.00 ERA (minimum 6 IP, otherwise it’s a Neutral Start).
Neutral Start: Greater than 3.00 – 4.50 ERA.
Poor Start: Greater than 4.50 ERA.
I tested a few specific metrics until settling on the ones above. A pitcher throwing at least 6 IP with an ERA 3.00 or lower should feel excellent (think 2 ER over 6 IP as the threshold). Additionally, starts don’t become toxic to your week until north of the 4.50 mark (3 ER in 6 IP). These ranges are subject to change from year-to-year, but currently represent how owners should feel given the pool of players.
Now that I had the raw data, I needed a way to express them elegantly. I created two metrics to help: Volatility % and VPR.
Volatility % expresses how often a pitcher performs an Excellent Start or a Poor Start:
[(Excellent Starts + Poor Starts)/(Games Started)]
For example, a pitcher with 9 ES, 4 NS, and 3 PS has a Volatility % of 75.0% (12 out of 16).
VPR stands for “Volatility Performance Ratio”, which is the ratio of Excellent Starts to Poor Starts: [Excellent Starts/Poor Starts]. A pitcher with 9 ES, 4 NS, and 3 PS has a VPR of 3.00 (9 to 3).
Think of Volatility % (or V-rate) as the sidekick to VPR. While VPR is the important stat at the end of the day, without understanding V-rate there is no way to properly apply it.
Ideally, you would want a pitcher with a 100% Volatility and a 999 VPR (infinite). That means he pitches an excellent start every time. On the other hand, you wouldn’t want a 100% Volatility and 0.25 VPR, which means he’s throwing a dud 3 out of 4 times.”
There is your definition of VPR, so let’s brings this back Teheran, Freeland, and Odorizzi. You need to catch your opponent in wins and strikeouts, but you don’t want to sacrifice ERA and WHIP doing so. What you need is to find the pitcher who’s least likely to go out and get crushed. It’s still early in the season so these numbers are all relatively small, but they still paint a decent picture.
| Name | GS | IPS | ES | NS | PS | VPR | Volatility |
| Julio Teheran | 8 | 5.38 | 5 | 0 | 3 | 1.67 | 100% |
| Kyle Freeland | 8 | 5.86 | 3 | 3 | 2 | 1.5 | 62.50% |
| Jake Odorizzi | 8 | 5.26 | 3 | 1 | 4 | 0.75 | 87.50% |
Looking at the table above, I certainly wouldn’t want to pick up Odorizzi for my needs, given his likelihood to post a Poor Start. I probably would pick up Freeland based on that data, as he’s the least likely to go out and have a bad outing. On the flip side though, if I was chasing my opponent in ERA and WHIP and wanted the pitcher who was most likely to get me a dazzling start, I’d be chasing Teheran, because he’s the one who’s most likely to have an excellent start. If he gets blown up, oh well, I was losing those categories anyway, but Teheran would give me the best chance based on how he’s gone this season.
The benefits of VPR go beyond just picking streaming pitchers. The stat was actually created due to David Price’s 2014 season, when he was undoubtedly an ace, and yet Nick just didn’t feel comfortable starting him. He couldn’t quite put his finger on why he didn’t feel comfortable with his ace, and the answer was clear after he looked at Price through the lens of VPR: Price had posted a whopping 10 poor starts in 2014 and had a VPR of only 1.6, whereas other aces like Johnny Cueto (6.0 VPR), Jon Lester (5.0 VPR) and Cole Hamels (3.6 VPR) were much more consistent.
We’re still working on all the different ways we could apply VPR, as we still see a lot of potential here for this stat. We’ll be updating our VPR table throughout the season to help see the big picture in performance predictions.
There is a counterpart to VPR called kVPR, which takes the same approach with strikeouts and WHIP performances. I’ll dive into kVPR shortly within the next week.
Any chance you could publish the VPR and Volatility in a table with all the starting pitchers?
Nevermind totally missed the giant link at the bottom!
Whoa this is a awesome idea