Macro Trends in RP Value: SVHD

Like a moth drawn to a flame, I find myself returning to the strange comforts of relief pitcher analysis.

In last week’s article, I looked at whether or not there seems to be big-picture movements in relief pitcher value and found that generally, despite the overall vibe shift and changing usage patterns, the proportion of good reliever seasons and distribution among the players accruing these seasons is not materially changing.

But – as one commenter noted – what about SVHD (saves + holds) leagues? It’s a fair question and one to which I am personally sympathetic. Despite all of my digital hullabaloo about standard 5×5 roto settings, I don’t actually play in a league where holds are not counted. As more and more home leagues shift away from saves dominance it might be worth a closer look.

SVHD Macro Trends

To achieve this, I essentially repeated my analysis with SVHD replacing saves. FanGraphs doesn’t record holds prior to 2002, so that year served as a lower bound in the analysis similar to how 1991 did in last week’s article. To create an apples-to-apples comparison, I also reran a saves-only version of the analysis with the same new lower bound.

The first thing to note is the values for what constitutes an above average and elite season. To refresh, I’ve calculated auction dollar values for all players from 2002 forward, focused on relievers, and excluded all who didn’t turn in at least one season of positive fantasy value. The threshold for “average” is the mean of this group, and the threshold for “elite” is one standard deviation above the mean. Here are the results:

This confirmed my first suspicion: that the typical bar for a good reliever in a SVHD league is lower because there are more paths to success, meaning the player pool deriving fantasy value is larger. This will cause a type of inflation where the same reliever’s season will deliver you less value in one league than it might another.

I also wanted to check on how the trend for SVHD leagues has behaved over time:

Again, pretty flat trends in both cases. Interestingly, there are slight upward slopes to both but given how gentle the slopes are and the high variances around the trend, we have to treat that with a grain of salt.

Looking also at the one-hit wonder rate, the pattern is overall similar to what we saw in a SV-only format. As another refresher, a one-hit wonder here is a pitcher who only registered one season by either threshold in his career. Results for the active player sample are almost certainly biased upward because there are young relievers who have not had many changes to accrue a second good season.

These are the numbers for a SV-only league. I’ve included the grouping for 1991-2025 for information only; bear in mind that they aren’t an exact apples-to-apples comparison to this analysis because the thresholds are different. There isn’t a huge gap (around $0.60 in value between elite relievers, half that for above average) so the impact on the data will be slight. Overall, the picture with a saves-only league is fairly consistent over time, which is what we’d expect based on the data we saw earlier. For SVHD leagues, the ratios are different but the pattern (more one-hit wonders in the earlier sample) is consistent:

The most notable difference is the jump in rates. If we assume that one-hit wonder relievers are, on the whole, less talented than their peers who register multiple good seasons it would follow that lowering the bar to what counts as good might introduce the pool to a lower caliber of player. That said, the effect is only particularly stark when considering elite seasons, and the sample sizes we’re dealing with are probably too small to draw definitive conclusions.

I also hypothesized that we aren’t seeing larger shifts in behavior for these data because the player pools are broadly similar. This seems basically correct:

Based on this, there is something like 75% overlap between the two pools. Translated to a per-season level, there are about five elite seasons occurring per setting, on average, and 13 above-average seasons. This translates to two or three total non-overlapping elite players and about six non-overlapping above average players in a typical year. Looking at the Auction Calculator for 2026, we had more elite seasons than we’d typically expect: 11 in a SVHD setting and nine for saves-only. Accordingly, we have a roughly proportional set of non-overlappers: David Bednar, Josh Hader, and Jhoan Duran for saves. In SVHD leagues, the non-overlappers depend on where you set the threshold cut off, but if you call it $13, we have Paul Blackburn, Didier Fuentes, Garrett Whitlock and Adrian Morejon.

Overall, the main takeaway here is that SVHD leagues are behaving in a broadly similar manner as their save-only counterparts. It seems as though any difference, to the extent there is one, is that SVHD leagues are actually introducing more heterogeneity in the pool due to an inflationary effect on value.

Value Composition

Another thing these settings share is the large dependance of a single category on the overall value for relievers. Or, put differently, water is wet:

While this finding is reductive on its face, it’s still important to measure. There’s a lot of research in more serious fields that relies on initial excavation of these basic truths. Here, relievers are deriving approximately 40% of their fantasy value from either saves or SVHD.

With this in mind, I wanted to look at whether this is changing over time. To do this, I plotted save and SVHD-specific dollar values over time. First, above-mean seasons:

In save-only leagues there is basically no change over time in dollar value attributed to saves. So, for a decent but not great reliever, saves are just as important now as they were two or three decades ago. For SVHD, the picture starts to shift. There is a pattern, but it’s faint, and the correlation is poor (-17%).

For elite reliever seasons, the picture becomes somewhat finer. The correlations still aren’t super high, but there’s a clear directional pattern in both, particularly in SVHD leagues. Another thing to note is the heteroskedasticity in both but especially in the saves-only data. Heteroskedasticity is a scary econometrics word but it basically just means that variance in your y-axis variable changes as your x-axis variable changes. You can visually detect it by looking for a cone-shaped splay in the scatterplot, which I do think we can basically see. This means that in addition to the overall decreasing proportion of value being attributable to saves or SVHD, elite reliever performance is also becoming more stratified – there are a handful of guys deriving a large portion of their value from saves or SVHD while more and more are deriving their value primarily from other categories.

Conclusion

This is the pearl of the oyster that I’ve spent 3000 or so words trying to pry open. We may not be seeing changes in the number of good reliever seasons, or changes in who is recording them, or even much difference in SVHD leagues with respect to this behavior. But we are seeing some changes in how the elite relievers are accruing value. This last bit maps closely with our priors and establishes the baseline we need to continue our analysis, so stay tuned for the next episode of Reliever Files.