The Change: Punt These Stats
It’s a tale as old as fantasy sports: pay attention to nine of the ten categories on draft day and you can dominate those categories and win. There are some accepted tenets to this strategy, called ‘punting,’ but as it is with all these things we hold for granted, it’s useful to look at what the numbers think of our Conventional Wisdom.
For one, which stats to punt?
Punt homers and surely your RBI will suffer. Punt steals and your runs should suffer. Punt batting average and everything will suffer.
We’ve settled on saves and steals, perhaps, but let’s see if that was the right thing to do. By correlating the regular 5×5 stats against each other, we can find out which stats are linked to each other. The stat with the lowest correlations to the other stats stands alone on an island and is the most puntable.
Hitting stats first. Shown are the r-squared numbers. All p values were less than .00001.
| HR | R | RBI | SB | BA | |
|---|---|---|---|---|---|
| HR | 1 | 0.451 | 0.725 | 0.048 | 0.118 |
| R | 0.451 | 1 | 0.701 | 0.140 | 0.296 |
| RBI | 0.725 | 0.701 | 1 | 0.0002 | 0.242 |
| SB | 0.048 | 0.140 | 0.0002 | 1 | 0.065 |
| BA | 0.118 | 0.296 | 0.242 | 0.065 | 1 |
| 0.468 | 0.518 | 0.534 | 0.251 | 0.344 |
Looks like we were right. Steals are indeed the least comingled, correlated stat in the regular 5×5 toolbox. They aren’t at all correlated to RBI, they barely have anything to do with homers or even batting average. It’s worth noting that you’ll have to keep an eye on your runs stat if you punt batting average, but once you keep that in mind, you’re ready to punt the stat and collect your glory. Batting average comes in second, but there’s still a fairly strong link to runs and RBI that should make you shy away from the strategy.
The pitchers!
| K | ERA | WHIP | Wins | Saves | |
|---|---|---|---|---|---|
| K (K/9) | 1 | 0.133 | 0.145 | 0.002 | 0.128 |
| ERA | 0.133 | 1 | 0.613 | 0.033 | 0.058 |
| WHIP | 0.145 | 0.613 | 1 | 0.052 | 0.077 |
| Wins | 0.002 | 0.033 | 0.052 | 1 | 0.160 |
| Saves | 0.128 | 0.058 | 0.077 | 0.160 | 1 |
| 0.282 | 0.367 | 0.377 | 0.249 | 0.285 |
Here’s an observation before we get into anything more detailed: all of these correlations are weaker. Makes you wonder if you shouldn’t just punt two pitching stats if you’re thinking about punting two stats.
And it also looks like maybe our Conventional Wisdom was off a bit. Wins are actually terribly correlated with other pitching stats. Strikeout rate, ERA, Whip — these things do not predict wins. Saves are negatively correlated with wins, so that’s not even a helpful .160 there. Generally, wins suck. Just don’t even think about them. Saves suck, too, but wins suck more.
Unless you change the minimums (from 30 IP to 100 IP) so that you’re looking only at starters. Then your r-squared on ERA and Whip changes to .283, and saves become the most puntable stat. This makes sense — if you’re going to punt saves, you’re not really going to look at relievers, and so then you can limit the pool to starters and get wins by finding ERA and Whip. So conventional wisdom was actually right on saves and steals.
The other punting truism we can sort of test is the idea that it’s really only an H2H strategy. I’ve said this often, because I think it’s really hard to win if you aren’t doing well in one category in roto scoring.
To look at this, let’s look at the average number of points it takes to win a 12-team mixed league with 5×5 categories. Thanks to Rudy Gamble at Razzball, we have that number: 101.5. Here is an example of how well you need to do in the other categories if you’re punting one category.
| Points from Punted Category | Points Needed | Finishes Needed |
|---|---|---|
| 0 | 101.5 | 3 1sts, 6 2nds |
| 1 | 100.5 | 2 1sts, 7 2nds |
| 2 | 99.5 | 1 1st, 8 2nds |
| 3 | 98.5 | 9 2nds |
| 4 | 97.5 | 8 2nds, 1 3rd |
In deeper leagues, with more points on the table, but more points needed to win, the calculus changes a bit. It’s still a terrible idea. In my 20-team league, 176.5 has been the average winning score over the last five years.
| Points from Punted Category | Points Needed | Finishes Needed |
|---|---|---|
| 0 | 176.5 | 6 1sts, 3 2nds |
| 2 | 174.5 | 4 1sts, 5 2nds |
| 4 | 172.5 | 2 1sts, 7 2nds |
| 6 | 170.5 | 9 2nds |
| 8 | 168.5 | 7 2nds, 2 3rds |
In head-to-head leagues, you can fight to win nine categories and give yourself and advantage by punting a category. Particularly if you get reliever-eligible starting pitchers, or put non-closers with elite rates in your reliever slots and put together a wicked lineup and staff, you can create an advantage for yourself by punting saves or steals.
But in roto leagues, you’ll have to do *really* well in the other categories to overcome a true punt. If you get fewer than four points in one category, you basically have to finish first or second in the other nine categories. That’s tough to do.
Score multiple times for Conventional Wisdom in fantasy baseball today. It’s still nice to see what the numbers have to say about these things.
With a phone full of pictures of pitchers' fingers, strange beers, and his two toddler sons, Eno Sarris can be found at the ballpark or a brewery most days. Read him here, writing about the A's or Giants at The Athletic, or about beer at October. Follow him on Twitter @enosarris if you can handle the sandwiches and inanity.
Punting can be useful in a rebuilding roto year (Only league). You can get 2nd or 3rd place in roto by punting saves. Pick up a couple of set up men as potential closers the following year for $1-$3. Go Stars and Scrubs where the scrubs are guys without current jobs. If things fall your way with the scrubs, you’re contending. Otherwise you can sell the stars for future value.
This is so awesome, thanks Eno!
What population did you draw from to generate the offensive correlations? MLB as a whole?
for hitters yes. pitchers were 30 IP minimum, re-run with 100.
I’m having difficulty replicating your numbers. I’m guessing you removed all the pitchers from the offensive totals. That sounds like work.
from the leaders page, choose “NP” on the far right instead of all to eliminate pitchers
They question lies in this: You can punt Saves and Steal in H2H
But SHOULD you punt them?
I would say probably. At least one of them anyway.
You generally have 3 options with SB’s. Draft an Elite stealer, draft a generally speedy team with no elite stealer, or punt SB’s.
— For H2H I’d say drafting an elite SB guy, followed by punting are the best options. Drafting a balanced team can leave you “weak everywhere” and cause you great headaches in H2H from week to week.
— For Roto, I’d say drafting a balanced team, followed by drafting an elite stealer are the best options. Punting should never be seriously considered if you are trying to contend.
I don’t think the same can really be applied for closers as there isn’t as big a statistical difference in the number of saves an elite closer would acquire over a less elite option (plus, inconsistent as all hell). Punting saves can work in H2H and you could just target elite rate RP’s instead… and again, I would never recommend punting for Roto.
Yes, punting in roto “can work” and can get you into contention… but the odds are extremely slim in a competitive league. I can tell you that in 7 years no one has ever won my roto league with less than a 2 in any category… and the highest anyone has finished with a 1 in a category is 3rd.
Yeah typically: I’m the guy that “punts” saves. I’ll try to balance the power speed guys, with going for a Hamilton type.But in H2H. I will balance Chapman/H. Alvarez type maybe, unless I should punt saves.
How about punting based on the fact that AVG/ERA are (relatively) hard to predict? Lower predictive ability –> more likely you can get ‘lucky’ or ‘unlucky’ and allocate your dollars poorly.
You can’t punt AVG, ERA, etc. but you can ignore them and grab the best players no matter their standing in those categories. I did that successfully with OBP by ignoring it for the first half of the draft and then adding it back to consider it during the second half of the draft.
i punted wins and k’s successfully last year in a h2h. relievers with excellent era/whip are always on the wire and i loaded up my bench with platoon batters. the result was a roster that dominated the 5 hitting cats and was usually winning some of saves/era/whip.
Another factor to consider in punting: the more *categories* you have, the more likely a punting strategy will work. There’s a simple math reason for this: punt a category in a 5×5 and you’re losing 10% of opportunity, punt the same category in a 7×7 and you’re only losing 7% of opportunity. I’d imagine running the correlations above with common add-on stats (OBP, OPS, K/9) would also reinforce correlation clusters.
Keeper leagues are also a factor, as they typically require more points to finish 1st, since in any given year, some owners are dumping.
I find it near impossible to punt a category and win, since all the other contenders will be ahead of the dumpers in most categories.
My primary H2H league uses IP, QS, K, SV, ERA, WHIP and has no IP floor or ceiling, and we start 10 pitchers. It’s not an ideal category mix but we’ve all grown used to is. Trying to win saves is a good way to lock in 3 losses each week in pitching.
Assuming nobody is gaming the system, drafting a balanced team is the way to go in all leagues. The marginal gains by improving in a weak category are always greater than the marginal losses by not stacking your team in a different category. If you think drafting a balanced team makes you weak all the way around, then you’re either bad or unlucky. Win your league with good player valuation. It’s fun.
The game changes in H2H when you incorporate pitcher streaming or an all-reliever strategy coupled with benching them when your rate stats are good enough in comparison to your opponent. Good leagues will have minimums/maximums to try and combat this, though.
Do you think that part of the reason for punting not working is that trying to finish 1st or 2nd in a category is much more likely to waste stats?
Examples while punting steals:
You win HRs by 30 and RBI’s by 60.
OR
You and another team are fighting for the most HR and both have at least 30 HRs than everyone else.
This is a good point. Although, SB is the most likely hitting category that people overshoot too much due to the high standard deviation, and shouldn’t be as big of an issue for the other categories. The most efficient place to finish in R/HR/RBI/AVG is 1st, believe it or not, and for SB it’s 9th or 10th. There’s an article by Todd Zola on this topic from a few years ago behind a pay wall, called “Category Efficiency Ratings”.
It may also be possible to adjust for deviation/unbalanced drafting by making trades. If you are crushing the league in HR or SB heading up to the trade deadline, identify someone who is desperate in those categories and trade some of your excess. (You do need to be careful they aren’t in a position to pass you as they move up the SB rankings though.)
For example, last year I somehow ended up drafting Billy Hamilton and Jose Altuve. By July I had a lead of like 30 in SB, and I dealt Hamilton and Keuchel for Adam Jones.
Now unfortunately Jones hit .258 the rest of the way with good-not-great counting stats, and I finished second in the league. Also, I still led the league in SB by 28 as Altuve ran wild. So, okay, not perfectly executed. But in principle I think this shows how you do have options to recover and extract value if you over-shoot a category.
One thing you didn’t look at is the impact of punting saves in an NFBC style league where you have no innings cap, 9 P slots with no SP/RP requirement and weekly line up changes for pitchers. The tactical execution of the strategy involves using 9 SPs every week where your competition uses only 8,7, or 6 gives you a very good shot at 1st in Wins and Ks. You lose some WHIP/ERA help you get from relievers. You gain draft slots/$ that would otherwise be spent on closers. You can also potentially punt on the top two tiers of SPs as you’re generally paying the premium for the high K rate (though the high K rate guys generally have the better ratios). So if in a 15 teamer you’re shooting for 9 in saves, you can hopefully make up 6 of those in K/W and you’re gunning for +3 on offense garnered by extra draft/$ for offense. And you have to hold your own in ERA/Whip. The 2012 NFBC overall auction championship was won by a save punter. In general, extreme strategies are easier to execute in auctions. I would say because saves are such a volatile commodity, punting saves actually builds a pretty repeatable strategy. I would bet that it reduces your chances of a 1st place finish but increases your chances of top 3 cash in that format.
In my H2H league last year we 9×9.
Offense (R,H,HR,RBI,SB,BB,K,TB,AVG);
Pitchers (W,L,SV,K,HLD,TB,ERA,WHIP,QS)
I focused on OPS for hitters, punting SBs.
For starting pitchers I focused on WHIP, Ks, and IP.
For relief pitchers I focused on WHIP, Ks, and HLD. (Easier to find good set up men than closers.)
I was a strong 2nd and will use this strategy again this year.
My opponents think I’m a robot!
In a H2H league I like the idea of punting saves and using your RP’s as rates specialist.
I also would like to have it set-up so that SB are not counted as a stat per se, but can be added to total bases.
Single + SB = 2 total bases (same point total as a double; not sure if it is possible)
4X4
ERA, WHIP, K, and W/QS
Average/Total Bases, R, RBI, and HR
If you are trying to track stolen bases after a single vs after a double (or walk or whatever), I think that would be impossible. But you could just have a category that is total bases plus stolen bases. You would also probably have to subtract CS. It would make guys with some power and great storm base rates really valuable. Like Kinsler and Beltran in their primes.
I love messing around with strategies like punting while I’m waiting for real baseball to start. I used the Auction Calculator to punt era & whip in a 10 team H2H 6×5 with daily changes (streaming capable). I think I’d end up with the following pitchers filling the rest of the draft with hitters that I think would dominate their 6 categories. What would you think of the following staff’s ability to win in H2H (round shown)
Chapman 5
Holland 6
Benoit 15
Carrasco 8
McHugh 14
Stroman 16
Kennedy 17
Hughes 18
Fiers 22
Salazar 23
Shoemaker 24
Hutchinson 25
Depending on the league and categories, if there are a few people who tend to punt a category, you can do better by making a token effort instead of being one of the punters. For example, having a single closer, or a single speed guy, in some leagues is a waste of time (you will lose the category every week) and in others has a benefit (one roster spot is getting you wins 3-4 times through the schedule). The marginal benefit of adding more steals/saves may end up not being worth it, so instead you use your other slots on the other categories.
The flip side of this is being aware of overkill situations; carrying 5-6 closers in a 12-team league positions you strongly in one category, but there’s a clear opportunity cost that tends to get overlooked.
I’ve decided to punt saves during the draft in my 6×6 roto keeper league (extra catagories: OPS and QS). I just traded Betances for deGrom, both of whom can be kept through 2018 in the 15th round. I’m hoping our format’s SP-heavy lean and grabbing a closer or two on the waiver wire for a couple of points will make this a viable strategy.
Hello Eno, great article I was wondering in a league with 7×7 categories Batting (AVG, Runs, HRS, RBIs, SBs, On Base % and Extra Base Hits) Pitching (ERA, WHIP, Wins, Saves, Strikeouts, Quality Starts and K Per 9) which category or categories would you punt. The league I am in is H2H and I have found it hard to find categories to punt in this pool due to how it has 7×7 categories.I was thinking about punting steals and saves but I am questioning if that would really give me an advantage since saves do correlate well with K per 9.
If your league doesn’t have dedicated SP positions, you should punt wins and strikeouts. Run out a huge bullpen of 7-8 relievers and dominate saves. The great thing about closer is that, on average, they have a lower ERA/WHIP than starters, so it gives you an advantage in that category. Also, because you can generally draft league-average closers late in your draft, you can focus on offense early (first 9-10 picks at least), so you can dominate on offense as well.
The point of the above strategy, if it wasn’t clear, was that you would have zero starting pitchers, ONLY relievers. Not only can you draft offense early, but you don’t have to carry as many pitchers (in some leagues 5-6 closers may be sufficient to dominate saves), so you can use extra roster space for bench offense too.