How do we know if we're right?
A few weeks ago on Fantasy Football Live Brad Evans suggested someone start Brandon Jackson (vs. the Bills) over Chris Johnson (@ the Steelers). I said something to the effect of: "That's insane."
As it turns out, Jackson had 11 carries for 29 yards and a touchdown (and one catch for 10 yards) while Johnson had 16 carries for 34 yards and five catches for 19 yards. In other words, while Johnson had nine more touches and 14 more total yards, Jackson was the better choice for that particular week, thanks to the touchdown.
So who was right?
One way to look at it is as if Evans played a marginal hold 'em hand, say 9-7 off-suit, and I played a strong one, say pocket eights. (Johnson is normally aces, but against a top run defense, you have to downgrade him). Just because Evans hit a nine on the flop and won doesn't change that he was the pre-flop underdog and made a bad all-in call. Sticking with the poker analogy, even 2-7 offsuit (the worst starting hand) can flop a full house and take down A-A. In fact, 2-7 beats A-A 14.15 percent of the time when all four cards are different suits. So it's not exactly rare for even a terrible call to pan out, but that a bad hand pans out obviously doesn't make calling with it a good decision. Can fantasy football decisions be looked at the same way - in terms of process and odds rather than results?
One difference is in poker we can look back and say with some certainty whether the decision to call was in fact the right play and by what margin. That's because the conditions for every pre-flop situation in poker are replicable, and we can simulate a billion hands and plot out the distribution of the results. Plainly this is not the case for fantasy football. Brandon Jackson, a career backup who hadn't shown much in limited duty and coming of a 65-yard game on 18 carries (3.5 YPC) in Philadelphia was suddenly given an opportunity to start in one of the league's most high-powered offenses at home against one of its worst run defenses (30th). Chris Johnson was the record holder for yards from scrimmage in the prime of his career, running behind a strong offensive line and coming off a 142-yard, two-TD game the week prior, but playing against one of the stingiest run defenses over the past several years. This situation existed under those precise conditions only that one time. While both backs might be in similar situations in the future, and Johnson has been in similar ones in the past (he had a poor showing in Pittsburgh in Week 1 of 2009), there will never be another first start on the Aaron Rodgers-led Packers for Jackson, or this precise matchup for Johnson at his exact health, age, skill and experience level again. So obviously it's impossible to define Jackson's projected output vs. Johnson's with anything approaching the precision of a heads-up poker hand. To say Evans played 9-7 against my 8-8 is beyond speculative and actually begs the question.
So how do we tell who was right? One thing we can do is come up with some rough baselines for each player. Johnson's 2,500 yards from scrimmage in 2009 are such an outlier that we can't project anything close to a repeat. I'd say 2,000 total yards and 14 TDs is a decent rough estimate if he's healthy. That gives him 125 yards and a 0.9 TD per game on average. But he was playing the Steelers in Pittsburgh, so that's got to come down a bit. Last year, the Steelers allowed 90 rushing yards per game (3rd in the NFL) and 7 TD. The average NFL team allowed 117 and 13 TD. So let's dock Johnson about half his TDs and 27 yards. That leaves him with 108 yards and .45 TD, or about 13 fantasy points in standard leagues. (I'm leaving out a couple factors, e.g., that the game was in Pittsburgh, that the Steelers offense was handicapped without it's starting quarterback, making a closer, more run-friendly game likely and whether the Steelers are better than average at defending running backs catching passes out of the backfield as there's no need to get overly precise for the purposes of this exercise.)
What about Jackson? I think the first baseline should be Ryan Grant's averages from last year in Green Bay's offense. Grant had 1450 yards from scrimmage in 16 games for an average of 91 yards per game. He also had 11 TDs, roughly 0.7 per game. That's about 13 fantasy points per game. But the Packers were playing the Bills - the 30th ranked run defense. While the NFL average was 117 yards per game, the Bills allowed 156. So let's add another 39 yards to Jackson, for a total of 17 fantasy points.
So far it looks like maybe Evans was right. But there are two major countervailing factors we've yet to take into account: (1) We have to presume Jackson is worse than Grant; and (2) Because Jackson had been with the team a long time, and the Packers had never slotted him ahead of the undrafted Grant, we have to question whether the Packers would make Jackson as large a part of their game plan. The first issue probably only downgrades Jackson a point or two - assuming he got Grant's 18 carries (and probably more looks as a receiver) - after all, we're not talking about replacing Adrian Peterson or Johnson. But the second factor could cut Jackson's production significantly. So if we dock Jackson a couple points for being worse than Grant and then project Jackson for say 13 carries (72 percent of Grant's average - Jackson actually wound up with 11), then Johnson comes out ahead in fantasy points about 13 to 10.
Now, of course, these numbers are just loose estimates, and if you tweaked them more favorably in Jackson's favor, he could come out closer or even possibly ahead. But at least it serves as a rough framework for evaluating who made the right call, regardless of how the cards flopped. And while I'd start Johnson over Jackson again in a similar situation, you can see how difficult it is to know with certainly whether that's the right call.
Things to Take Away from Week 4
Things to Watch For in Week 5
Beating the Book
Saints -6.5 at Cardinals
Arizona's looked like a doormat, but so has Carolina, and that didn't stop the Saints from nearly giving that game away - at home. There's something not quite right about New Orleans, and the Cardinals will be focused coming off of last week's embarrassment. Back the Cards at home.
Saints 24 - 23
Last week we won with the Giants to go 3-1 in this forum and 32-29-3 on the season. We were 10-7 in this forum last season, 131-122 overall. We were 12-5 in this forum in 2008. From 1999-2009 we've gone 1439-1262 (53.3%, not including ties).
The full article comes out on Thursday morning.
Surviving Week 5
I got by last week with the Packers, though it was a little closer than I would have preferred.
This week, there are no double-digit-favorite slam dunks, which of course makes it more interesting. The teams I'm considering are Indy at home against the Chiefs (this probably looked like a slam dunk before the season started), the Bengals at home against the Bucs, the Ravens at home against the Broncos, the Cowboys at home against the Titans and the Saints in Arizona. I was using the Yahoo! game for pick distribution data, but I'm switching to Officefootballpool.com because I think there are less casual, i.e., free, pools there to skew the data. In other words, nothing against Yahoo!'s game which is perfectly good and in which I personally have an entry, but if you're in a pool with a significant payout, I think the OFP distribution data will be more reflective of what you're up against. Here are the current numbers as of Tuesday night:
I'll crunch more numbers when the full article comes out Thursday night, but my first instinct here is to take the Colts despite the large number of people on them. For starters, Indy will be desperate for the win, and they match up well against a speedy finesse team like Kansas City as opposed to a more power-oriented one. Moreover, the 36.8 percent doesn't tip the scale in terms of expected payout that severely. (In Week 2, for example, 54 percent were on the Packers). If I had already used the Colts, my second choice would probably be Dallas. Of course, I reserve the right to change my mind by Thursday.