In Week 1, the Ravens watched Mark Andrews get double covered, and Isaiah Likely went on to have a huge game.
In Week 2, Andrews wasn't double-covered very much by the Raiders, and he went on to have a so-so game while Likely did very little.
It's already Week 3, and there's serious concern about how dominant Andrews still is, and whether or not Likely was fool's good in Week 1.
Let's start with Andrews, who, despite two straight games of under 10 PPR points, has a route rate of 75% in each matchup and should be counted on for something similar in Week 3. Against the Raiders in Week 2, he was open over the middle a good amount of time when they played zone coverage. And with an average route depth of 8.37 yards, Andrews consistently worked further downfield than Likely -- for the second game in a row, in fact.
This figures to be an important factor against a Cowboys defense that's played zone coverage on at least 83% of their snaps in each game. It means they won't double-team Andrews much, if at all, and it means they could leave the middle of the field open for Andrews, which Lamar Jackson could find optimal if he's getting pressured off the snap quickly like he did last week against the Raiders.
By comparison, Likely wasn't used as much in Week 2 (31 snaps) as in Week 1 (52 snaps). That seemed to be for a couple of reasons: The Ravens opted to put more speed on the field with their three-receiver sets, and they got the run game going in the second half for pretty much the first time this year. Likely saw a little more playing time in the first half and then again once the Ravens abandoned the run game late.
Additionally, the Raiders had some answers for Likely, typically putting a defensive back on him instead of a linebacker. Remember his long touchdown against the Chiefs? That was on a linebacker he out-ran. Las Vegas didn't chance that.
So anytime the Ravens run game has a favorable matchup, Likely's playing time could be limited. And even if his snaps and routes aren't limited, a defense that uses a defensive back on him could keep his numbers under wraps. These are not problems Andrews has, at least not to the same degree.
The guess this week is that Baltimore's run game finds success. Dallas' run defense got smashed by Alvin Kamara and a really good Saints O-line last week, whipping the Cowboys silly with zone runs to the tune of five yards per rush over 25 rushes (nearly half gained at least five yards). Derrick Henry has been much better on zone runs (5.3 yards per rush) than gap scheme runs (2.8), so it would make sense for the Ravens to lean on that scheme. And not that massive Dallas D-tackle Mazi Smith has been great through two weeks, but he missed practices to start the week, and if he's out, then Dallas' depth will for sure get tested by Henry.
So if the Ravens can run, both Andrews and Likely's numbers could be a little dicey, save for if they score. If they can't run, or if the Cowboys find a way to put up numbers on the Ravens defense (which is a possibility), then Andrews and Likely could really have strong games.
Andrews plays too much and is more than a basic tight end, so he's a must-start. You knew that already.
I'd have reservations about starting Likely unless you're out of decent options. Taysom Hill, Dallas Goedert, and Cole Kmet are both touchdown-dependent choices. Zach Ertz doesn't have any more upside than Likely. Jonnu Smith is fun but risky since his role and his receiving averages could be impacted by Miami's new quarterback. Mike Gesicki isn't as appealing if Tee Higgins plays. No one else stands out.
So, Likely is a low-end starter by default, but expectations should be mild. At least in his case, there are paths to him seeing more targets than he did last week, but probably not as many as Andrews will have. That seems like a one-week moment.