The narrative about the new NBC drama “Smash” this spring, promulgated largely by critics and bloggers, centered on the show’s serial failings: it did not live up to its pilot; it alienated viewers with sappy writing and acting; and the only reason NBC ordered a second season was because Bob Greenblatt, the top NBC entertainment executive, decided its fate based on his heart, not his head.


The bottom line was that “Smash,” after weeks of deteriorating ratings, was a flop.


So how did such a disaster wind up tying for the top position at 10 p.m. among viewers most coveted by network executives — those 18 through 49 years old? The answer lies with the digital video recorder, or DVR.


Negative reports about the performance of “Smash” generally took into account only how it had fared on the first nights that it was shown on NBC. As is now the case in prime time, however, a show’s true popularity can no longer be assessed by instant analysis.


The first-place 18-49 ranking for “Smash” occurred in what is known as the “live+7” ratings, which include playback of recorded television shows over one week.


On average, “Smash” added 2.5 million viewers over seven days of playback and averaged a 3.3 rating in the 18-49 category. No 10 p.m. network show did better in that category. Two CBS dramas, “Hawaii Five-O” and “C.S.I.,” each scored a 3.3 also.


“Playback is a big factor in determining how many people like a show, there’s no question about it,” said Leslie Moonves, the CBS president, who has more programming experience than any other current television executive.


Brad Adgate, the longtime head of research for the media buying firm Horizon Media, acknowledged being astonished that “Smash” had a 3.3. “That’s a good number for a show in any time period,” he said. “It’s especially good at 10 o’clock.”


The viewers that the show consistently added over seven days, Mr. Adgate said, meant that it surely deserved to be renewed for a second season.


“A lot of insights can be gleaned from playback,” he said. “It is an indication of how passionate people are about shows.”


Increasingly, delayed viewing must be factored in by networks trying to decide which shows to retain and which to reject, Mr. Moonves said. He pointed to how many viewers were added by “Hawaii Five-O” each week: more than 2.4 million. “That is a big, big number,” he said.


What the “live+7” performance does not do is make more money for the networks. The accepted system of advertising sales continues to be based on what is called a “commercial+3” rating: that is, how many viewers watched the commercials within three days of a show’s first broadcast. (Advertisers do not want to pay for viewers who skip through all the commercials.)


But looking at how many viewers were added on delayed viewing does give programmers a better sense of how ardently a show is being followed.


“Smash” ranked fifth among shows in terms of number of viewers added. ABC’s “Modern Family” was on top with 3.3 million. But as often is the case in television, the results should contain an asterisk. “Smash,” which is designed to run only original shows, did not present any repeats in the season that ended in May. Measured against only the original episodes of other shows, it would rank behind several other 10 p.m. dramas, though its 3.3 is still a strong performance for that time slot.


The example of “Smash” mainly proves that initial ratings are not a particularly valid currency, though they still are the ones most widely reported and discussed in the industry.


Preston Beckman, the longtime programming scheduling executive, first at NBC and then Fox, said, “The relative strength of shows doesn’t change all that much,” when playback is considered. Hits only get bigger; failures improve, sometimes by startling percentages, but never add enough viewers to change the overall impression.


The NBC drama “Awake,” for example, had one of the highest percentage increases of any show on television over seven days — 61 percent. But that only meant it edged up from a rating of 1.08 to 1.74, still cancellation territory.


On the other hand, another NBC drama, “Grimm,” gained 69 percent over seven days, but that was a lift from 1.53 to 2.58, strong for a show on Friday night.


“Some marginal shows suddenly start to look more viable,” Mr. Beckman said. “That’s one of the reasons why some of the networks are still trying to build scripted shows on Fridays. Friday numbers become more acceptable with playback factored in.” Friday is now the second least-watched weeknight, after Saturday, so networks tend to run less costly shows like newsmagazines and reality shows.


But the analysis is mainly important in terms of how much a show is valued by viewers — not by advertisers.


One senior NBC executive, Ted Harbert, suggested to advertisers at the NBC upfront sales presentation last month that it might be time to consider basing ad sales on seven days of viewing instead of three.


That is not likely to happen, Mr. Adgate said. “If Nielsen reported live plus 21 days, the networks would want to get paid for that,” he said. And Mr. Beckman conceded, “A lot of delayed viewing is a function of commercial avoidance.”


But Mr. Adgate suggested that another trend may hold out more hope for making a profit from delayed viewing. He noted growth in network series being watched on a video-on-demand basis: viewers ordering episodes they have missed. “In V.O.D. they can include commercials that you can’t skip through,” he said.