In 2015, they moved the extra point attempt back, and kickers converted 94.1% of their extra points.
That same year, 2-point conversions were converted at 47.9%. That means assuming all things being equal, we're talking about potentially 2 extra points over 100 TDs. Considering the average team scored just under 38 offensive TDs for the entire season, we're talking about potential 1 extra point, if that.
There are obviously specific situations where that point might have mattered. And there are teams that are worse and teams that are better than others at conversion rates for both kicking and 2-point conversions.
For example, the Steelers went 8 for 11 in 2-point conversions after going 4 for 4 the year before (and 2 for 2 this season so far). In 2015, they scored 42 offensive touchdowns. If they had gone for it every time and we use that 8 for 11 figure, that's roughly 61 points vs. 40 points, a difference of 1.3 points per game potentially.
Now one might argue the 2.36 offensive TDs per game they had scored might have generated some extra points if they had converted all of their attempts in a specific game. That's not how life works, you don't get to decide the distribution of the conversions.
But just trying to understand the impact, that might potentially work out to an extra 3 points in a specific game. That might have helped them win the 2 games they lost by 3 last year, but it might have cost them the 2 games they won last year by 2 and 3 points respectively. So depending on distribution, that could have potentially won them a few more games, evened out, or even cost them 2 more games.
In reality, those plays actually didn't do much to help win any games, and could have cost them. The first attempt and conversion was against us opening week in 2015, and it helped them normalize the score a bit but they trailed by 14 until a late TD with 2 seconds left. But potentially that might have helped. The next 2, both conversions, came in a 43-18 blowout of the 49ers.
The first incomplete attempt came against the Rams, and that could have been costly. The Steelers won 12-6, but a late TD would have potentially cost them the game. They also failed against the Cardinals to leave the score at 12-10, though they would win the game 25-13.
The next attempt did help against the Raiders, putting them up by 4 instead of 3 (11-7 instead of 10-7 at the time). It didn't make a ton of sense at the time, but they did win by 3 points and that extra point helped them avoid a loss by FG, though obviously it wasn't planned that way. They also converted against the Browns in a 30-9 win, so that hardly mattered.
They had 2 attempts against the Seahawks, converting early and failing late after scoring to take a slight lead, 27-26 in a game they would lose by 9 points. Neither conversion nor failure was the reason for the loss as they pretty much evened out.
They converted against the Colts to push an early lead to 14-10 in a game they would win 45-10. And they would convert in their last game against the Browns to go up 13 vs. 12 in a game they'd win 28-12.
So throughout all of that, it might have got them back into the Patriots game if they had scored early enough and helped them protect a late FG loss. It also potentially put them in position to lose another game against the Rams with a late score. The rest more or less didn't matter, although some different distribution in the Seahawks game might have helped or hurt them.
High-variance strategy like this often favours the trailing or weaker team. 6 of those 11 attempts came after the Steelers had scored while trailing, either to close the gap or take the lead. It is also often used to tweak the score in certain game situations. But that's often why teams like the Patriots don't use it much; we're rarely in that weaker/trailing position. Obviously it happens, but not as often as most teams.
So there is a nerdy numbers side to this that suggests it will mean more points overall. But this isn't like the discovery of 3-point vs. 2-point shooting. There are very few attempts to apply this, and the potential impact may not be as significant as first believed.
