With fairness this metric is not counting his two relief performances which have both been excellent. Mopping up is hard to do. That brings the figure to 10 solid games .
Not familiar with the metric you are using but it seems to be arbitrary cutoffs. As bad as Talbot has been in 6 games he's had zero GP that approach the toxic, get pulled 2 starts that Koski has had. In an arbitrary metric this could mean Koski has had only two very bad starts but they were horrifically bad. If one digs further the 6 games in which Talbot was bad, the whole club was brutal most of those games.
It's based off games where their save percentage is above league average, or above .885 if the opposition had less than 20 shots against (I think Vollman created it). It isn't perfect and it does lose context a bit though as it doesn't count a game like Kosko vs. Avalanche due to some late goals, despite that being subjective a quality start, but it is a quick way to look at things.
Agreed Talbots non-"Quality Starts" weren't really that bad though when viewed with proper context, (i.e. St. Louis's strange goal followed by a 5 min penalty kill dropped his percentage below average on the games despite a decent start, the entire team belly flopping vs. San Jose and Winnipeg going ballistic on Edmonton in the third period) and certainly nowhere near the belly flops of Kosko vs. Vancouver and LA.
Agreed, his relief performance were very good and ultimately they are responsible for the overall swing though in GAA and Pct. I posted a few posts back. Especially when coupled with Kosko's poor start in each of those as you'd think the opposition wouldn't continue to score at the same rate if Kosko stayed in. (excluding those games Kosko is 4-3, 2.60, 0.924 and Talbot, 3-2-1, 3.13, 0.910). Understanding the swing in these case show that Talbot has been a little more consistent the past month, but less spectacular.