Poll has no actual undecided respondents (a corner of the electorate who are likely Schmidt voters at this juncture) which could be crucial to this race if they exist in large enough numbers.
Doesnt the graphic have 5% undecideds which are then pushed to choose someone aka leaners?
Yes, but that push here forces respondents to pick a candidate doesn't allow them to express that they are genuinely undecided (that's why I said "actual undecided" respondents). Have been seeing a lot of public pollsters doing this recently, it can be messy in close races.