From loquacious to laconic, I confessed I had no clue. I've looked at what I've deemed relevant data from several angles relating to the nominating contests but have only glanced at the general election.
My unstudied guess was that if Bloomberg got in, he'd hurt Hillary and help Trump. Yes, Bloomberg was a Republican more recently than he was a Democrat, but his brand of Republicanism doesn't sell west of the Hudson.
A quick check of the data suggests my lazy assumption is wrong. Reuters-Ipsos conducts several daily tracking polls, one of which examines a hypothetical three-way race between Trump, Hillary, and Bloomberg. The following charts show how the latest results shake out among "likely general election voters" who identify as Republican, Democrat, and independent/other party.
In recent election cycles, Democrats have been more likely to vote outside of their party than Republicans have been. In Obama's easy re-election victory in 2012, Democrats were slightly less likely to vote for him than Republicans were to vote for Romney. The same thing happened in 2008. Consequently, I'd expect Bloomberg's numbers to be higher among Democrats than among Republicans, but that's not the case.
More importantly, Bloomberg cuts heavily into an independent field that Trump dominates. Given that Republicans are modestly more inclined towards Bloomberg than Democrats are and that independents strongly prefer Trump over Hillary, Occam's Razor suggests that a sizable majority of independents who back Bloomberg in a three-way race would go for Trump if it was just between him and Hillary.
Fortunately for Trumpians, the markets don't think Bloomberg is going to throw his hat in (currently 4:1 odds that he won't). Unless Bernie pulls off an upset--and he's running through a potentially fatal gauntlet right now, with South Carolina today and a slew of heavily black southern states on Tuesday--that money seems smart to me.