Bringing back Fernando Tatis for 2010 was never a bad idea, as Ted Berg has noted this offseason, but it became especially attractive after the Gary Matthews Jr. (GMJ) acquisition. Tatis can play the corner outfield and infield positions ably. This flexibility would make him more valuable than signing an outfield-only player like our beloved Endy, or an infield-only player like Ryan Garko. The feasibility of signing a second base free agent like Felipe Lopez or Adam Kennedy would increase, despite Luis Castillo's presence. Take a look at this position player roster, which assumes Castillo remains with the team:
Each position is sufficiently covered. Five outfielders, a cheap Tatis/Murphy platoon at first base and passable backups at all infield positions. By the time Carlos Beltran returns, either someone will get injured or GMJ will become release bait.
The New York Post noted that the Mets might pursue Tatis if Carlos Delgado is not signed. As awesome as it is watching Delgado hit 430 foot bombs, he's a one dimensional player at this point. His projected negative defensive value makes him little more than a DH. What Tatis lacks with the bat, compared to Delgado, he makes up for with defense and versatility.
As for the double play "issue" from 2009 -- Tatis had the highest clutch score on the Mets in 2008 (per Fangraphs), and the 3rd lowest in 2009. Did he lose mental fortitude between the seasons and become a choking double play master? No, not at all. It's just the kind of thing that's impossible to predict, as far as I know, and is heavily dependent on luck and other factors we cannot understand. Tatis was indeed "unclutch" in 2009, but did anyone out there predict that before the season? The double plays seem like an anomaly. After being paid $1.7 million in 2009, 35 year-old Tatis shouldn't command a deal pricier than that. His old friend Omar Minaya would be wise to re-sign him.