On Tuesday, Bernie Williams went 5-for-5, raising his season rates to.294 AVG/.332 OBA/.445 SLG. This is still on the weak side as far as major league right fielders are concerned. American League right fielders are batting .291/.352/.488, while their National League counterparts are batting .274/.349/.466. Nevertheless, given the depleted state of the Yankees outfield, the team can hardly afford to turn its back on any player’s hot streak. Clearly the Yankees must let Williams hit until he goes cold. The difficult question is trying to decide when that will happen, or if you’re a particularly optimistic sort, even if it will happen.

Deciding what Williams’s short-term future entails has a greater potential impact than deciding when the old man sits and Bubba Crosby rises, for the greater the faith in Williams the less urgent the acquisition of a slugging outfielder seems – at least in comparison with the need to add another starting pitcher. Williams’s hot streak is not just about a few hits and home runs, but about setting priorities for the stretch run.

After batting .333 as a 33-year-old in 2002, Williams’s bat finally withered to the point that it caught up with his rapidly declining defense. His 2003 and 2004 seasons were virtually identical to each other; combined, he hit .262/.363/.424. Williams’s patience (he walked 156 times over the two seasons) helped offset his declining power and falling batting average. If he had only retained the ability to play center field, he still would have been a valuable player.

But last season, Williams gave every impression of being finished. Batting .249/.321/.367 and playing center field with the range of an umbrella stand, Williams contributed nothing to the Yankees’ pennant pursuit. In fact, he hindered it. When he opened this season by batting .217/.262/.283 in April, it was easy to believe that it was all over.

Williams had always been a slow starter, so it was certain there would be some rebound, though given the last three years, it seemed doubtful that the bounce would be that high. Nevertheless, in June he hit .301. It was a “soft” .300, lacking power or patience, and seemed most likely a function of luck – sometimes even bad players can get one in three balls in play to fall in for a few weeks. Williams has sustained the act for long enough now that luck won’t suffice as the sole explanation; he’s legitimately hot. Williams is batting .362/.371/.672 in June, which is good enough for right field or any other position you might name.

Of course, Williams was never that kind of hitter, and it’s unlikely he’s found untapped reserves of power at age 37. In fact, Williams isn’t hitting like he used to in any way – his younger version had a higher on-base average thanks to his good eye at the plate. With 13 walks all season, Williams is no longer working the count. He’s just getting his pitch and letting it rip, doing all of his hitting on the first or second pitch of his at bats.

This approach has worked far better from the right side of the plate, where Williams has batted .364/.413/.561, than it has from the left, where he’s hitting .262/.292/.393. His left-handed numbers have been slowly rising, but his newfound impatience will suppress his on-base numbers all season long, largely neutering whatever contribution he might make against right-handed pitchers.

As a portent of the future, the lack of walks is the strongest predictor of how the Williams renaissance will end. If his average and power have revived, his eye hasn’t. Williams has drawn just two walks this month. This puts a tremendous amount of pressure on him to get it right every time – to swing at the right pitch and place it just so. That’s where luck comes back into play.

At some point, Bernie’s swings won’t be as unerring as they presently are, leading to line drives hit at fielders instead of around them and grounders that find the second baseman’s glove instead of scooting up the middle.

At that point, like all hitters, Williams will have to fall back on other skills to make a contribution. When Jason Giambi slumps, he still puts runs on the board by hitting the odd home run and by taking walks. Williams lacks Giambi’s power and has forsaken his patience. That won’t leave the Yankees much. In fact, it will leave them Williams’s April; but this time in August or September, when the damage done by a slumping player won’t be so easily shrugged off.

In order for the Yankees not to get burned, they have to remain skeptical of Williams even while cheering him and hoping that he stays hot. For the safety of the team, GM Brian Cashman and company need to view Williams’s last month and a half not as a miraculous return to form but as an unlikely encore by a star who took his final bow some time ago. They should enjoy it for as long as it lasts, but once the applause dies down they’ll need to have another act ready to go on stage.

***

After Tuesday’s column on the variability of relievers, a reader wrote in to remind me that Casey Stengel knew that relievers weren’t reliable from year to year. Indeed, he did; perhaps it was something he learned from Joe Page, who could have won a Cy Young award in 1949 (if the award had existed) but fell apart in 1950.

Late in life, Stengel told Sparky Anderson, “Young man, if you’ve got two relief pitchers, one of them will go bad next year. Get another one.” Anderson didn’t know why the statement was true, and chances are Stengel didn’t either – but it is true, which is why each winter, inept general managers like Ed Wade and Steve Phillips sign themselves a quiverfull of name relievers and generally wind up with nothing but high payrolls. There’s a reason those guys are ex-GMs.

Mr. Goldman writes the Pinstriped Bible for www.yesnetwork.com and is the author of “Forging Genius,” a biography of Casey Stengel.