In some games, Jeff Deslauriers looks like a bona fide NHL starting goaltender. At other times, he looks more suited to sitting on the end of the bench and cheerleading in the AHL or selling insurance.

That juxtaposition is jarring on the eyes, if nothing else.

Case in point, I don’t recall seeing Deslauriers look much better than he did in Saturday’s 3-0 loss to the Colorado Avalanche, a game in which he saved the sleep-walking Edmonton Oilers from embarrassment on the scoreboard through the first 40 minutes.

Even allowing that Deslauriers should have squeezed the 2-0 goal — he faced at least 20 shots that were more difficult at the Pepsi Center — he finished up allowing two goals on 41 shots (a .951 saves percentage) before Colorado closed the deal with an empty-netter.

Deslauriers, 25, was square to shooters. His angles were good. To my eye, he crept out of his crease further than he usually does and made himself look bigger to shooters. He covered the top half of the net better than at any time I’ve seen. He was composed. He was OK handling the puck.

In terms of mechanics, I liked him in Denver every bit as much as in shutouts against Philadelphia and Phoenix. On Saturday, at least, it was fair to ask, "Tell me again, why did Steve Tambellini see the need to throw four years and a big stack of dough at Nikolai Khabibulin?"

Well, he didn’t have to, but . . .

The problem with Deslauriers is that for all the times he’s looked like a guy capable of being Pat Quinn’s starter — like the past four games and a stretch that included a five-game winning streak on the road — he’s had periods when it looks like Andre Racicot has taken over his body.

The swings from his best performances to his worst are huge. Too often, it doesn’t even look like the same person. When his Red Light persona takes over, Deslauriers is too deep in the net, shaky on his skates, guessing and scrambling like a man having a seizure. His puck-handling is atrocious.

Suffice to say, if Deslauriers has any intention of convincing Tambellini and Quinn he can be the go-to guy, he’s got to tighten up the difference between his best and his worst games.

That conjures the cliché of "consistency," which is threadbare from over-use, but is nonetheless true. There’s got to be a middle ground between, "This guy is unbelievable" and, "Man, that was brutal." It’s a happy medium Deslauriers has to find. I meant to ask goaltending coach Frederic Chabot for his thoughts about how Deslauriers can be so good one night and so bad the next before the team hit the road. but I didn’t get to it. I’ll pick his brain on that later.

Even with those big swings, by the numbers, I’m thinking Deslauriers has probably been better than a lot of his critics thought he projected to, at least if you take into account how bad the Oilers have been.

In 32 games this season, Deslauriers is 11-18-3. He’s got a 3.02 goals-against average and a .903 saves percentage. His GAA is better than the 3.33 he registered in 10 games last season and the saves percentage is a marginal improvement on his .901 of a year ago. What Deslauriers must do is work that saves percentage above .910, which looks to me like a reasonable cut-off point for starters, and then lock old Red Light in the closet long enough to keep it there.

I think he will.

As I’ve admitted more than once, I’ve been wrong on players before. It’s a sure bet I’ll be wrong again.

Not long ago, blogger Tyler Dellow was at the front of a pack of like-minded people voicing offence at my suggestion Deslauriers had "shoved it up the backsides of his critics." I never saw that coming . . .

His argument (aside from the fact my conclusion differed from what a lot of stats guys project) was that my button-pushing with the numbers crowd was based on too short a sample size — even though I said my faith in JDD was based on many seasons, as well as his hot streak.

When I did a follow-up after Racicot took over and Deslauriers went in the tank with everybody else, I was accused of doing a flip-flop. If I was really basing my opinion on the long run, Dellow asked, how could I "abandon" Deslauriers based on a short stretch of bad games?

I wasn’t abandoning Deslauriers, but, as I’ve referenced,I was certainly taken back by how he struggled so mightily with the puck — handling it, as expected, and stopping it.

Even with his swings in performance, I’m sticking with Deslauriers. Time, as it always does, will tell if I’m wrong. If I am, it’s comforting to know I’ll have company as esteemed as Dellow himself.

In July 2008, the following appeared on Dellow’s website about Ryan Miller of the Buffalo Sabres, who signed a five-year extension: "As an aside, it should go without saying, but I think that the Ryan Miller contract was ridiculous. I have him as a pretty average NHL goalie, maybe slightly above. That’s not worth $6.25MM annually.

"He’s exactly the kind of goalie who Ken Holland is talking about. Faced with a choice of him or Daniel Briere, I take Briere. Maybe that’s not fair because the Sabres learned from the loss of Drury and Briere not to let their stars walk but if Darcy Regier was talking to Kevin Lowe these days, Lowe might have been able to advise him that you don’t fix your unwillingness to pay guys in the past by overpaying lesser players."

Going into play today, Miller was tied for the lead among NHL goaltenders with a .931 saves percentage, the most telling stat for stoppers, as Dellow has often said. Miller was sixth in wins (29), third in goals against average (2.16) and fifth in shutouts (5).

A ridiculous contract? An average goalie?

That’s not a take I’ve seen referenced by those keen to congratulate Dellow (as well as each other) when he gets it right, as he often does. From where I sit, that looks like a total whiff on a blue-chip stopper who is one of the best five at his position in the NHL, no? And Briere? Oh, my . . .

It happens.