The first two weeks of an NHL season tend to produce a lot of information and even more conclusions.

Some will eventually prove useful. Many will not.

A player will score six points in four games and look like a breakout. Someone drafted three rounds earlier in fantasy will start with one assist. A backup goalie will win twice. A rookie will score on his first NHL power play. By the middle of October, there will already be fairly confident explanations for most of it.

The problem is that four or five games are not very good at telling us how good a player is.

They can, however, tell us quite a lot about how teams intend to use their players.

For Fantasy Hockey Intel, that is probably the main purpose of the first two weeks.

We already have projections. Now we start replacing assumptions with observations.

Roles become real

Throughout the summer and preseason, fantasy analysis relies heavily on projected line combinations and expected roles. Once the regular season starts, some of those assumptions become actual data.

Who is really on PP1?

Who actually plays 19 minutes rather than the expected 15?

Which winger stays in the top six when the coach has a full lineup available?

Which defenseman is on the ice when his team needs a goal?

And who is on the ice when the net is empty at the other end?

Those questions can become useful surprisingly quickly.

Empty-net usage is particularly easy to overlook. Coaches tend to trust specific players when protecting a lead, and those minutes create real fantasy opportunity. They can mean extra ice time, additional shot attempts and, occasionally, relatively cheap goals and assists that still count exactly the same in fantasy.

This does not mean a line combination or deployment pattern used for three games is permanent. Coaches change things constantly. But there is a meaningful difference between guessing that someone might receive an opportunity and watching him receive it repeatedly in NHL games.

This is particularly important for players whose fantasy case depends heavily on deployment. A player does not need to score immediately for an improved role to matter. In many cases, the role change is the more useful information.

Injuries will create opportunities that did not exist in September

Opening-night rosters rarely remain intact for long. This year, some changes have already happened before the first puck has been dropped.

Florida begins without Brad Marchand. Mathew Barzal’s availability has been uncertain for the Islanders. St. Louis has already lost defensemen from its planned opening lineup, while Buffalo has had injury concerns around both Ukko-Pekka Luukkonen and Alex Lyon.

For fantasy purposes, the injured player is only half of the story.

The other question is who receives the missing minutes.

An injury can move a winger from the third line to the first, create a temporary PP1 opening or turn a nominal backup goaltender into a starter for several games. Most temporary opportunities disappear when the injured player returns.

Some do not.

A replacement can perform well enough to retain part of the role. A power-play unit can work better than expected. A coach can discover a combination that was not part of the original plan.

The first two weeks are therefore especially useful for identifying opportunity changes, even when it is too early to make strong conclusions about the players themselves.

The rookies worth watching are the ones receiving real roles

There will be a lot of rookie scoring updates during the first weeks.

The point totals are interesting. The usage is usually more interesting.

Bradly Nadeau is expected to get an early NHL opportunity in Carolina after producing more than 50 points in consecutive AHL seasons. James Hagens enters the season with a realistic role in Boston. Konsta Helenius appears positioned for regular NHL work in Buffalo after a strong AHL season and a brief but productive NHL sample. Roman Kantserov arrives in Chicago after leading the KHL in goals and has been discussed as a potential top-line option.

These are very different players in very different situations, which is precisely why simply comparing their first five point totals will tell us relatively little.

What matters initially is whether the NHL role is real.

If a rookie remains in the top six after the first round of line changes, receives meaningful power-play time and begins generating his own offense, the information becomes more useful.

If he scores three points while playing 11 minutes on the fourth line, the points still count. They simply tell us less about what should happen next.

The distinction is particularly important in dynasty leagues, where a temporary scoring streak can create a surprisingly large change in perceived value.

Goaltending will probably create the most confusion

Goaltenders are difficult enough over a full season. Two weeks does not improve the situation.

Several teams enter the year without a completely obvious distribution of starts, and injuries can complicate those plans immediately. A goalie who starts three of the first four games may be establishing himself as the No. 1. He may also simply be benefiting from scheduling or an injury to his partner.

That makes workload more informative than early save percentage.

A .938 save percentage after three starts looks impressive. It tells us relatively little about whether that goalie will make 48 starts.

If one member of a tandem consistently gets the more difficult opponents, starts both ends of favorable scheduling situations or begins receiving two starts for every one given to his partner, that is more useful.

The first couple of weeks can therefore materially alter goalie value without telling us much about actual goalie performance.

For fantasy purposes, starts are a resource. We should pay attention when teams begin distributing them differently from what was expected.

Power-play units matter more than early power-play results

Special teams are another area where results can be misleading.

A new PP1 can go 0-for-11 and remain completely intact. Another can score four times in its first eight opportunities and be changed a week later.

The important early observation is generally who is receiving the opportunity, not whether the puck happened to go in.

For defensemen in particular, PP1 access can dramatically change fantasy value. The difference between 45 seconds of secondary power-play time and two or three minutes with the first unit is large enough that it can matter before the player has recorded a single power-play point.

The same applies to forwards moving into net-front, bumper or shooting roles.

If a preseason projection was based on anticipated PP1 usage and the player is actually on PP2 from opening night, that is meaningful new information.

Four games without a power-play point is considerably less meaningful.

Watch minutes, but watch the right minutes

Average time on ice is useful, but the total can hide quite a lot.

Two forwards can both play 18 minutes. One receives five minutes of special-teams time and offensive-zone deployment. The other kills penalties and takes defensive-zone faceoffs.

Their fantasy opportunity is not equivalent.

The early questions should therefore be more specific.

Is the player’s even-strength role expanding? Is he playing with the team’s best offensive players? Is he receiving power-play touches? Is he on the ice when his team is trailing late?

Those situations are more closely connected to future fantasy production than simply noticing that average ice time increased from 16:42 to 17:31.

The goal is not to collect more numbers.

It is to determine whether the numbers describe a materially different role.

Schedule can temporarily make players look better or worse

The first couple of weeks also contain a basic problem that is easy to overlook: teams do not play the same number of games.

Some fantasy players will accumulate five or six games while others have played four. That difference can substantially affect early total-value rankings.

Schedule strength also varies.

A team playing several home games against weaker opponents can produce an early cluster of fantasy-relevant players. Another can begin with a difficult road schedule and look unexpectedly poor.

Neither needs to represent a meaningful change in talent.

This matters particularly when looking at league-wide rankings after one week. Total points will already look like a sample. In reality, part of what we are seeing is simply opportunity volume.

Rate statistics help, but rates over four games have their own obvious problems.

There is no perfect correction. Mostly, it is worth remembering that six games are more games than four.

Shooting percentage will lie to us

Someone will start the season shooting 30%.

Someone else will take 20 shots without scoring.

There will immediately be explanations for both.

Occasionally those explanations will even be correct.

Usually, two weeks is far too early to know.

This is where process data becomes useful. If a player starts scoring more while also generating more shots, more individual expected goals and better opportunities, there may be something underneath the production.

If almost nothing changes except the percentage of shots entering the net, there probably is not much reason to rebuild a season projection.

The same applies in reverse.

An offensively involved player with zero goals through six games is not necessarily struggling in a way that matters for the next 76.

Sometimes the puck simply did not go in.

Line combinations are signals, not conclusions

Fantasy managers tend to like lines because they are visible and easy to understand.

Player A is now playing with Player B.

Therefore Player A is more valuable.

Sometimes.

The important question is whether the combination survives.

Teams experiment heavily at the beginning of seasons. A coach can change a line after one bad period or leave it together for ten games despite mediocre results.

A new combination becomes increasingly interesting when it is supported by other information: increased ice time, offensive-zone usage, power-play involvement and improved individual process.

The line itself is a signal.

The surrounding evidence tells us how seriously to take it.

The waiver wire will contain real opportunities and fake ones

This is probably the most practical part of the first two weeks.

Some players who were barely drafted will acquire legitimate fantasy value.

The difficulty is separating them from players who merely scored early.

A useful early-season waiver candidate usually has something structural behind the production: a new role, more minutes, PP1 access, an injured teammate, increased shot generation or a roster change that creates a path that did not previously exist.

The less interesting version is a player with four points on six shots while playing exactly the same role he had before.

Both can help a fantasy team in the short term.

Only one gives us much reason to alter the player’s underlying value.

This distinction matters even more in dynasty formats, where acquiring a temporary hot hand at a permanently higher price can be expensive.

What Fantasy Hockey Intel should know after two weeks

We probably will not know who is having a breakout season.

We should know considerably more about who has received breakout opportunity.

We probably will not know which goalie tandem will still have the same workload distribution in January. We should know whether the preseason expectation was correct.

We will not know which rookies are going to score 60 points. We should know which ones have actually been trusted with roles capable of producing 60 points.

And we will have a first set of evidence on which preseason assumptions were simply wrong.

That is useful.

The objective of the first two weeks is not to replace an 82-game projection with a six-game projection.

It is to update the inputs.

Some preseason assumptions will survive unchanged. Some will need minor adjustments. A few will become obviously obsolete almost immediately.

That is what should happen.

A projection made in September should not know more about October than October does.

The useful part is recognizing which new information deserves to change it.