Guide
NFL preseason explained: what those three games are for
The NFL preseason explained: three games, a 90-man camp, one cutdown to 53, joint practices, and what coaching staffs are actually watching for.
By CricketTaken EditorialPublished Guide22 min read
The NFL preseason is a strange product to sell. Three matches a year in which the most famous players on each roster are deliberately kept on the sideline, the results are recorded and then ignored, and the fourth quarter is contested by men most of the audience will never see again. Nothing else in professional sport asks its customers to watch a competition in which the competitors have been told not to try.
Which makes it easy to write off, and wrong to. The NFL preseason explained properly is not an exhibition schedule at all. It is a five-week selection process with a hard deadline at the end of it, and the games are the least important part of it. Something like a fifth of every roster in the league is decided during these weeks, and the decisions are made on evidence that mostly does not appear on television.
Three games now, and there used to be six
Start with the shape of the calendar, because the shape has been changing for fifty years in one direction.
Through the 1970s a club played six exhibition games before a fourteen-game regular season. From 1978 the preseason was cut to four and the regular season went to sixteen. In 2021 the preseason went to three and the regular season went to seventeen. Each time, the league moved a game from the part of the calendar that generates modest revenue into the part that generates a great deal.
Add the columns up and the joke tells itself. Twenty games, every era, without exception. The players have never once been asked to play more football. They have simply been asked to play more of the football that people pay full price to watch.
That framing matters for the argument about what to do next. Every few years somebody proposes an eighteen-game regular season, and the assumption in the proposal is always that the preseason drops to two to pay for it. If that happens, the pattern holds again, and the preseason will have gone from six games to two inside a working lifetime while the total has never moved.
The reduction has consequences the schedule change does not advertise. Four preseason games gave a coaching staff a rhythm: one game to look at everybody, two to build combinations, one dress rehearsal and one throwaway. Three games removed a rung from that ladder. Evaluation that used to be spread across four Saturdays is now compressed, which is a large part of why joint practices have become as important as they have.
- 3Preseason games each club plays
- 4Games for the two Hall of Fame Game clubs
- 1Roster cutdowns between camp and the opener
- 20Games in a club's year, preseason and regular season
The Hall of Fame Game adds a fourth fixture for the two clubs selected for it, and it is generally played on a Thursday in the window between 31 July and 6 August.
The preseason is a filter, not a competition
Here is the arithmetic that explains every strange thing about how these games are played.
A club carries ninety players through training camp. It may keep fifty-three when the season starts, and sixteen more on a practice squad. So at most sixty-nine of the ninety can be retained in any capacity, and in practice fewer, because no club fills its practice squad entirely from its own releases. Twenty-one men per club, at minimum, are leaving the building.
Now look at who those men are. The fifty-three is not seriously in doubt for perhaps forty of them. A starting quarterback, a first-round pick, a player on a large guaranteed contract, an established starter at any position: none of these people are competing for anything in August. Their places were settled in March by the accounting, and the way guaranteed money locks a roster in place does more to decide a squad than any preseason performance will.
That leaves roughly a dozen genuinely contested places on the active roster and sixteen more on the practice squad, and around fifty players in camp with a plausible claim on them. Every snap of preseason football is therefore being played by, and for, the bottom half of the roster. The starters are not resting because the games do not matter. The games do not matter to the starters because the games are not about them.
Why the starters do not play, and why that is the correct decision
The complaint about resting starters is that customers pay to watch good players and are shown bad ones. It is a fair complaint about the pricing. It is a bad complaint about the coaching.
Work through what a coaching staff would gain. It would learn nothing about a player it has watched for years and coached every day since April. It would gain no competitive information, because opponents are not showing anything real either. It would produce no result worth having, because the result is not recorded anywhere that matters.
Now work through what it risks. Any snap of full-contact football carries a chance of a serious injury, and the chance does not care whether the game counts. A club that loses a starting left tackle in the third preseason game has traded nothing for a season's worth of damage, and that trade has been made often enough, and publicly enough, that no head coach wants to be the next one explaining it.
The economics point the same way and are rarely stated plainly. A starter's value to the club is measured in regular-season and postseason availability. There is no version of the calculation in which exposing that asset in a game with no stake is defensible. Resting starters is not laziness or contempt for the audience. It is the same reasoning that governs modern fourth-down decision-making, applied to bodies rather than yards: work out what the decision is worth, work out what it risks, and stop pretending that tradition is an argument.
What has changed is that coaching staffs found a substitute for the competitive work the games used to provide, which is what made resting starters practical rather than merely tempting.
Joint practices have quietly become the real preseason
Two clubs meet at one of their facilities for two or three days and practise against each other. Offence against defence, live but controlled, with the coaching staffs agreeing in advance what will be run and how hard anybody will be hit.
For evaluating a football team this is better than a preseason game in almost every respect, and the reasons are worth listing because they are not obvious.
The reps are targeted. In a game, a coach gets whatever situations the flow of play produces. In a joint practice he can call for two-minute work, or red zone, or third and long, or a full period of nothing but the goal line, and he can run each one fifteen times against a genuine opponent rather than once.
The reps are controlled. Tempo is agreed. Quarterbacks are usually off limits. The players who cannot afford an injury take the work they need at the intensity they choose, and stop.
The reps are competitive in the way that actually tests players. Practising against your own defence for four weeks produces familiarity: the receivers know the coverage, the corners know the route tree, and everybody is subconsciously calibrated to everybody else. A different club's defence is a genuine unknown, and a young receiver who has looked good all camp against team-mates who know exactly what he runs is a different proposition on day one against strangers.
And the volume is enormous compared with a game. A starter might take a handful of snaps in a preseason match. He will take dozens across two joint practice sessions.
The consequence is that the most informative football of the preseason now happens in front of a few hundred people at a training facility on a Wednesday morning, and the least informative happens in a stadium on a Saturday night in front of a paying crowd and a television camera. That is an uncomfortable fact about the product and nobody involved has an obvious answer to it.
Camp is not the beginning, and that changes what it is for
By the time players report in late July, most of the teaching has already happened, and a reader who thinks of camp as the start of the football year will misread everything that follows.
The offseason programme runs in stages through the spring. It begins with conditioning and meetings and no football at all, moves through individual work, and finishes with organised team activities in which offence and defence work against each other without pads and without live contact. Almost all of it is voluntary, which is a word doing a great deal of quiet work in a sport where attendance is noticed. A single minicamp is mandatory. Everything else is a decision a player makes with his contract status in mind.
The consequence is that a scheme is installed in the spring, not in August. A rookie has already seen the playbook, sat in the meetings and run the plays at walking pace against a defence in shorts. What he has not done is any of it against someone trying to hurt him.
That is precisely the gap camp exists to close, and it explains the shape of the whole five weeks. The question in July is never whether a player knows the play. It is whether he can still execute it with contact, with fatigue, and with a crowd. Coaches talk about players who are excellent in shorts, and the phrase is not a compliment. It describes somebody whose spring was convincing and whose August was not.
It also explains why the route from the draft to a roster runs through these weeks rather than through April. Every club signs a large group of undrafted rookies within hours of the draft ending, and those players arrive with no guaranteed money and no claim on anybody's patience. The spring gives them the install. The preseason gives them the only live evidence they will ever get to submit, and the number of them who make rosters each year is the strongest available argument that the seven rounds of the draft measure certainty rather than talent.
What training camp is actually allowed to do
The camp itself is heavily regulated, and the regulation is the reason a modern camp looks nothing like the ones in the archive footage.
The collective agreement sets an acclimation period at the start of camp, during which players work without pads and gradually build up to full contact. Two-a-day padded practices are prohibited outright. The number of practices in pads is capped. Contact work in the regular season is capped separately and much more tightly, which is why the padded work available in camp is the only real physical preparation a club gets all year and why coaches guard it so jealously.
The safety architecture around all this was not designed in the abstract. It was written after Korey Stringer died of heat illness during a Minnesota camp in 2001, a death that changed how the sport treats hot-weather training. Medical personnel and an ambulance are required at practices. Hydration and heat protocols exist as rules rather than as recommendations. Every August camp report about a light non-padded session in the heat is describing an arrangement that was paid for.
The trade-off inside all of this is real and worth stating without taking a side. Less contact means fewer training injuries and fresher players in December. It also means less tackling practice, which shows up in the tackling. Coaches who complain that fundamentals have declined are not simply being nostalgic. They are describing an actual consequence of an actual rule, and the rule was written for reasons that outweigh it.
The Hall of Fame Game, and the one preseason fixture with a reason to exist
The preseason opens in Canton, Ohio, with a game that predates most of the rest of the structure around it.
The first Hall of Fame Game was played in August 1962, the year the ground was broken for the Hall itself, and it began as a fundraiser for the building. It has stayed attached to enshrinement weekend ever since, is played at Tom Benson Hall of Fame Stadium, and since 2017 has generally fallen on a Thursday in the window between 31 July and 6 August. One AFC club and one NFC club are selected by the league in advance of the rest of the preseason schedule, and the selection usually tries to put a club with a newly elected inductee on the field.
The two clubs involved gain a fourth preseason game, which they mostly treat as an extra week of evaluation for players who need it rather than as a fixture.
The game has been called off three times, and the list is a decent summary of the last fifteen years of the sport. In 2011 the labour dispute wiped it out. In 2016 it was cancelled at the last minute because the field had been painted in a way that made it unsafe to play on, which remains the most embarrassing thing that has ever happened in Canton. In 2020 the pandemic took it along with the rest of the preseason.
The single cutdown, and the day the preseason is really for
Everything in the preceding five weeks resolves on one afternoon in the last week of August.
Clubs go from ninety to fifty-three in one step. There is no staged reduction, no intermediate limit to hide behind, and no way to defer the awkward decisions. Every club in the league hits the same deadline on the same day, which means the entire supply of released players arrives on the market simultaneously.
What happens next is governed by the waiver system, and the sequence has a rhythm that repeats every year. Released players with limited service time have their contracts offered to the other clubs in a set order before they are free to sign anywhere. Players with enough service are released straight into free agency without passing through that order at all. Only once the claiming window has closed can clubs start assembling practice squads, which is why the whole exercise takes two days rather than one, and why the second day produces as much roster movement as the first.
For the players involved this is the point of the entire preseason. Everything they have done since reporting to camp has been evidence submitted towards one decision made in a room they are not in. Most of them find out by phone call. The ones who clear waivers and get a practice squad offer are back in the same building by the end of the week, and the sixteen jobs on the other side of that cut are what a large share of the preseason was genuinely being contested for.
- Rookies and then veterans report to campThe ninety-man roster assembles. Rookies and selected players arrive first, veterans follow, and the club has its full complement for the first time since the spring.
- An acclimation period before any padsWork begins without pads and builds towards full contact over a set number of days. Two-a-day padded practices are not permitted at any point.
- The first preseason gameStarters play little or nothing. The purpose is to get bottom-of-roster players onto film in live conditions, which is the one thing practice cannot supply.
- Joint practices with another clubTwo or three days of controlled work against an opponent, at agreed tempo. This is where the most useful evaluation of the whole preseason happens.
- The second and third preseason gamesThe compressed dress rehearsal, and then a final game handed almost entirely to players fighting for the last places.
- The cutdown to fifty-threeOne deadline, one step, every club at once. Around a third of each camp roster is released on the same afternoon.
- Waivers, then practice squadsClaims are settled in order, and only then can clubs sign the players who cleared to a sixteen-man practice squad.
- The regular season opensThe fifty-three that starts the season is rarely the fifty-three that finishes it, because the churn that began on cutdown day does not stop.
The acclimation period, the limits on padded practices and the prohibition on two-a-days are set in the collective agreement rather than by individual clubs.
What coaching staffs are actually watching
Preseason evaluation is far more specific than the general impression the broadcast gives, and it varies enormously by position.
Offensive line is the position group that most needs preseason snaps and gets the least attention for it. Pass protection is a coordination problem: five men passing off stunts and twists in a fraction of a second, and you cannot assess it in a practice where the defensive line has been told not to make it hard. Backup combinations that have never played together get their only live test in these games, and a club's real depth chart on the line is usually settled by what happens in the second half of preseason matches.
Specialists get the cleanest evaluation of anybody, and the fairest. Kicking is one of the few things in football that transfers directly from preseason to the real thing, because a forty-eight yard attempt is the same task in August as in December. A kicking competition decided in the preseason is decided on evidence.
Special teams coverage units are where most of the bottom of the roster actually earns its living, and they are the reason a club will keep a linebacker who will never play a defensive snap. A player who cannot contribute on those units has to be a genuine contributor elsewhere, and the preseason is where that gets sorted. The rewriting of kickoff coverage over recent seasons has made this evaluation harder and more interesting at once, because the alignment and coverage rules on kickoffs have changed what a coverage player is even being asked to do.
Backup quarterback is the position where preseason play is most misread. A second-string quarterback playing against a third-string defence with third-string protection is being asked to do a job that does not resemble the job he would have if he were needed. Clubs know this. The evaluation is about processing and decision-making rather than the result of any given throw, which is why a staff can be encouraged by a passer who completed half his attempts and unmoved by one who looked tidy.
And across every group, the thing being assessed is often not football ability but retention: whether a player can hold a game plan installed on Wednesday and execute it on Saturday without being told twice. That is invisible from the outside and it decides an enormous number of these jobs.
Why preseason statistics tell you almost nothing
There is a structural reason preseason numbers do not travel, and it is worth understanding rather than just being told to ignore them.
In a real game, both sides are approximately at full strength, so a performance is measured against a stable baseline. In a preseason game the baseline moves continuously. A receiver catching six passes might have done it in the first quarter against players who will start on opening day, or in the fourth against players who will be selling insurance by October. The same statistical line represents two completely different achievements, and the box score does not distinguish between them.
Sample size makes it worse. A player might get twenty meaningful snaps across the whole preseason. No conclusion drawn from twenty snaps of anything is stable.
And the tactics are deliberately uninformative. Clubs run vanilla schemes in preseason games precisely so that opponents learn nothing: base defences, simple protections, no exotic pressures, none of the situational packages that will define the season. A defensive back who looks excellent in a coverage his club will never call again has demonstrated something narrower than it appears.
What the film gives a coaching staff, and the box score does not, is technique. Hand placement, footwork, whether a player's eyes go to the right place, whether he finishes a block. Those things are visible in a single rep and they do transfer. This is why coaches sound so uninterested in preseason results and so interested in preseason tape, and it is not evasion. They are watching a different object.
Preseason games can end level, and that tells you something
Overtime was removed from preseason games in 2021, alongside the reduction to three fixtures. A preseason match that is tied after the fourth quarter simply stops.
That is a small rule with a large admission inside it. Overtime exists to produce a winner where a winner is required. Nothing in the preseason requires one, and asking players who are fighting for jobs to take extra full-contact snaps to settle an exhibition is a cost with no benefit attached to it. The league looked at the trade and concluded there was nothing to trade.
It is also a neat illustration of how differently the two halves of the calendar are legislated. The rules that decide the shape of a real match, including the way the extra period now guarantees each side a possession, are the product of decades of argument about fairness. The preseason gets the version where nobody bothers, because fairness in an exhibition is not a live question.
The preseason is where the league tests its rules
The other genuine function of these weeks is experimental, and it is the one the league talks about least.
New rules arrive in preseason play first. Officials get to enforce something for the first time in conditions where a mistake costs nobody anything real. Coaches find out how a change alters their preparation. Broadcast and replay procedures get tested. The competition committee gets to watch a rule behave in the wild rather than on a whiteboard, and to see which parts of it produce arguments.
This is more valuable than it sounds, because rule changes almost always have effects nobody predicted. A change to an alignment requirement turns out to create a coaching workaround. A change to a spot turns out to invert an incentive. The preseason is where those discoveries are cheap, and a rule that survives three weeks of professional football played by people actively looking for edges in it is a rule that has been tested properly.
Officiating crews need the same runway. Officials are graded, assigned and, at the margins, employed on the basis of accuracy, and a crew calling a new rule for the first time in a game that decides a playoff place is a bad arrangement for everybody. The preseason gives them live repetitions on the same schedule the players get.
Why the preseason still exists at all
Given the games are barely watched, largely uninformative and mostly played by people who will not be employed in a fortnight, the obvious question is why anybody bothers.
Three answers, in ascending order of honesty.
The first is evaluation, and it is real but overstated. Clubs need live football to decide the bottom of a roster, and no amount of practice replaces the specific test of a player performing under lights against strangers with his employment at stake. That is a genuine function. It does not obviously require three games.
The second is availability. Camp injuries and joint practice injuries happen, and a club that has only two evaluation windows and loses one to weather or illness has a serious problem in a system where the cut deadline does not move. Redundancy has value.
The third answer is the money, and it is the one that actually holds the structure up. Preseason games are inventory. They fill stadiums, they fill broadcast slots, and they have historically been bundled into season ticket packages at prices bearing no relation to what is on the field. The reduction from four games to three was not a gift to player welfare. It was a trade, in which a low-value fixture was exchanged for a seventeenth regular-season game worth several times as much, and the trade was made because the arithmetic was overwhelming.
Understanding that makes the whole institution legible. The preseason is exactly as long as the league's revenue model wants it to be, and every argument about its length is really an argument about what replaces it. That is also why an eighteen-game regular season keeps coming back: the same trade is available again, and it is worth more the second time.
How to watch a preseason game and actually learn something
Ignore the score entirely. It is not evidence of anything and it is not intended to be.
Watch the first quarter for who is on the field rather than what happens. The composition of the opening units tells you more about a club's real depth chart than any depth chart it publishes, because a published depth chart is a public relations document and a starting eleven on a Saturday night is a decision.
Watch the offensive line combinations across all three games. If the same five backups play together repeatedly, the staff has settled on its emergency unit. If the combinations keep changing into the final week, it has not, and the club is going into the season without knowing what it does if a starter goes down.
Watch the fourth quarter with the roster in mind rather than the game. The players on the field then are competing directly against each other for the same two or three places, and their week is genuinely about to be decided. It is the only part of the broadcast where everybody involved is trying as hard as they can.
Then watch the special teams units in every game, which is the closest thing to a job interview the preseason contains. Somewhere between six and ten places on a fifty-three are held by players whose primary qualification is that they can cover kicks and punts, and those places are settled here. Follow the rest of the American football coverage into September and you will recognise those names in November, usually on the coverage unit, occasionally starting, because somebody ahead of them got hurt and a club that did its August work properly already knew who was next.
Common questions
How many NFL preseason games does each team play?
Three, and the two clubs chosen for the Hall of Fame Game play four. The preseason dropped from four games to three in 2021, at the same time as the regular season expanded from sixteen games to seventeen, so the total number of matches a club plays did not change at all.
Do NFL preseason games count for anything?
Not in the standings, not for playoff seeding and not for any individual record. What they count for is roster places: the results of preseason games are irrelevant, but the performances inside them decide which players survive the cut to fifty-three and which end up on a practice squad or out of the league.
Why do NFL starters not play in the preseason?
Because there is nothing for them to win and a season to lose. A starter's place is not in question, the coaching staff already knows what he can do, and every snap he takes carries a real chance of an injury that costs the club games that actually count. Joint practices against another club now supply the competitive work that the games used to.
Can an NFL preseason game end in a tie?
Yes. Overtime was removed from preseason games when the schedule was cut to three, so a game that is level at the end of the fourth quarter simply ends there. Nothing in the standings depends on it, and playing an extra period to settle an exhibition would mean asking players fighting for jobs to take additional snaps for no purpose.
When do NFL teams cut down to 53 players?
On a single league-wide deadline in the last week of August, after the final preseason game. Clubs carry ninety players through camp and go to fifty-three in one step, on the same afternoon across the league, and the players released then pass through waivers before practice squads can be assembled a day or two later.
Filed under American Football·nfl · preseason · roster rules · training camp · schedule