The NHL playoff format explained, from points to the Cup

Two teams play sixty minutes of hockey, the horn goes with the score level, and the league hands out three points instead of two. That is where any honest account of the NHL playoff format has to start, because the bracket everybody argues about in April is filled by a standings table that runs on arithmetic no other major North American league uses.

A win is worth two points, whether it arrives in regulation, in overtime or in a shootout. A loss after sixty minutes is worth one. A loss inside sixty minutes is worth nothing at all. So a game that ends in regulation distributes two points between the two clubs, and a game that reaches overtime distributes three. The extra point is manufactured by the clock reaching zero with the score tied.

Everything downstream of that inherits it. The playoff race, the tiebreakers, the wild card cut line, the seedings that decide who hosts game seven. You cannot read an NHL standings table honestly without understanding that not every game is worth the same amount.

• 2Points for a win, in any manner
• 1Points for a loss after regulation
• 2Points shared out in a game that ends in regulation
• 3Points shared out in a game that reaches overtime

Rule-defined values from the league's points system, not figures from any season.

The three-point game, and why the standings do not add up

Add up every point in an NHL season and the total will not equal twice the number of games. It will be higher, by exactly the number of games that went past regulation.

This is a deliberate design, and it arrived in stages. For most of the league's history a tie was a tie: both clubs took a point, two points went out, and overtime either did not exist or was a short period that usually settled nothing. The problem was obvious to anyone watching. A club that reached overtime had already banked its point, and had every reason to protect it rather than risk a fourth goal against.

The fix was to guarantee the point rather than remove it. From the end of the nineties, a club that lost in overtime kept the point it had earned by getting there, which meant the extra period was played for a second point at no risk to the first. Attack in overtime and you might gain. Concede and you lose nothing you already had. Overtime got noticeably more open, which was the point.

Then the shootout removed ties from the regular season altogether, and the structure locked into its current shape: every regular season game produces a winner, the winner takes two, and the loser takes one if the game got past regulation and nothing if it did not.

The loser point has a formal name in the box score, the overtime loss, and an informal one in every arena in North America. Both describe the same thing. A point awarded for not losing quickly enough.

Because the totals are inflated, raw points are not comparable across eras and are not comparable within a season between clubs who have played different numbers of games. The league's answer is points percentage: points earned divided by points available, where points available is twice the number of games played. A club on ninety-six points from eighty-two games has a percentage a little under six tenths. That figure, not the raw total, is what the standings actually sort on when games played are uneven, and it is the first tiebreaker in the rulebook for exactly that reason.

What the loser point does to a playoff race

Sixteen of the thirty-two clubs qualify. Half the league. That sounds generous until you notice that the last two places in each conference are usually decided by a handful of points across eighty-two games, and that the loser point is the reason the gaps are so small.

Every overtime game in the league quietly compresses the table. A club that loses eight games in overtime and eight in regulation has sixteen defeats and eight points to show for them. A club with the same number of wins but sixteen regulation defeats has nothing. Those eight points are frequently the difference between a wild card place and an early summer, and none of them were earned by winning anything.

The consequence for a chasing club is worse than it looks, because your result and your rival's result are not variables of equal weight. Here is an invented scenario, with round numbers chosen so the arithmetic is visible. Club A trails Club B by four points with ten games left, and both play tonight.

Club A's result	Club B's result	Gap after tonight
Win in regulation	Loss in regulation	2
Win in regulation	Loss in overtime	3
Win in regulation	Win, any manner	4
Win in overtime	Loss in overtime	3
Loss in overtime	Loss in regulation	3
Loss in regulation	Loss in overtime	5

Read down the second column. Club B losing is not one outcome, it is two, and one of the two costs it almost nothing. A chasing club needs its rivals to lose inside the first sixty minutes, which is an event it cannot influence and which happens in only some of the games it needs.

That is why March in the NHL produces scoreboard-watching of a specificity you do not see in other leagues. A supporter is not hoping a rival loses. A supporter is hoping a rival loses in regulation, which is a materially different and rarer wish, and it is why the phrase "three-point game" is part of the ordinary vocabulary of the sport. When two clubs directly above you play each other and the game reaches overtime, the pair of them have taken three points off the board between them, and you took at most two.

Clubs on the wrong side of the line know all of this in February, which is one of the things that drives the trade market. A front office looking at its own points percentage and a schedule full of divisional opponents is making the buy-or-sell decision against this arithmetic, and the way the deadline market prices a club's real chances is downstream of a standings system that keeps almost everybody mathematically alive for longer than they deserve.

The objection: it pays to survive to overtime

The complaint is not that the loser point is unfair. It is that it changes how the last five minutes of a tied game are played, and not for the better.

Work through the incentives from inside a tied game in the third period. Both clubs currently hold one point each in escrow, guaranteed by the clock. Pressing for a winner in regulation puts a second point on the table but also risks the first, because pushing numbers forward is exactly how a team concedes. The gain from a regulation win over an overtime loss is one point. The cost of a regulation defeat compared with an overtime loss is also one point. That looks symmetrical, and it is not, because the two outcomes are not equally likely: a club that opens up against a locked-in defensive structure is more likely to concede than to score, especially against an opponent doing the same sums.

So both benches back off. The last three or four minutes of a tied game become a mutual agreement not to lose, the puck spends its time along the walls, and the crowd waits for the horn. It is the least interesting hockey the sport produces, and it exists because the rules pay for it.

There is a counterargument, and it is stronger than the objection usually allows.

The compressed standings are a product, not an accident. The loser point keeps more clubs within touching distance of the cut line in March, which keeps more buildings full and more broadcasts meaningful in a league where a large share of revenue is local. From the league's side, a rule that turns most of the membership into contenders in February instead of only the sixteen who will actually qualify is doing its job. That revenue is not an abstraction either. It feeds directly into the ceiling every club has to build under, which is how the cap is derived from what the league earns.

There is also a fairness argument. A club that plays sixty-five minutes without being beaten has demonstrably done something a club beaten in regulation has not. Rewarding it with nothing treats a coin-flip loss and a comprehensive one as identical, which is its own kind of distortion.

The obvious fix is on the table and has been for years: make every game worth three points. Three for a regulation win, two for a win after regulation, one for a loss after regulation, nothing for a regulation loss. That is the system most international and European competitions use, it removes the three-point game entirely, and it restores the incentive to win inside sixty minutes because the third point only exists there. The NHL has not adopted it. The reasons offered are usually about continuity of the record book and about not wanting to give up the compression the current system produces, which tells you the compression is regarded as a feature rather than a defect.

Regulation wins are the league's own answer to the loser point

The league knows what the loser point does. You can tell, because the tiebreaker order is built to undo it.

When clubs finish level on points, the standing between them is settled in this sequence:

• Fewer games played. A points percentage test in disguise. It exists because postponements happen and because the table has to mean something in March when clubs are two or three games out of step with each other.
• Regulation wins. Games won inside sixty minutes, with overtime and shootout wins excluded entirely.
• Regulation plus overtime wins, usually written as ROW. The same figure with overtime wins added back, so only shootout wins are stripped out.
• Total wins, in any manner.
• Points earned in the games between the tied clubs. If those clubs have not played each other an equal number of times at each rink, the first game in the city with the extra home game is thrown out of the calculation, so the comparison is even.
• Goal difference across the whole season.
• Goals scored across the whole season.

Two things in that list deserve more attention than they get.

The first is the order of regulation wins and ROW. Regulation wins sits above ROW, which means a club that racks up overtime wins is penalised relative to a club that finished its games in sixty minutes, even though both were worth the same two points at the time. ROW existed first as the senior tiebreaker, to stop shootout specialists climbing the table on the back of a skills competition. Putting regulation wins above it extends the same logic one step further and says, plainly, that the league would rather reward clubs that win before the format starts inventing points.

The second is the treatment of shootout goals in the last two criteria. A shootout victory is recorded as a single goal for the winner, and a shootout defeat as a single goal against the loser, so a shootout cannot inflate goal difference by the number of pucks that actually went in during it. It is a small piece of housekeeping that stops a tiebreaker being gamed.

Here is what that means in practice, with two invented clubs constructed so their points are identical.

• Club A
• Club B

Regulation wins4032

Wins after regulation410

Losses after regulation812

Regulation losses3028

Both invented clubs play 82 games and finish on 96 points. Club A has 44 wins and 8 losses after regulation, Club B has 42 wins and 12. The first tiebreaker after games played is regulation wins, which Club A takes comfortably.

Show the numbers

Worked example: two invented clubs, level on points, split by the tiebreaker

Item	Club A	Club B
Regulation wins	40	32
Wins after regulation	4	10
Losses after regulation	8	12
Regulation losses	30	28

Both invented clubs have ninety-six points. Club B has fewer regulation losses, which looks like the better season by one reading, and its supporters will have spent the winter being told the club was hard to beat. Club A has eight more regulation wins, and under the rules that is the only comparison that matters. Club A gets the higher placing, home ice in round one, and the softer half of the bracket.

This is why regulation wins is a live column on the scoreboard in the last three weeks of a season, and why coaches will occasionally pull a goaltender in a tied game with under two minutes left. The tie is worth a point. The regulation win is worth two points and a tiebreaker, and if the empty net costs you the game you drop only from one point to none. Under a 3-2-1-0 system that call would be automatic. Under this one it is a judgement, and the judgement is finer than it looks.

Half the league qualifies, and the NHL playoff format decides which half

Thirty-two clubs. Two conferences, East and West, sixteen clubs in each. Two divisions inside each conference, eight clubs in each division. Sixteen playoff places.

The places are filled in two passes, and the order of the passes is the whole story.

The first pass is divisional. The top three finishers in each of the four divisions qualify, and nothing outside their own division can take those places away. Twelve clubs are in.

The second pass is by conference. The two clubs with the best remaining records in each conference take the last two places, wherever they finished in their division. Those are the wild cards. Four more clubs are in, and the sixteen is complete.

38%13%50%

• Top three in each of the four divisions12
• Wild cards, two per conference4
• Clubs that miss out16

Structural counts under the current format. Twelve places are decided inside a division, four are decided across a conference, and half the league goes home.

Show the numbers

Where the sixteen places come from

Item	Value
Top three in each of the four divisions	12
Wild cards, two per conference	4
Clubs that miss out	16

The consequence of doing it in that order is that division strength is unequal by construction. A strong division can place five clubs in the postseason, its top three automatically and both of its next two as wild cards. The other division in that conference then sends only three. No rule prevents this and no mechanism corrects it, and it happens often enough that it is not a curiosity.

Note also what the NHL does not have. There is no play-in round, no eliminator between the eighth and ninth placed clubs, nothing resembling the extra week basketball has added at the bottom of its bracket. When the eighty-second game finishes, the sixteen are the sixteen. A club that misses the last wild card place by a single point, earned by somebody else for losing in overtime in December, is finished.

The divisional bracket, and how the wild cards cross over

Now the sixteen have to be arranged, and this is where the format stops resembling a ladder and starts resembling a bracket.

Inside each conference, the four first-round series are set like this. The division winner with the better record of the two plays the wild card with the worse record. The other division winner plays the wild card with the better record. The clubs that finished second and third in each division play each other.

That crossover is the part people get wrong. The wild cards are not assigned to their own division. They are assigned by record to whichever division winner is available, so a wild card that finished fourth in the East's stronger division can be dropped into the weaker division's half of the bracket, and will spend the round playing an opponent it barely saw during the regular season.

Follow the consequence through. Suppose one division produces the conference's best record and both of its wild cards. Call it Division A, and the other Division B. The first round in that conference is A's winner against A's lower wild card, A's second against A's third, and B's winner against A's higher wild card, with B's second against B's third completing the set. Rearrange that and you find four clubs from Division A sitting in the same half of the bracket. The strongest division in the conference has been made to eat itself, and it will contribute exactly one club to the conference final.

The sharper version of the same problem is this. The bracket pairs second against third inside each division. If the second and third best records in a conference happen to belong to two clubs in the same division, those two clubs meet in round one. Not in the conference final, not in the second round. Round one, in April, with one of them finished by the end of the month.

1. 1Eighty-two gamesEvery club plays the same number, collecting two points for a win in any manner, one for a loss after regulation and none for a loss inside it.
2. 2Sort each division on pointsFour divisions of eight. Points percentage does the sorting when clubs have played different numbers of games.
3. 3Break the tiesRegulation wins first, then regulation plus overtime wins, then total wins, then the points earned in games between the tied clubs, then goal difference, then goals scored.
4. 4Twelve automatic placesThe top three in each division qualify, whatever anybody else in the conference has done.
5. 5Four wild cardsThe two best remaining records in each conference take the last places, regardless of which division produced them.
6. 6Build each conference's two halvesThe division winner with the better record faces the lower wild card, the other division winner faces the higher wild card, and second plays third inside each division.
7. 7First round, best of sevenThe higher-placed club hosts games one, two, five and seven. Four wins to advance.
8. 8Second round, best of sevenThe two survivors from each half meet. Nothing is reseeded, so an upset in round one simply hands the beaten club's slot to whoever caused it.
9. 9Conference finalThe winners of the two halves meet. Home ice goes to the club with the better regular season record.
10. 10Stanley Cup FinalEast against West, best of seven, home ice again by regular season record. Sixteen wins in total, no more and no fewer.

The structure of a full season under the current format. No standings, results or clubs are asserted.

Why the NHL playoff format is a fixed bracket rather than a seeded ladder

Once the bracket is drawn it does not move. There is no reseeding between rounds. If the lower wild card wins a first-round series, it takes the slot of the club it beat and plays whoever wins the other series in its half, no matter what the records say.

This is a real design choice, and it is worth being clear about what it costs. Under a reseeded system, the best surviving club plays the worst surviving club in every round, which is what a seeding is supposed to guarantee. Under a fixed bracket, a club can earn the best record in its conference, watch its half of the draw collapse into upsets, and still find itself facing the strongest team left because that team happens to be on its side of the sheet.

The postseason is also a genuinely fresh start in a way the seeding does not capture. Home ice is the only advantage the regular season buys, and it is a modest one. The Presidents' Trophy, given to the club that finishes with the best record in the league, guarantees home ice in every round it plays and nothing else. It is the one trophy in North American sport whose holders routinely have to explain it away.

The case for seeding purely by record, and the case against it

The argument for scrapping the divisional bracket is straightforward. Rank the qualifiers one to eight within each conference, pair first with eighth and second with seventh, reseed after every round. The best club then gets the easiest available path at every stage, which is what the eighty-two games were supposedly for. Under the current arrangement it can get the hardest one, and the reason has nothing to do with anything it did.

The argument against is better than it is usually given credit for, and it has three separate parts.

The first is that records are not comparable, because the schedule is not balanced. Clubs play their divisional opponents more often than they play the rest of the conference, and far more often than they play the other conference. A club in a weak division banks points against weak opposition and arrives in April with a record that overstates it. Seeding one to eight on that basis is not neutral arbitration, it is rewarding the accident of geography with a softer draw. The divisional bracket at least makes clubs beat the opponents they actually played most, which is the fairest available reading of what their record measured.

The second is rivalry, and it is the reason the format exists at all. A divisional bracket produces the same matchups repeatedly, year after year, between clubs that already play each other four or five times a winter and travel a short distance to do it. Familiarity is what turns a series into a feud, and a feud is worth more to the sport than a theoretically correct pairing between two clubs whose supporters have no history at all. Hockey's postseason reputation is built on those repeat meetings. A record-seeded bracket would scatter them at random and hope something grew.

The third is legibility. Sixteen clubs, four fixed paths, a picture you can print in April and follow to June. Supporters can see the road in front of their club from the first day, which is a large part of why the bracket is a piece of media in its own right. A reseeding system produces a better tournament and a worse story, and the league has decided which of those it is selling.

There is a middle option nobody has taken: keep the divisional first round and reseed from the second onwards. It would preserve the rivalries where they matter most and stop the strongest half of a conference destroying itself before the semi-final. That it has not happened suggests the fixed picture is valued more highly than the objection to it.

It is instructive to hold this up against a league that went the other way. American football seeds strictly within each conference and hands the top finisher a bye and the lowest surviving opponent in every round, and the argument there is about how much that advantage is worth rather than about who should get it. The way a strict seeding rewards the regular season shows what hockey has traded away, and what it bought with it.

Best of seven, home ice, and the 2-2-1-1-1 pattern

All four rounds are best of seven. Four wins advances, sixteen wins takes the Cup, and there is no shorter round anywhere in the structure. The first round is the same length as the final, which is unusual and which is part of why the opening fortnight of the postseason is the best television the sport produces.

Home ice follows a fixed pattern. The higher-placed club hosts games one, two, five and seven. The lower-placed club hosts games three, four and six. Written out as a rhythm it is two, two, one, one, one, and everybody in hockey refers to it by that string of digits.

Through the first two rounds, home ice goes to whichever club finished higher in the standings. In the conference final and the Stanley Cup Final it goes to the club with the better regular season record, which is not always the same as the higher seed, because those two clubs came out of different halves of the draw with different paths behind them.

What home ice actually buys is worth stating precisely, because it is not the crowd.

It is the last change. Under the rules governing line changes after a stoppage, the visiting team must put its players on the ice first, and only then may the home team respond. That single procedural detail hands the home coach the final word on every matchup in the game. He can put his shutdown pair and his checking line out against the opposition's best forwards, and the visiting coach cannot counter, because he has already committed.

The visiting coach is not helpless, but his tools are indirect. He controls who starts periods. He controls deployment after an icing, because the team that iced the puck may not change and the opposing bench can send out anyone it likes. He can load his best players onto offensive-zone faceoffs and accept whatever comes back over the boards. Mostly, though, he spends the night reacting, and over sixty minutes that adds up. The way coaches deploy lines against each other is the single most valuable thing home ice confers, and it is why a series that shifts to the lower seed's building for games three and four so often changes character rather than merely changing venue.

The 2-2-1-1-1 shape matters as well. It means the higher-placed club has the last game at home if the series goes the distance, and it means a road club that wins one of the first two games has taken the structural advantage away in a single night. That is what commentators mean by stealing a game, and the arithmetic behind the phrase is simple: the higher seed's edge is four home games out of seven, and losing one of the first two reduces it to three out of five with the decider still on its own ice.

Playoff overtime has no shootout and no clock that saves you

This is the rule that separates hockey's postseason from everything else in North American sport, and it is astonishingly simple.

If a playoff game is tied after sixty minutes, the clubs play another twenty-minute period at full strength, five skaters a side, sudden death. The first goal ends the game. If nobody scores, there is a normal intermission, the ice is resurfaced, the teams change ends and they play another twenty. Then another. Then another. There is no shootout, there is no reduction to three-on-three, there is no point at which the game is settled by anything other than a goal scored in ordinary hockey conditions.

The regular season does none of this. There the extra period is short and played at three-on-three, and if that settles nothing a shootout does. Two different rulebooks govern the same situation depending on the month, and the gap between regular season overtime and the postseason version is the largest in-season rule change any of the major leagues has.

The effects run deep.

A game can last the length of two games. Benches shorten as coaches lean on the players they trust, which means a top defensive pair can play the equivalent of a full extra game inside one night while the fourth line effectively disappears. Goaltenders face volumes of shots that have no regular season equivalent. Everyone on the bench is thinking about the next game as well as this one, because the next game is usually two days away and the fatigue is transferable.

There is also a psychological asymmetry with no parallel elsewhere. In sudden death, one mistake in your own zone ends your season. Not costs you a possession, not gives away a point. Ends it. Players describe the experience in terms nobody uses about any other part of the sport, and the hockey itself becomes visibly more conservative: fewer pinches, more pucks put deep, defencemen holding the blue line rather than joining the rush.

And it removes every excuse. A club knocked out in a shootout can tell itself the format did it. A club knocked out in the third overtime cannot, because the game was decided by exactly the thing the sport is for, and it took as long as it took. That is why an overtime goal in a series-deciding game is the most valuable single event in hockey, and why the trophy waiting at the end of it carries the weight it does.

What actually changes between game one and game seven

A best-of-seven against a single opponent is a different competition from eighty-two games against thirty-one of them, and the differences are structural rather than atmospheric.

The matchup becomes the whole game. In the regular season a coach faces a given opponent a few times a year and cannot plan around one line. In a series he faces the same twenty players for two weeks, and the tactical exercise reduces to which of your groups is on the ice against which of theirs. The home coach's last change is the lever, and both coaches spend the series trying to create favourable overlaps and avoid unfavourable ones. Faceoffs become deployment decisions as much as possessions.

Adjustments compound. Both coaching staffs watch the same video overnight and both fix something. By game three, the forecheck that worked in game one has been countered, the breakout that beat the forecheck has been scouted, and the power play's zone entry has been solved by a penalty kill that now knows which side the puck is going. This is the part of a series that most rewards the depth of a coaching staff, and it is why the same two clubs can produce four games that look nothing alike.

Special teams stop being a rounding error. The power play is the highest-scoring state in hockey by a distance, and in a tight series a single unit that solves the other side's kill can decide the whole thing. Both benches also manage their own discipline harder than they do in November, because the cost of a needless penalty in a one-goal playoff game is nothing like the cost of one in game twelve.

Depth is exposed rather than hidden. Over eighty-two games a weak third defensive pair gets sheltered by the schedule and by rotation. Over seven games against one opponent whose coach has last change, it gets hunted. The same is true of a fourth line that cannot handle a defensive-zone start. Playoff rosters get shorter because coaches trust fewer players, and that concentration is exactly what turns an injury to a top-four defenceman into a series-ending event rather than an inconvenience.

Rest and travel change shape. The 2-2-1-1-1 pattern with days between games removes the back-to-backs that define the winter and gives coaching staffs time to implement changes properly rather than sketch them on a whiteboard. It also means a triple-overtime game is not just a lost night, it is a compromised next game, and sometimes a compromised round.

Goaltending is why hockey's postseason is less predictable than basketball's

Every sport's postseason has upsets. Hockey's has more of them, and there is a mechanical reason rather than a mystical one.

Start with how a hockey game is decided. Goals are scarce, and a large share of games turn on a one-goal margin. Each individual event therefore carries an enormous share of the outcome, and the number of decisive events per game is tiny. Basketball is the opposite: a game contains a large number of small events, and the better team's per-possession edge has time to express itself inside a single night, let alone across seven of them. A seven-game series in basketball is a big sample of small things. A seven-game series in hockey is a small sample of big ones.

Now add the goaltender, who is the only position in team sport with that much unilateral leverage over the score. He faces most of the shots that matter, and the difference between a good night and a great one is a couple of saves nobody else on the ice could have made. Across a series that is a handful of events, produced by one player, and they can comfortably outweigh two weeks of territorial superiority by the other club.

The honest version of the hot goaltender idea is more careful than the broadcast one. Nobody can summon a run of exceptional performances on demand, and a goaltender who plays brilliantly for two weeks may simply be a very good goaltender getting the outcomes his process deserves. What is true, and what is enough on its own to explain the upsets, is that the range of plausible results around any goaltender's real ability is wide over a sample as short as six or seven games. The favourite is not being beaten by magic. It is being beaten by the fact that seven games is not many.

This is also why the shot and chance numbers matter more in the playoffs than the scoreline does. A club that is comprehensively outplaying its opponent and losing the series is in a genuinely different position from one that is being outplayed and losing, even though the series stands the same, and the case for measuring chance quality rather than counting goals exists precisely because two-week samples lie. Most of how the sport is now built and analysed, from roster construction down to the deployment of a fourth line in April, sits on top of this one uncomfortable fact about sample size, and it runs through nearly everything else in the modern game.

How to read a series that is level after two games

Two games each in the higher seed's building is the most commonly misread situation in the postseason, because the scoreline says nothing and the underlying picture usually says a great deal. Five things will tell you where the series is actually going, and all five are visible from the sofa.

Who won the games, and how. A road win built on a goaltender making a large number of difficult saves is a different asset from a road win built on controlling the puck. The first says the visiting club got the outcome it needed from the least repeatable part of the game. The second says it has a structural edge that will still be there in game five.

Whether the top line has been matched out of the series. Watch who comes over the boards against whom. If the higher seed's best forwards have spent two games against the opposition's shutdown pair and produced very little, that is what the last change bought, and it is about to be taken away for games three and four.

Which coach is about to gain the last change. The series is moving to the lower seed's rink, where the matchup lever changes hands. A club that struggled to create anything at home is not necessarily in trouble. It may simply have been on the wrong side of the deployment, and the road games are where you find out.

Special teams volume, not just conversion. A power play that has had six chances and scored once is in better shape than one that has had two and scored one. Chances drawn are more repeatable than finishing, and the club drawing more penalties is usually the one carrying more of the play.

Who has spent more. Overtime minutes, blocked shots, defencemen pushed past thirty minutes, a goaltender who faced forty shots twice in four nights. Depth is finite, and the club that has already dipped into its reserves is the club that will find game six harder than the standings suggest.

Take those five together and a level series stops being level. It usually has a direction, and it is usually visible two games before the scoreboard admits it.