Poll
No votes (0%) | |||
No votes (0%) | |||
No votes (0%) | |||
No votes (0%) | |||
1 vote (6.25%) | |||
1 vote (6.25%) | |||
2 votes (12.5%) | |||
7 votes (43.75%) | |||
1 vote (6.25%) | |||
4 votes (25%) |
16 members have voted
To start, I want to examine a theory I have that in the regular season the goal is not just to win, but there is a secondary goal to at least tie. Why? Who advances to the playoffs is determined by the teams with the most points. Points are awarded in the regular season as follows:
Win = 2 points.
Lose in overtime = 1 point.
Lose after third period = 0 points.
My first question is why is there this bonus for making it to overtime? I know of no other sport that does this. I think this would give teams tied late in the third period to play conservatively, killing the clock, to make it to overtime. If a game goes to overtime, three total points are awarded between the two teams, 50% more than if it doesn't.
My hypothesis is that in the regular season more games go into overtime than would be statistically expected.
My first piece of evidence is simple. The probability of a regular season game going into overtime is 22.5%. In the playoffs it is 21.1%.
Is this just by chance and explained by a small difference compared to sample size?
I'll stop for now.
I welcome thoughts on other angles to test my hypothesis. Of the advanced statisticians of the forum, isn't there a test specifically for this kind of situation? Perhaps an F test? I learned of the F test 35 years ago (man I'm old!) and hardly remember it at all. Any other angles? Any comments among the hockey fans of the forum, just based on observation? For example, the game is tied with 3 minutes to go, does the game slow to a crawl?
The question for the poll is which is your favorite team in the western division?
Quote: unJonI love this. I’m not sure how detailed the data you purchased is. But I think it would be very interesting to investigate the effect of pulling the goalie when down 1 goal, since losing by 2 is no different than losing by 1, but getting to overtime gains an extra point in the standings. That might have an interesting effect on the “puck line” for otherwise evenhanded games.
Thanks! I was hoping to find the third period had an unusually low number of average goals, as evidenced by both teams killing the clock in a tie game. However, it isn't the case. Here are the average goals by period in the regular season.
Period 1: 1.725957257
Period 2: 2.010017809
Period 3: 2.015138023
Overtime: 0.225289403
I find it interesting that period 1 is as low as it is. I'm open to discussion about why that is.
You can see periods 2 and 3 are nearly the same. What I think may be going on is the clock killing effect to get into overtime is offset by empty net goals in the last minute. My data does not show anything on empty net goals, unfortunately.
Quote: WizardThanks! I was hoping to find the third period had an unusually low number of average goals, as evidenced by both teams killing the clock in a tie game. However, it isn't the case. Here are the average goals by period in the regular season.
Period 1: 1.737582781
Period 2: 2.020695364
Period 3: 2.017935982
Overtime: 0.112720751
I find it interesting that period 1 is as low as it is. I'm open to discussion about why that is.
You can see periods 2 and 3 are nearly the same. What I think may be going on is the clock killing effect to get into overtime is offset by empty net goals in the last minute. My data does not show anything on empty net goals, unfortunately.
Does it show the time of the goal? You could probably set up an instrumental variable for goalie pulled based on the score and time remaining in 3rd period.
Playing for a tie in the NHL is definitely a "thing".
The one "loser" point is valuable when trying to make the playoffs.
Quote: unJonDoes it show the time of the goal? You could probably set up an instrumental variable for goalie pulled based on the score and time remaining in 3rd period.
The data does not show that.
That regular season, they played twenty two ties in eighty games.
More than 25% of the games ended in overtime. Hopefully the NHL never returns to those days.
Quote: WizardAccording to morehockeystats.com, in the regular 2020 season, there were 128 goals made and 257 goals allowed in empty net situations. That is an average of 0.159354 empty net goals per game.
Whats the difference between goals made and goals allowed?
Quote: WizardThe data does not show that.
What’s the summary description of the data?
Quote: billryanQuote: WizardAccording to morehockeystats.com, in the regular 2020 season, there were 128 goals made and 257 goals allowed in empty net situations. That is an average of 0.159354 empty net goals per game.
Whats the difference between goals made and goals allowed?
Goals you scored on offense after going empty net = goals made.
Goals you allowed on defense after going empty net = goals allowed.
I guess I missed it, but what was the poll question of the thread?
And this year, every game was a division game.
I think this is a great project, and I can help with whatever you need.
There was a story I read once of a guy who “solved” hockey betting. But then he died and no one could read his notes. I can’t remember where I read it. Was it here?
The other big factor I know in hockey betting is you sometimes have situations where a team who played yesterday is playing against a team with 1-2 days rest. Of course it’s factored into the odds, not sure about the total.
2. This research paper contains data from a decade earlier on the timing of hockey goals in a game -including looking at the increase in scoring due to the effect of power plays.
hockey goals timing
3. The above paper seems to infer that scoring in the first period is lower due to the fact that in the first two minutes of the game it is relatively common for the teams to always be at equal strength.
Quote: gordonm888
3. The above paper seems to infer that scoring in the first period is lower due to the fact that in the first two minutes of the game it is relatively common for the teams to always be at equal strength.
sorry to get off the subject a little but scoring in the NFL in the 1st quarter is also considerably less (see link)
I've never heard an explanation of the reason for this but I would attribute it to both teams not yet being fully warmed up
I could be wrong - not really sure
could be the same reason for NHL games - also not sure - just guessing
Edit: also, just looked at link again and scoring in the NFL is also considerably less in the 3rd quarter
https://www.boydsbets.com/scoring-by-quarter-in-the-nfl/
*
With BetMGM, there have been some more "no risk tokens" [that have some risk] to be used in Euro2020 in particular. A tie bet seems to be one of the better bets. Is it in hockey too?
Quote: WizardI just purchased three years of NHL data, for the seasons starting 2017 to 2019, and plan to use it to make a future page on betting hockey. This consists of 7,248 regular season games and 598 playoff games.
To start, I want to examine a theory I have that in the regular season the goal is not just to win, but there is a secondary goal to at least tie. Why? Who advances to the playoffs is determined by the teams with the most points. Points are awarded in the regular season as follows:
Win = 2 points.
Lose in overtime = 1 point.
Lose after third period = 0 points.
My first question is why is there this bonus for making it to overtime? I know of no other sport that does this. I think this would give teams tied late in the third period to play conservatively, killing the clock, to make it to overtime. If a game goes to overtime, three total points are awarded between the two teams, 50% more than if it doesn't.
The NASL (the predecessor to MLS) did something like this. Originally, it was 6 points for a win, 3 for a draw (plus 1 point per goal, up to 3 per game). When shootouts were added, rather than having games end in draws after 90 minutes, eventually it became 4 if you won a shootout and 2 if you lost one.
Quote: WizardMy hypothesis is that in the regular season more games go into overtime than would be statistically expected.
There's a saying in English soccer; play for the win at home; play for the draw away. Something similar may apply here.
Quote: WizardMy first piece of evidence is simple. The probability of a regular season game going into overtime is 11.27%. In the playoffs it is 9.03%.
Is this just by chance and explained by a small difference compared to sample size?
My "gut answer" is, to a degree, yes. It's like in football; if the game is tied late, teams "take a knee" and run out the clock in regulation rather than press for a win and take a risk that it backfires on them - and remember that there is no "bonus" for losing a game in overtime in football.
Quote: lilredroostersorry to get off the subject a little but scoring in the NFL in the 1st quarter is also considerably less ... also, just looked at link again and scoring in the NFL is also considerably less in the 3rd quarter
I attribute this to the fact that the first and third quarters start with the ball in lousy field position. The second and fourth quarters will start with the ball in a random position.
Quote: ThatDonGuyMy "gut answer" is, to a degree, yes. It's like in football; if the game is tied late, teams "take a knee" and run out the clock in regulation rather than press for a win and take a risk that it backfires on them - and remember that there is no "bonus" for losing a game in overtime in football.
We see something similar in basketball. If a team is down by two points, with seconds on the clock, they usually go for a two-point shot and go to overtime as opposed to taking their chances with a 3-point shot and the win. This is why you don't see many games with a one-point margin of victory in basketball.
That’s partially it. It’s also that teams will use time outs to try to score near the end of the half or the game.Quote: WizardI attribute this to the fact that the first and third quarters start with the ball in lousy field position. The second and fourth quarters will start with the ball in a random position.
Let's have a look at betting on the underdog vs. favorite. All bets are against what is initially a 10-cent line and one unit bet on all games, including favorites. The "win" column includes the return of the original wager. As expected, the house edge on underdogs is much less. In fact, you can expect to lose 5.5 times as much betting favorites compared to underdogs.
Type | Win | Games | Return | House edge |
---|---|---|---|---|
Underdog | 3861.26 | 3887 | 99.34% | 0.66% |
Favorite | 3744.32 | 3887 | 96.33% | 3.67% |
Pick | 70.27 | 72 | 97.59% | 2.41% |
Total | 7675.84 | 7846 | 97.83% | 2.17% |
How about road vs. home teams? In the NFL I have found road teams do a little better against the spread. That is the case here too. The house edge is 10.3 times higher on home teams.
Type | Win | Games | Return | House edge |
---|---|---|---|---|
Road | 3908.00 | 3923 | 99.62% | 0.38% |
Home | 3767.84 | 3923 | 96.04% | 3.96% |
Total | 7675.84 | 7846 | 97.83% | 2.17% |
Next, over or under? The win column is also based on what starts as 10-cent lines. Here we see a very slight player advantage betting the under.
Type | Win | Games | Return | House edge |
---|---|---|---|---|
Over | 3826.96 | 3923 | 97.55% | 2.45% |
Under | 3924.02 | 3923 | 100.03% | -0.03% |
Total | 7750.98 | 7846 | 98.79% | 1.21% |
So, in conclusion, the underdog and under basic strategy definitely holds true in the NHL.
You might wonder why the combined house edge is less on over/under bets than on the money line. I think this is partially because there are 4.3% pushes on over/under bets, which have a 100% return. Also, over/under bets are always close to -105/-105. Money line bets tend to be spread further apart, which decreases the overall return. Still, it seems fishy to me and I'll look more carefully at it.
Quote: WizardI attribute this to the fact that the first and third quarters start with the ball in lousy field position. The second and fourth quarters will start with the ball in a random position.
not quite sure what you mean
the ball is kicked off the tee - so if there is a good return there is not lousy field position
maybe you're saying most of the time there is not a good return and they start around the 25
.
Teams will let the clock expire on offense inside the red zone at the end of the 1st and 3rd, it would almost never happen at the end of the 2nd and 4th.
Quote: lilredroostermaybe you're saying most of the time there is not a good return and they start around the 25
.
Yes, that is what I'm saying. It seems to me the receiver usually takes a knee and they start at the 20.
Quote: WizardThanks! I was hoping to find the third period had an unusually low number of average goals, as evidenced by both teams killing the clock in a tie game. However, it isn't the case. Here are the average goals by period in the regular season.
Period 1: 1.737582781
Period 2: 2.020695364
Period 3: 2.017935982
Overtime: 0.112720751
I find it interesting that period 1 is as low as it is. I'm open to discussion about why that is.
You can see periods 2 and 3 are nearly the same. What I think may be going on is the clock killing effect to get into overtime is offset by empty net goals in the last minute. My data does not show anything on empty net goals, unfortunately.
I would say Period 1 is the lowest for three possible reasons, or a combination of the three and there are probably more:
1.) The teams having to, "Feel each other out," in order to perhaps find holes in the defense.
2.) The goaltender is fresh during the first period.
3.) You don't start the first period of a game with a Power Play, but you sometimes start the second and third periods with one.
---Good observation on the third period/overtime.
Quote: WizardI just purchased three years of NHL data, for the seasons starting 2017 to 2019, and plan to use it to make a future page on betting hockey. This consists of 7,248 regular season games and 598 playoff games.
To start, I want to examine a theory I have that in the regular season the goal is not just to win, but there is a secondary goal to at least tie. Why? Who advances to the playoffs is determined by the teams with the most points. Points are awarded in the regular season as follows:
Win = 2 points.
Lose in overtime = 1 point.
Lose after third period = 0 points.
My first question is why is there this bonus for making it to overtime? I know of no other sport that does this. I think this would give teams tied late in the third period to play conservatively, killing the clock, to make it to overtime. If a game goes to overtime, three total points are awarded between the two teams, 50% more than if it doesn't.
That point did not always exist and hadn't started until 1999, so you might have seen more third period scoring at that time as both teams would have probably just been trying to end it prior to the rule being instituted.
In fact, there used to not be overtime at all until the early 80's. Prior to overtime, the game would just end in a tie at the end of regulation. It's perhaps because of that they went with the one point rule for an overtime loss. Prior to the institution of overtime, a team would have received that point anyway without the need to play overtime.
Quote:My hypothesis is that in the regular season more games go into overtime than would be statistically expected.
My first piece of evidence is simple. The probability of a regular season game going into overtime is 11.27%. In the playoffs it is 9.03%.
Is this just by chance and explained by a small difference compared to sample size?
I think both are probably what one would, "Statistically expect." In the playoffs, there's simply no real benefit to going into overtime as opposed to trying to play aggressively and just win the game right there in the third period. As you have mentioned, there's an incentive just to get to overtime during the regular season.
Quote:If we assume the 11.27% in the regular season is to be expected in the playoffs, then the standard deviation in overtime games in 598 playoff games would be 7.73. We would expect to see 598*0.1127 = 67.41 playoff games go into overtime. Instead, only 54 did. That is 1.73 standard deviations less than expectations. The probability that 54 or less games in OT is 4.15%.
I disagree and would maintain that no probability can be stated. There's simply no reason to want to go to overtime in the playoffs.
okay, some good answers from Mr. Wizard & Bill on that one
the link also shows that the home team outscored the visitors by more in the first half than they did in the 2nd half every single tracking
what's the reason?
the obvious answer is that the winning home team slowed the game down by running the ball in the 4th quarter
but is that really the answer? because a lot of times the team that won came from behind to win
and if they were only up by 3 points or less I doubt they would slow it down in the 4th quarter
surely they did sometimes slow the game down - but if they were down in the 4th quarter then they would speed things up
and the home team wasn't always the fave - and home teams don't dominate in the NFL like they do in the NBA - I'm pretty sure they only win about 55% of the time
https://www.boydsbets.com/scoring-by-quarter-in-the-nfl/
.
Quote: Mission146I disagree and would maintain that no probability can be stated. There's simply no reason to want to go to overtime in the playoffs.
I'm not sure I see your point here, especially about no probability can be stated.
Let's say I shoot a free throw 1000 times and make it 250. No other data is available. Can I now say that my probability of making a free throw, based on the evidence at hand, is 25%?
Quote: WizardQuote: Mission146I disagree and would maintain that no probability can be stated. There's simply no reason to want to go to overtime in the playoffs.
I'm not sure I see your point here, especially about no probability can be stated.
Let's say I shoot a free throw 1000 times and make it 250. No other data is available. Can I now say that my probability of making a free throw, based on the evidence at hand, is 25%?
I should have said no absolute probability can be stated in such a limited sample size. My main point was that I'd expect OT to have a lower probability in the playoffs simply because there's no incentive to make it to OT, but there's no way to quantify what the probability, 'Should be.'
There definitely will be more OT games in the regular season because BOTH teams have an incentive at the end of a tied game to NOT LOSE. To play more conservatively.
Goals per minute in hockey are higher during power plays than regular 5 on 5. The likelihood of the actual two minutes of power play occurring in the first period is lower than the likelihood in periods 2 and 3. Additionally, in games that have an empty net situation, there are far more goals per minute than non empty net situations.
Also went to a VGK playoff game last month (game 6 vs. Avs). Awesome experience!
Quote: tringlomaneInteresting stuff Wiz. I decided to toss a couple "free bets" on the canadiens ml and u4.5 tonight. Since I don't get the value of the stake if i win the free bet, better to go with underdogs anyway.
Thanks! May the odds be ever in your favor.
Getting +1.5 goals on the underdog: 98.06%
Laying -1.5 goals on the favorite: 97.42%.
Not surprisingly, betting on the underdog is the better value, much like in the NFL betting against the spread.
Quote: SOOPOOThere definitely will be more OT games in the regular season because BOTH teams have an incentive at the end of a tied game to NOT LOSE. To play more conservatively.
Thank you. I'm wondering if I should make a big deal proving the data shows this to be the case, or do you think to most hockey fans the overtime incentive in the regular season is obvious?
Quote: WizardThank you. I'm wondering if I should make a big deal proving the data shows this to be the case, or do you think to most hockey fans the overtime incentive in the regular season is obvious?
I’d be interested to see if division games have a different OT likelihood.
Quote: WizardThank you. I'm wondering if I should make a big deal proving the data shows this to be the case, or do you think to most hockey fans the overtime incentive in the regular season is obvious?
My two cents says that it is obvious to those that watch hockey. I don't watch a lot of hockey, maybe 100 games a year, but I sure think it is obvious.
For example, in the NBA, when the #1 seeds play the #8 seeds the games tend to be less competitive when compared to later rounds.
I understand that many people (i.e. Americans?) don't like going to games and not seeing a resolution, but I think it's acceptable that teams play evenly through the regulation and overtime periods. I find it more satisfying than the manufactured resolution of a shootout. This scenario is especially true of soccer, where they play — what, 90 minutes? — and then decide the outcome in a way that's totally different than all the strategy and beauty of the regular game.
I know the NHL decided they never wanted to see a situation like my coworkers' first hockey game, which ended in a 0-0 tie — an anomaly but a mar on the sport for casual fans.
Anyway, back to the great hockey analysis. Thanks for that!
Quote: WizardThank you. I'm wondering if I should make a big deal proving the data shows this to be the case, or do you think to most hockey fans the overtime incentive in the regular season is obvious?
Quite obvious. The announcers use the phrase ‘both teams trying to secure a point’ or some such all the time.
Quote: KeeneoneSeeding in the playoffs could potentially be one explanation for the difference in playoff/regular season overtime games. The top seeds play lower seeds in the early rounds resulting in less competitive games?
For example, in the NBA, when the #1 seeds play the #8 seeds the games tend to be less competitive when compared to later rounds.
I think that gets balanced somewhat by closer seeds playing each other and good teams no longer playing teams that failed to make the playoffs.
As already mentioned, scoring increases in the third because of teams pulling the goalie. Coaches almost always wait too long to pull the goalie (similar to football coaches being too conservative on 4th down); if they ever become better at that decision, it should increase third period scoring a little more.
Teams definitely play for ties. I've always thought three points for a regulation win, two for on OT win, and one for an OT loss would have made more sense. Or even five for a win, four for an OT win, three for a shootout win, two for a shootout loss, and one for an OT loss. Having some games be worth three points and other games be worth two doesn't make sense.
Like most sports, the spread, moneyline, total are pretty efficient. Underdogs and unders may be less bad, but generally still not good for most bettors. There are lots of other bets on the board that can be are great. Laying the price on player prop unders can usually be good; been great in these playoffs. Just looking in Las Vegas sportsbooks saw a +400 on OT and -420 on no OT. That's like a 0.2% house edge, which should be beatable in sports. Could have even been better numbers outside of NV. I think the value was on the +400, but I bet the -420, because of reasons.
Another thing about betting hockey is the huge number of pushes that can be found. Totals on 5, obviously. Moneylines for a period even more so. Period totals on the 2, or team totals on the 2 or 3. There is one reason I have been able to take advantage of these, even if -EV. There might be a few others as well.
Whenever a state first legalizes sportsbetting (or a new sportsbook comes in), always check to see if someone is letting people parlay pucklines and totals. There have been a few places that have done it, as recently as 2019. Usually don't last long. I knew one place in Las Vegas that allowed it, and I bet it for more than one full season. I was young and dumb and didn't try to go scorched earth, but it was still one of the few times gambling was the majority of my income.
Heard something recently about the value of +1.5 in baseball, and might apply to hockey. For a long time (up until around 2010), there was always value betting the biggest underdogs at +1.5. One reason was a lot of players just didn't want to lay -300 on the biggest favorites. So they would lay -1.5 instead. But the bigger the favorite, the more likely they are to win by exactly one run. That's because the smaller favorites only win 55% of the time; if they are at home, that means we might expect them to win by exactly one run about 16-17% of the time. Some of the bigger favorites win more than 75% of the time; if they are at home, that means they would win by exactly one run 22-23% of the time. A change of 5-6% is a pretty big difference, when the price to buy the 1.5 was the same (the market has made some corrections). The same might be true in hockey. The most likely time a team will win by exactly one goal is when they are a very big favorite. I would theorize it is possible that this means the most value on the +1.5 is with the biggest underdogs.
However, the numbers are different. I just purchased the 2020/21 season, which was played entirely in 2021, which mixed more games in the data. I also weighted the returns to achieve an overall 95.45% return, which you would expect against a 20-cent line.
I plan to add more fun facts to the page tomorrow. For now, I welcome all questions, comments, and corrections. Thank you.
Quote: WizardPlease have a preview of my new page on Betting the NHL. So far it presents my analysis of the three major ways to bet I already addressed here.
However, the numbers are different. I just purchased the 2020/21 season, which was played entirely in 2021, which mixed more games in the data. I also weighted the returns to achieve an overall 95.45% return, which you would expect against a 20-cent line.
I plan to add more fun facts to the page tomorrow. For now, I welcome all questions, comments, and corrections. Thank you.
This last season had teams playing within their divisions and Canadian teams not traveling south. Are you concerned that will give false patterns for next season when things are normal?
Quote: billryanThis last season had teams playing within their divisions and Canadian teams not traveling south. Are you concerned that will give false patterns for next season when things are normal?
What kind of false patterns do you think it would cause?
Quote: WizardWhat kind of false patterns do you think it would cause?
The more you play a team, the more familiar you are,
The longer the road trip, the weaker most teams play
A team like the Islanders was able to bus or train to most of their games this year and often were home the same day. I'd think those results would differ from a West Coast/Vegas trip. This season, the Islanders had a couple of occasions where they played a road game, then stayed in the city a couple of days and then played another game. That never happens in a regular season.
Some divisions seem to embrace a more wide-open style of play while others are more defensive. Next year the teams will mix it up more, and have long road trips.
I think the lack of fans also had an effect. No one will ever convince me that 20,000 rabid fans have no effect on the teams on the ice.
I'm a casual fan who doesn't bet but if I did I'd be looking beyond the last two seasons if I was searching for patterns.
Quote: billryanThe more you play a team, the more familiar you are, ...
Good points. I tend to poo-poo such factors, but I am not a handicapper.
Quote: josephrevellI can say that management and logistics affect the result of the team. Long bus journeys and sluggish staff can affect the mood and motivation of the team. Good sports clubs rent hotels near the play area.
I think the home field advantage, for whatever the reason, is already baked into the line.
I have bet all the major sports
I have found:
the NHL the most difficult to win at - so difficult that I've given up trying
MLB is the 2nd most difficult - haven't quite given up yet
NBA is the 3rd most difficult
NFL is the 4th most difficult or 2nd easiest to win at if you look at it that way
NCAAB - college basketball is the easiest for me to win at
and yes, I know that I may not have made a sufficiently large # of bets for this to have any value to anyone else
but I believe it is true for me
.