Week 3 NHL Betting Report
Week three of the NHL season has been logged into the history books and it was my worst of the first three, posting a 10% loss (coming entirely from over/under, specifically when the total was 6.5). It could have been a fantastic week had I just tailed my best over/under model and best win/loss model exclusively, but sometimes vanity can be the deadliest sin. This was a massive week for my “Travelnomics” model, which comes as a bit of a surprise considering the previous iteration finished dead last in my 4th quarter model tournament. Last year’s champion “Shorting Value” crushed the early week before having a bad Saturday, giving me some bad advice in the process. When Shortsy and Travsy disagreed, I chose the wrong side.
Before we go any further, it’s time for my regularly scheduled obligatory *DISCLAIMER* it needs be noted that I’m not betting with real money. These are all fictional wagers in a spreadsheet. My mission is to engage in a mass betting campaign, picking a winner of every single game, every over/under, because it provides a complete dataset for macroeconomic analysis, which can be shared with you, shedding light on what worked and what failed. I’m also tracking the results of betting every outcome, to help me (and you) uncover previously unknown or newly emerging profit vectors. However, my sister has begun betting hockey based on my advice and I’ll be sharing all those picks with you.
My Weekly Profit: -$774
My Season Profit: -$1,324
Unlike me still playing with fake money, my sister has been betting real money based on my advice and early in the season the amounts have been very small ($1 or $2) She was breaking even in single-game outcomes until Saturday then had a bad night thanks to too many underdogs, chasing the higher payout at a point in the schedule that favored dogs in the past. I’m 0-10 on parlays and might get out of the parlay business altogether. The polls are not picking winners, so voters have as much difficulty picking from my best bets as I do. Maybe the games that are too obvious are the wrong ones to pick? The negative correlation between my confidence in a pick and the profitability is troubling, but can overcome that by sharing picks for every game.
Granted, my small bets on dogs were bigger losers than my larger wagers on favorites in the week that just wrapped (I’m counting Sundays as day one of the new week while I’m working Mondays), so that negative correlated confidence paradox didn’t hold in the last 7 days. In that case, it would have been advisable to share my most confident picks. That whole paradox is likely explained by dogs vs faves, and which is thriving any given week. Don’t start reading my pick write-ups and making giant bets when it seems like I’m unconfident. As my models continue improving, my confidence in a pick will be correlated to theirs. My emotions won’t be necessary.
Over/Under Struggles
Overs went 23-22 with a pretty even split at the 3 most common totals. All my losses came when the total was 6.5, which was only 38% of games. Keeping in mind, most of the numbers you see on these reports are based on lines very close to opening on Draft Kings. Today they opened a few games at 6.5 with a -125 line, then quickly dropoed some to 6 when the early money is too one sided. It just so happens that I also track closing lines on Draft Kings, or at least as close as I can get to closing while I have the availability to do so. Some of them are several hours before the game because I have to go to work, but there is always at least 24 hours between my two recorded values. There is a little more work to do on my line-movement worksheet before my model that bets based on line movement can be born, but it’s still on the to-do list.
Week 3 Results
*Note* “Overall Market Bets” based on betting exactly $100 on every outcome.
The autopsy of my bad week comes down to 3 strikes, under 6.5, over 6.5, and underdog moneylines. Historical precedence revealed past examples of dogs surging in late October, but alas history did not repeat. On the whole, favorites had a great week -1.5 goals, which fueled my model success, especially from road teams. It’s probably not a coincidence that my betting models collectively lost a large sum on underdogs moneyline, given they were responsible for compelling me to pick a dog more often than not. While my success betting favorites ML can be entirely credited to my intuition, with the models only posting a small gain on that category.
My Team of the Week: Los Angeles Kings, $589
The LA Kings were dead last in my profitability ranks last week, but have pulled themselves out of the basement after going 3-1. Following their loss to Ottawa in week 2, I Tweeted a “note to self: Kings might suck” when they were suddenly leaking goals and Kuemper was injured. But Big Save Dave Rittich stepped up and played well in his absence without Pheonix Copley making a single appearance, much to my disappointment. My success betting the Kings came mostly from tailing my top models, who dispensed good advice (Ottawa to win Sunday notwithstanding). Their unders went 3-1, and I went 3-1 picking the right OU bet.
My second best team of the week was the New York Islanders and the primary contributor was back-to-backs, their victory against a tired New Jersey squad, followed by their loss the Florida the next day. Both games were the rare home team with the disadvantage, which historically has a higher rate of return shorting than rested home teams against tired visitors. The Isles are 3-5 and are similar to last season, a bubble playoff team. Ilya Sorokin has been playing better with a .936 SV% through 4 games. If they continue getting elite goaltending, they’ll be a dangerous team to bet against because you never know what the goalie will steal one.
My Worst Team of the Week: Pittsburgh Penguins, -$975
The Pittsburgh Penguins climbed to #5 in my profit ranks in week 2 after a strong performance by me picking their games, but that momentum proved unsustainable. They went 0-4 on a western Canadian road trip and my betting models strongly recommended them in most of the games, dragging me into the dumpster. It certainly wasn’t a matter of me believing this team is playoff-bound, or even good. It was just listening to my advisors and following orders. I also went 0-4 on their over/under, with the total set at 6.5 for all their games. They went over that 3 times, with Pittsburgh making the biggest contribution to my struggles with 6.5 totals.
The Tampa Bay Lightning are historically one of my best teams to wager (possibly #1 since 2019), but they were my #2 worst in the last 7 days and have replaced the Kings for dead last in my profit rankings, which will get mentioned again down in the Monday picks section. This is another instance where I can invoke the “Nuremberg defense” and claim that my error was following the orders of my betting models, freeing me of responsibility for my actions. But that won’t ease the pain of those who lost money on that advice…okay, I need to stop this Holocaust metaphor before it goes too far (it was about to). Sorry, but I might be using the following orders excuse often this season.
Team By Team Profitability Rankings
These profitability rankings are based on the sum of all my bets per team, including where the money was won or lost. Each week my new Profitability Rankings will be based on all the games in the season, not just what happened this week.
Betting Models Results
My picks this week generated an 10% loss, coming entirely from over/under. My win/loss models collectively generated a 5% profit, over/under a 1% loss. Below are their stats from week 3, including where the money was won and lost. My own stats are included in there for context, despite the humbling performance by the human competitor who sees where all the best models have their money invested before making my own picks. I need to be better at reading the tea leaves, but it’s still very early in the schedule and plenty of time to improve.
Top Model Week 3: Travelnomics
Seeing my travel-based model soaring like a jet has come as a surprise considering the previous version (then named Shorting Travel) was my worst by a mile in the 4th quarter last season. That one crashed and burned, with the replacement rebranded Travelnomics because it doesn’t always bet against the travel. Creating this model required calculating the distance travelled for every road trip for every team for the whole schedule. What game of the road trip, how far they travelled from the last match, and cumulative distance of the current road trip to date. Given last year’s struggles, I’m hesitant to trust my new top model, but the longer this lasts, the more trust will build between us. Travelnomics has a big lead for my top model on the full season after Shorting Value struggled on Saturday (sadly with me following many of its picks after a scorching Tuesday to Thursday).
My best over/under model was the “Profit Hedger”, while Shorting Goalies remains at the top for the full schedule. The Profit Hedger is a close cousin to Max Profit, and uses the same input, total model profit last 30 days when making that pick with these teams. So, if 5 models are betting under 5.5, it adds up their profit betting these teams when the total is 5.5. Except unlike Maximus, the Hedger can bet against positive profit, Max is too greedy (but is also used like an indicator variable). Aside from the complexity of calculating the input, it’s a simple concept that only sorts games into 6 different angles, 4 of those sorted according to the “Angle Heaters” model input data. Are teams betting under having success with that angle regardless of the teams?
Worst Model Week 3: Expected Goals Value
Evidently the usefulness of expected goals inputs took at hit this week, as all 3 of my models using that as a primary input registering big losses this week. The value hedger took the biggest loss, getting equally crushed betting home teams and road teams, but most of the damage inflicted by underdogs. It uses xGF% to estimate a probability of either side winning, then subtracts that from the implied probability. My “Expected Goals Last 10 Games” model was really good last week and collapsed this week, while this sucked both weeks. Hopefully as the sample size increases, the stat based models improve, as my Corsi model isn’t doing well either, but maybe I should have used 5v5 instead of all strengths.
My worst over/under model this week by a wide margin was my best at OU last week, the model simply called “Prime” who lost $3,544, deep in the red for both overs and unders. It lost money at everything, over 6.5, under 6.5, over 6, under 6, over 5.5, and broke even on 3 under 5.5 picks. I’m not going to worry about performing an official autopsy beyond “leaking from all sources” and am happy to let it continue in its current form. It could heat up later. If it continues imploding, it joins the “5 worst models” aggregate to help advise me on picks not to make. It has different commands in the 2nd quarter which is a few weeks away. I don’t really have the time for full scale overhauls.
Tomorrow’s Picks
Some of these lines and totals have moved since I made my picks this morning. If the total is different, my models may no longer like it. You can always bet the alt total, or just wait until I get better with my own picks.
FLA @ BUF
The Florida Panthers went 3-2 without Matthew Tkachuk and Aleksander Barkov in the line-up, with wins against Vegas and Boston. Tkachuk is back (with 4 PTS in 3 GP) and Barkov is nearing return. My top betting models have more total money on Buffalo, but my best model has Florida -1.5 goals and hit max bets on Panthers -1.5 in their most recent 2 games in New York. Collectively my models performed very well betting V-1.5 goals in the last 7 days (aside from Oilers today), so I’m comfortable placing a small wager on Florida -1.5 PL at +180. The only concern being that Buffalo has won 3 straight at home including wins against Dallas and Florida (who was missing Tkachuk, but now are not). The champs should be hungry to avenge that loss. My best models have 100% of their max in under 6.5 goals, so that’s my pick. They just aren’t doing well on that bet with these teams at that total in the last 30 days.
EDM @ CBJ
I’m generally reluctant to bet teams with a back-to-back disadvantage, but so are my betting models and they have 53% of their total money in the Oilers moneyline, which feels expensive at -230 for a btb, but Pickard is starting Sunday so Skinner will be in net Monday. That’s why I’m also betting the under. My top models have more on over 6.5, but Shorting Goalies has the under, so I’ll tail that position. It’s Skinner in net and the mighty Oilers offense might be a step slower on the btb. That’s my hope anyway.
NSH @ TB
Tampa is dead last in my profitability rankings, posting a loss betting them to win, lose, over, and under. They appear to be my early kryptonite with difficulty getting a read on them either way. The Preds are 30th in those same ranks, so I’m not doing well solving that puzzle either. I’d skip this one and move on to the next if that was permitted. These reports cover EVERY game tomorrow. That’s the site policy. But you’ve been warned to proceed with caution, maybe to even bet the opposite of my recommendation. I’m just tailing my hottest models with minimum bets.
TOR @ WPG
The Jets drive to go 82-0 is still on track while the Leafs are 4-5 and 1-3 on the road. Now we find ourselves sitting at a roulette table after one side has hit 8 times in a row. Do you think that makes it more or less likely to repeat on the next try? This question is where the “Law of Averages” meets the “Gambler’s Fallacy”. Unlike roulette or true 50-50 propositions, there are quasi-predictable patters in team sports streaks of all shapes and sizes. My “Going Streaking!!!” model does exactly that and has a max bet on Toronto -1.5 goals. Although it cares more that Leafs lost their last game and Jets won theirs. In the previous 3 seasons, 17 home teams were on 8-game win streaks. They went 12-5 in the next game, 14-3 on the puckline +1.5 goals. But betting $100 on all those moneylines only yielded a $95 profit because of expensive prices. Jets opened at even odds. Yet my top models have all their money on Toronto, so I’ll have to take that. Under 6.5 was a tougher choice with my models split, but that line has moved to 6. My pick would be over 6, but I bet the open.
CHI @ COL
The Chicago Blackhawks are travelling to Denver for a meeting with the suddenly resurgent Avalanche, but an Avs team that has been devastated by injuries, and playing on back-to-back nights at high altitude. They might live there, but that doesn’t mean they are completely immune from the impacts. Okay, so they are 4-1 on their last 5 home back-to-back disadvantages, but unlikely they had this many injuries for those games. Justus Annunen who won their last 3 games is starting on Sunday, meaning the struggling Georgiev should be in net Monday. What’s interesting is that 76% of my model money is on Chicago -1.5 goals at an estimated line of +431 (Draft Kings doesn’t put their alt pucklines up until later, or I just can’t find them). I’m going to put a minimum bet on that one for fun, but will probably recommend Chicago ML to my sister, which opened at +190 but has dropped to +180. Could move lower as more people realize it should be Georgiev. That’s also why I’m placing a minimum bet on over 6.5 (which has moved from +100 to -110) despite my best models preferring under 6.5.
SJ @ UTA
The San Jose Sharks travel to high-altitude for a match-up with the Utah Coyotes (if they officially announce the new name for next season, I’ll stop calling them the Coyotes). The Yotes have lost 3 in a row, which is would be a bigger concern had the Sharks not lost 9 in a row. My best models have most of their money on Utah, while my worst have most of their money on San Jose. Picking the Utah moneyline was an easy choice, but I locked in at the -270 open price and that has moved to -250. There is public money coming in on the winless Sharks. Good luck with that. The over/under total opened at 5.5 and has moved to 6. All I did was tail my best model.
CGY @ VEG
The Flames scorching start to the season has cooled, losing 3 of their last 4 games, though 2 of those were to really good teams (Carolina and Winnipeg). They are getting good goaltending with Dustin Wolf posting a .924 SV% through 4 GP. Vegas on the other hand has won 3 in a row after losing 3 in a row, scoring 19 goals in those 3 wins. This was a tough game to choose because my models have 70% of their money on Calgary, but this might be a team on the way down vs a team on the way up. I’m going to tail my models and put a minimum bet on the Flames ML at +160 (that has moved to +180), but don’t feel good about it. My worst models really love that Flames ML, while two of my top models have max bets on Flames -1.5 goals. These models are like my children, but I can’t recommend Flames -1.5 goals in Las Vegas. I do like over 6 goals considering it should be Vladar tomorrow and the Vegas offense is humming.
CAR @ VAN
The city of Vancouver will get hit with a metaphorical hurricane on Monday as Carolina comes to town (they are currently 4-1 on the road, while Van is 1-2 at home). My Travelnomics model is making a max bet on Canucks -1.5 goals (because eastern team on a long western road trip), but Canes are 5-0 covering +1.5 on the road while Canucks are 0-3 covering -1.5 at home. My top model largest position is the home moneyline so I’ll just make a minimum bet and cheer for the local team. That being said, I did just hit a max bet on Canes +1.5 in their last game and that’s my models second largest position overall, slightly behind home moneyline. It was initially tempting to try that one more time. Also taking under 6 goals.
That wraps another picks piece. I’m not giving out any parlays until my own game-to-game picks improve and increase my effectiveness choosing poll options.