NHL Best Bets for Tuesday, October 6: Blues and Golden Knights Clear the Bar, the Ottawa Back-to-Back Trap, Eighteen Goalies and the Shots Board

Nine games, every one of them a one-goal line, and a model that has learned something about itself in a week: at 55% and up it is 12-5 this season, and below 55% it is 8-14. So this is not nine bets. It is two picks that clear the bar, a card that will carry four legs because that is what the card does, a back-to-back trap in Detroit that Monday night's number has not priced yet, eighteen goalies worth a look before you touch anything, and a shots board that is where the real edges live on a night like this. Numbers are Monday night's. The model rebuilds every morning and the card freezes at 11 AM ET at that morning's lines, so the live version is always on the Model Cards page.

The board

Our number is the model's win probability blended with the market, the one the record grades. Market is the vig-free probability the price implies. Edge is ours minus theirs, in points: green means we like the side more than the price does, red means the market is firmer on it than we are. Lean is the card's own tier for anything under 58%; nothing on Tuesday reaches Play (58%) or Lock (65%). Pass means the model's side is still the more likely winner but at that price we would not bet it, and the one Watch is explained below.

The two that clear the bar

St. Louis Blues over Chicago Blackhawks, 57%, -135. The firmest call on the slate, and the only one where our number and the price agree to within a point. Recent form is the biggest driver (+10.4 points), with the venue line (+5.7) and goal margin (+4.9) behind it, and the case against is the streak (-5.8): St. Louis walked out of Denver on Saturday on the wrong end of 6-1, Jordan Binnington stopping 30 of 36. The thing is, Chicago is 0-3 and being outscored by more than three goals a game, Connor Bedard is on injured reserve until November, Andrew Mangiapane is out, and Spencer Knight's last five starts read 1-4 with an .829 save percentage and 4.4 goals against a game. The Blues are missing Logan Mailloux and Tyler Tucker on the blue line and Alexey Toropchenko up front, and the model still finds +0.8 points of injury edge for them. The Blues have won four of the last six meetings. Line STL -1.5, total 5.5.

Vegas Golden Knights over Seattle Kraken, 55%, -170. The model alone has this closer to a coin flip (it makes Vegas 53%); the market makes it 60%, and the blend lands at 55%. What drives it is shot share (Corsi, +6.9), the venue line (+4.5) and the power play (+2.7); the case against is goal margin (-2.9), because Seattle is 2-1 and +2.7 a game after hanging six on Calgary on Sunday, while Vegas edged Vancouver 3-2 the same night. The goalie line is the thing to check before you bet it: DailyFaceoff projects Adin Hill, and the logs we hold have no start for him since April 9. Joey Daccord stopped 24 of 25 in his last start. If Hill is confirmed and looks like himself, the pick holds; if the Kraken get Daccord and Vegas gets a rusty Hill, the market's 60% is generous. Line VGK -1.5, total 6.5.

The card as it stands

Four legs make every NHL card, the four most likely winners on the slate, whatever the price. By Monday night's numbers that is STL, VGK, TOR and OTT, and the honest version is that two of the four are there on rank, not value: Toronto at -155 and Vegas at -170 are both priced five points firmer than we make them.

Toronto over Nashville, 53%, -155. The streak (+5.9), the venue (+2.4) and shot margin (+1.1) carry it; shot share (Corsi, -5.8) is the case against, and it is a loud one. The Leafs are 1-2 and just lost 3-2 at home to Ottawa, with Sergei Bobrovsky stopping 26 of 29; he is 0-2 with an .887 through two starts in Toronto, with Anthony Stolarz the alternative if the Leafs blink. Juuse Saros stopped 29 of 30 in Nashville's 2-1 win over Dallas on Saturday. Toronto is without Dakota Joshua and Max Domi. The last six meetings are 3-3. At 58% the market is firmer than we are; the card takes Toronto because 53% is still fourth-best on the board, not because the price is kind.

Ottawa at Detroit is the one to re-read in the morning. Monday night's number, OTT 53% against a market 49%, was computed before Ottawa's Monday game in Boston was counted, so the model's schedule features read Ottawa as rested. They are not: Boston on Monday, Detroit on Tuesday, their third road game in four nights. The second of a back to back is the single heaviest schedule feature in the NHL model, and teams in that spot have won 43.7% of their games over the last two seasons against 51.1% otherwise. When the cache rebuilds Tuesday morning expect this number to move toward Detroit, possibly through 50%, which would swap New Jersey (53%) onto the card in Ottawa's place. The hockey case for Ottawa is real (recent form +7.6, shot share +4.7, Linus Ullmark 5-0 with a .940 over his last five starts, Detroit 0-2 and without Dylan Larkin, Simon Edvinsson and Marco Kasper, an injury edge worth +1.8), and so is the case against (Jake Sanderson, Artem Zub and Andre Burakovsky out, a streak reading of -8.1). Watch the morning number; do not bet Monday night's.

The five we would pass on, and what would change our mind

New Jersey over Utah, 53%, -130. Head to head is the biggest driver (+3.7; the Devils have won the last four meetings), with rest (+2.3) and the power play (+1.1). Goal margin (-6.7) is the case against, and it is earned: the Devils were shut out 6-0 by the Islanders on Saturday. Utah is 2-1 and +2.3 a game, Karel Vejmelka has a shutout already this season and Jake Allen stopped 23 of 25 in his last start. At -130 you are paying 56.5% for a 53% team. Pass, unless the price comes to -115.

Carolina over Montreal, 52%, -125. The biggest single feature reading on the whole slate is here: shot share (Corsi) is worth +12.0 points to the Hurricanes, before venue (+4.7) and head to head (+3.1). Against: shot margin (-3.6) and an injury line (-0.9) because Seth Jarvis and Alexander Nikishin are out, with Kaiden Guhle missing for Montreal. Pyotr Kochetkov stopped 16 of 18 on Saturday; Jakub Dobes gave up six on 30 shots in a 6-5 overtime loss in Pittsburgh. A total of 6.5 is the model's way of saying this one could be loud. Pass at -125; a coin flip at the Bell Centre is not where the edge is.

Rangers over Islanders, 52%, -148. The model alone actually leans Islanders (it makes the Rangers 48.5%); the market's 57% pulls the blend to 52%. Motivation (+4.6), the power play (+3.9) and form (+2.6) for the Rangers; the streak reading (-5.9) is against them, odd on its face for a team on three straight wins until you look at the other bench: the Islanders arrive off a 6-0 win in which Ilya Sorokin stopped all 37 shots. The Islanders are without Mathew Barzal; the Rangers without Adam Sykora, Joe Veleno and Joonas Korpisalo in goal. Islanders have won four of the last six at this rivalry. Six points of gap between us and the market is the widest on the slate. Pass.

Florida at Los Angeles, FLA 51%, -125. Recent form (+3.2), head to head (+2.7, Panthers have won four of six) and goal margin (+2.5); streak (-3.1) against, with the injury line (-0.9) because Aleksander Barkov is on long-term injured reserve and Aaron Ekblad is day to day. Florida is 1-0-2, lost 3-2 in overtime in Anaheim on Sunday, and this is the fourth game of their road trip, the one spot our body-clock study found to be the weakest on a long trip. The reason to look at it anyway is Darcy Kuemper: 21 saves on 29 shots and eight goals against in his last start, .782 over his last five, while Jacob Markstrom opened his Florida season with a 15-save shutout. If the goalie gap is real, a 51% team at -125 is still a pass; the 10 PM puck drop is for the props below.

Buffalo over Minnesota, 51%, -112. A true pick 'em on the price (MIN -108, BUF -112) and a true coin flip on our number (BUF 51%). The streak (+3.1) carries Buffalo, with shot margin (+0.8) and the penalty kill (+0.6); goal margin (-4.2) is the case against, because Minnesota is 2-0 and +2.5 a game after beating Boston 4-1 on Saturday. Jesper Wallstedt has stopped 45 of 47 this season (.957) with Filip Gustavsson out until November; Ukko-Pekka Luukkonen is 1-1 after stopping 12 of 15 against Chicago. Buffalo is without Jason Zucker and Conor Timmins; Minnesota without Brock Faber. A total of 6.5 again. Pass on the side; this one is a shots game.

Eighteen goalies in one table

Every starter below is a DailyFaceoff projection, not a confirmation; confirmations land through the afternoon on the goalies page. Form is each goalie's last five starts, which for most of them reaches back into last April.

The headline number is built from team results and the schedule; the goalie matchup lives in the model's secondary read, the one that writes the drivers, so it is not fully priced into the percentages above. That makes an Upgrade a reason to look twice at a Pass and a Downgrade a reason to be careful with a Lean, and the three games where the gap is widest are Florida at Los Angeles, Nashville at Toronto and Minnesota at Buffalo.

The shots board

Shots on goal are the model's best prop in hockey, because volume is sticky and the model counts it from each player's own logs. The line shown is the half-point line where our distribution gives the over at least 57%; the probability is our number at the common 2.5 line. A caution in the model's own voice: these printed numbers run hot. The props on the Model Cards have hit 65% of the time against printed figures in the 70s and 80s, which is still a profitable rate at -120, but it is the realised rate to bet on, not the printed one.

Three more lines from the same projections that are not shots. Jack Eichel is projected for 1.5 assists against Seattle, which makes one assist a 78% proposition on our number; Clayton Keller for 1.3 against New Jersey (72% for one); Nick Suzuki for 1.0 against Carolina (62%). And in goal, Igor Shesterkin projects to 21 saves against the Islanders, so the over is a play at 19.5 or lower, and Jesper Wallstedt to 22 saves on 1.7 goals against, the most goalie points on the slate.

Anytime goal

Our probability is from each player's expected goals in this game; Fair is the price that probability deserves and Take at is the worst price that still leaves an edge after a book's margin. The list is the model's, unedited, which is why Cole Reinhardt is on it: 0.77 expected goals on 1.6 projected shots is a shooting rate nobody sustains, and it is the small-sample projection of a player with a handful of logs, so we would not pay -116 for it. The rest of the list is ordinary.

Three tickets

The same three the Model Cards build every night: the games as one ticket at the market's prices, the props at fair odds (props carry no market price in our data, so take the legs at -120 or better), and the best of both. Our probability is the product of the legs; the fair price is what that probability deserves.

Read the first row the way it is written. Our number says the four-leg card is worth +1038 and the book pays about +788; the difference is the price of the two legs the market likes more than we do. The card will still play it, because the card's job is to log the four most likely winners every night and grade them in public. Your job can be more selective than that. One construction rule we hold to: no leg sits on every ticket, and no game side goes on a props ticket.

Schedule spots

No team on Tuesday's slate is on the second of a back to back by the model's Monday-night read, and one actually is: Ottawa, as above. The rest of the schedule is clean. Minnesota flew 1,161 km east and crossed a time zone for Buffalo; St. Louis went 1,475 km east and a zone from Denver to Chicago; Nashville 1,044 km and a zone to Toronto. Florida is on the fourth game of its road trip (40.9% for the visitor in that spot over five seasons) and Utah the third. The base rates for reference: NHL teams on the second of a back to back win 43.7% against 51.1% otherwise, and with three days off 57.6%. Travel distance and body clock, measured properly over five seasons, added nothing to the model once team strength was controlled for, which is why they are colour here and not a reason.

Injuries that move the number

The model values every out skater by his own production and prices the loss above replacement, discounted for absences the team has already been playing through. The lines that moved a Tuesday number by more than half a point: Detroit without Larkin, Edvinsson and Kasper against Ottawa without Zub is worth +1.8 to Ottawa; Montreal without Guhle against Carolina without Jarvis and Nikishin is -0.9 against Carolina; Florida without Barkov is -0.9 against Florida; Chicago without Bedard against St. Louis without Mailloux and Tucker is +0.8 to St. Louis; the Rangers without Sykora against the Islanders without Barzal is +0.7 to the Rangers. Everything else on the report (Buffalo's Zucker, Timmins and Lyon; Minnesota's Faber and Gustavsson; Toronto's Joshua and Domi; Utah's Lamoureux; Seattle's O'Brien and Melanson) is either already priced in or too small to show.

How the last cards landed

The NHL card had its two best nights of the young season back to back. Saturday's card went 4-0 (Colorado 65%, Buffalo 64%, Edmonton 59%, Minnesota 57%) and the games parlay hit at +321. Sunday's went 4-0 again (Vegas 67% at Vancouver, won 3-2; Seattle 56% over Calgary, won 6-1; the Rangers 52% over Utah, won 4-2; Anaheim 51% over Florida, won 3-2 in overtime) and that parlay hit at +777, with the shots props splitting: Cutter Gauthier cleared 2.5 with four, Jack Eichel missed with two, three others still to grade. Monday's four legs (Tampa Bay -205 over Philadelphia, Pittsburgh -170 over Winnipeg, Dallas -175 over San Jose, Ottawa -115 at Boston) were still on the ice when this went to press; they grade themselves on the Model Cards page as the finals come in.

The record

That middle row is the whole philosophy of this piece. The model is good at the games it is sure about and has been worse than a coin flip at the ones it is not, and on Tuesday it is sure about two. The live record, the hit rate by confidence band and every voided pick with its reason are on the NHL accuracy page; the ledgers, every past card and every graded ticket are on the Model Cards page. Units are a flat one unit on every game leg at the card's frozen price, and the ledgers are never added across leagues.

Responsible play

This is model output, not betting advice. Verdexed publishes its model's probabilities and edges for informational and entertainment purposes only. Nothing here is a recommendation to place a wager, and no outcome is guaranteed, even the model's highest-confidence picks lose regularly.

You must be of legal betting age in your jurisdiction and located where sports wagering is legal to bet. Sports betting is not available everywhere; know and follow the laws where you are. Never bet more than you can afford to lose, and never chase losses.

If gambling is affecting you or someone you know, help is available. In the US, call or text 1-800-GAMBLER (1-800-426-2537), or visit ncpgambling.org. Please play responsibly.