Game Predictions – 4/1/2006

Sorry for being a little later than usual. Here are tonights game predictions. I am calling the Florida-Tampa game the game of the night with Calgary-Edmonton a close second. The reason Florida-Tampa is the game of the night is because the outcome of this game will tell us if Florida has any chance at the playoffs. Florida currently trails Tampa (and Montreal and New Jersey) by 8 points which is quite a lot to make up in just 9 games. But tonighs game is the first of 3 that the Panthers and Lightning have against each other before the season ends. If Florida can win all 3 of those games they would only have to make up 2 points in the other 6 games which is certainly possible. A loss tonight and we can pretty much put the final nail in the coffin of what seems to be a never give up Panthers team which is 10-2 in their last 12 games. If you like cheering for the underdog, cheer for the Panthers. They are actually a pretty good team now that they are mostly healthy and Luongo is as good of a goaltender as there is.

Home Team Fair
Road Team Fair
Predicted Winner Confidence
Philadelphia -141 New Jersey 129 Philadelphia Some
Toronto -101 Buffalo 101 Buffalo Some
Montreal -138 Boston 127 Montreal Some
Ottawa -253 Washington 160 Ottawa Strong
Atlanta -138 Carolina 128 Atlanta Good
Columbus -129 Chicago 122 Columbus Some
Florida 124 Tampa Bay -131 Tampa Bay Good
Nashville -306 St. Louis 167 Nashville Strong
Edmonton -159 Calgary 137 Edmonton Good
Los Angeles -142 Dallas 129 Los Angeles Good
San Jose -135 Phoenix 126 San Jose Some

This article has 18 Comments

  1. For a while now, I’ve heard a lot of people talking about how reliable a teams is on any one player for their goal output. So, this morning, I decided to see which team goal-scoring leaders have produced most of the team’s goals.

    (In order starting with most goal production):
    1. Alexander Ovechkin, 23.1%
    2. Jaromir Jagr, 21.8%
    3. Brian Gionta, 19.9%
    4. Jonathan Cheechoo, 19.6%
    5. Ilya Kovalchuk, 18.8%
    6. Simon Gagne, 17.1%
    7. Jarome Iginla, 16.4%
    T8. Marian Gaborik, 15.8%
    T8. Eric Staal, 15.8%
    10. Sidney Crosby, 15.6%
    11. Olli Jokinen, 15.5%
    12. Teemu Selanne, 15.2%
    13. Dany Heatley, 15.2%
    14. Ryan Smyth, 14.0%
    15. Vincent Lecavalier, 14.0%
    16. Miroslav Satan, 13.9%
    17. Nicolai Zherdev, 13.8%
    18. Steve Sullivan, 13.4%
    19. Anson Carter, 13.0%
    20. Michael Ryder, 13.0%
    21. Henrik Zetterberg, 12.9%
    22. Jere Lehtinen, 12.9%
    23. Patrice Bergeron, 12.3%
    24. Marek Svatos, 12.3%
    25. Shane Doan, 12.1%
    26. Mark Bell, 11.8%
    27. Darcy Tucker, 11.0%
    28. Michael Camalleri, 10.5%
    29. Chris Drury, 10.1%
    30. Scott Young, 9.3%

    Scott Young (of St. Louis) has only produced 17 goals, 9.3% of the teams 183 goals. The team leader officially is Mike Sillinger, but he isn’t on the team anymore, which is why I listed Scott Young. The three offensive Rookie of the Year candidates (Ovechkin, Svatos, and Crosby) are all team goal leaders. Dion Phaneuf (Calgary) and Henrik Lundqvist (NY Rangers) are a defenceman and goaltender respectively.

  2. Messy, messy yesterday 😯

    Strong: MTL, CGY, DAL
    Some: PHI, TOR, FLO, NSH, PHX
    Upset: NONE
    Home: OTT, ATL, CBJ


  3. David. I have asked before but you never got back to me..i was curious on why your algo system never predicts a team who played the night before to win the following night?? Is this a glich in your system? As we all know…not all teams who play back to back games lose the 2nd night. Its close to a 50/50 split. Your system did pick Columbus to win tonight but Chicago did also play last night but your algo had to predict someone. A lot of teams have a really good record on the 2nd night of a back to back game…Calgary and Dallas being two of them tonight. I was just curious on this question, get back to me if you can.

    Today’s Predictions and Proline Pools:

    -Buffalo (BOXED)
    -Ottawa (BOXED)
    -St. Louis (BOXED)
    -Phoenix (BOXED)

  4. I Know Mike Camillari personally, childhood friends with Stajan, Eminger, and Coliaicolvo. Nice to see him on the list above, Kevin.

  5. Mr. Hockey: It does. See last nights prediction of Detroit over Chicago after Detroit played the previous night. It’s just rare because teams who are playing on back to back nights when their opponent had the previous night off win only about 40% of the time. That is a significant hurdle to overcome even when the opposing team is decimated by injuries (see Rangers-Ottawa game Thursday).

  6. According to my algorithm it never lost money on Saturday or Sunday, and never won on Monday during the three weeks 😕

  7. David. Yes your system does pick winners on the 2nd night of a back to back game, but very very seldom. Just noticed the trend from a few weeks back now. I wanted you to know. If you go back thru these pages and count all the predicted winners of all the games involved, the ratio is very staggering. You said only about 40% of teams who played the night before win the next night. I just did the stats and calculations and I got 48%, which is closer to my 50% as orginally said above. All im saying is a team who played the night before in a nut shell has aprox. 50/50 shot according to my stats and records to win the next night. From what Ive seen and noticed for about week and half now, your system piredicts aprox. 10% of teams who played last night to win the following night. Hense, this is why i asked if this is a glitch in your system.

  8. FYI.. St. Louis is a NHL worst 1-12-3 on the 2nd night of a back to back game, which drastically brings the percentage to 48%. If it was for St. Louis, teams playing back to back games would have an above .500 percentage to win the next night.

  9. And for all those who say I sound rude…thats not at all what this is meant to be. Just noticed a suprising stat and wanted to bring it to Davids attention in case he wasn not aware. This is Hockey , and thats what Im doing…Analyzing Hockey.

  10. Over the course of the whole season teams who are playing on back to back nights when their opponent had the previous night off are 119-175 for a .405 win %.

    In 2003-04 teams were 119-166- 56, 0.431.
    In 2002-03 teams were 106-177- 47 0.392.

    I don’t mind people questioning my algorithm. It only forces me to think about it more and look at different ways of improving it. I have some ideas and hope to have time to play around with them over the summer and have a new and improved algorithm for next season.

  11. It’s not a surprise that Ovechkin and Jagr are the top 2 on that list, as they have a huge lead on 2nd place on the team. I realized that I didn’t list the team that each player plays on, so I’ll list those right now:

    (In order of 1-30) Washington, NY Rangers, New Jersey, San Jose, Atlanta, Philadelphia, Calgary, Minnesota, Carolina, Pittsburgh, Florida, Anaheim, Ottawa, Edmonton, Tampa Bay, NY Islanders, Columbus, Nashville, Vancouver, Montreal, Detroit, Dallas, Boston, Colorado, Phoenix, Chicago, Toronto, Los Angeles, Buffalo, St. Louis

  12. David. I see where we diff in the stats we calculated. I did not include overtime or shootout losses into the equation because its not a complete loss …walking away with a point is still a point. But it does count for a loss in streaks and such. So if you did add overtime and shootout losses than yes…the stats are aprox 40%. Its still a lot higher than your system predicts for back to back games. Thats all i wanted to bring to your attention. I think your system would be alot more efficent once you improve it over the summer as you said. Keep it up.

  13. My algorithm, and any successful algorithm, will rarely predict teams playing back to back games to win. Why? Because doing so will be predicting upsets and predicting upsets is next to impossible to do. We all know upsets happen. They happen a lot. But trying to figure out when that upset is going to happen is next to impossible. I mean, if it were reasonably easy to predict an under dog to win, it probably means that the under dog really shouldn’t be classified as an underdog.

    Let’s assume that the team who is playing on back to back nights wins the second game 40% of the time. If I always predict that team to lose, I’ll be correct 60% of the time. If I always predict that team to win, I’ll be correct 40% of the time. I’d rather be 60% right than 40% right. The challenge is trying to improve of the 60%. If upsets were completely random by switching some of your picks of the back to back team to lose to predictions of a win, you will on average only lower your 60% success rate.

    So the question is, is there something we can grasp on to which makes upsets not random. Is there a reason why upsets happen? Is there some other piece of information that I can use that will help identify when a back to back team will win aside from the fact that random upsets will happen? That is the question that needs to be answered but you will never find me predicting anywhere close to 40% of teams playing back to back games to win.

  14. David. that totally makes scense. I totally understand what you are saying. I agree in that upsets ae hard to predict. Taking all favs night in and night out, people will be around 60% correct. Hard part is determining where upsets will occur. In reality there is no system or numbers or stats that really matter in picking games. It really is all random. Thats the sad part about it..its sports anything can happen.
    -An idea I had was this….
    We all know home teams tend to pull out wins more statistaclly than away at about a 2/3 ratio. Taking that into consideration, maybe you can input into your system when a team is playing back to back games and the 2nd game is at home maybe theres a way to NOT let your system know it is a back to back gamer. A 2nd game for a team has more likelyhood to go in favour of the home team. An away team on a 2nd night of a back to back game has more chances to lose. See what Im saying?
    In eliminating the awareness of your system to know a home team is playing back to back might make it a more even decision for your algo and maybe more fair????? Just a thought. Not the best idea but its the first that popped into my head.

  15. i think maybe thats just one small way to increase your percentages …and even a little improvment makes a world of a difference.

  16. Changes to my predictions for tonight:

    Proline Pools Ticket:

    -Montreal over Boston (Back2back SOs for Huet vs Boston)
    -Atlanta/Carolina *BOXED* (leaning towards Atlanta)
    -Washington/Ottawa *BOXED* (maybe Morrison will start??)
    -Buffalo vs Toronto (Leafs are 1-6-0 on 3 days rest or more)
    -Columbus over Chicago
    -Florida over Tampa Bay (Florida 3 striaght wins vs TB)
    -St. Louis/Nashville *BOXED* (Blues might snap losing skid)
    -Calgary over Edmonton (Flames won 6 of last 9 in EDM)
    -Dallas over L.A. (Stars 12-1-2 on 1st gm back on road)
    -San Jose over Phoenix

  17. I wanted to comment on your algorithm predicting Los Angeles to beat Dallas in the first place, and secondly, do they deserve to be considered a Good favorite? Since firing Andy Murray, they have gone 1-3, with the 3 losses against teams looking for playoff seeds. Their win came against Nashville, who has practically locked up the 4th seed. Losses coming against Calgary, who is competing with Vancouver, Edmonton, and Colorado for the 3rd seed, while the others will likely get one of the others. I see all four of those Northwest division teams to make the playoffs, with two of the three Pacific division teams to miss out. I’ll make a prediction now and say Los Angeles and San Jose will be spectators. But let’s take a look at future schedules, shall we?

    (These schedules only list the games each team plays against the other teams in the list. Numbers in parenthesis represent multiple games against a team)

    Anaheim: Calgary (2), Edmonton, Los Angeles (2), San Jose, Vancouver
    Calgary: Anaheim (2), Colorado, Edmonton, Los Angeles, Vancouver
    Colorado: Calgary, Edmonton, San Jose, Vancouver
    Edmonton: Anaheim, Calgary, Colorado
    Los Angeles: Anaheim (2), Calgary, San Jose (2), Vancouver
    San Jose: Anaheim, Colorado, Los Angeles (2), Vancouver (2)
    Vancouver: Anaheim (2), Calgary, Colorado, Los Angeles, San Jose (2)

  18. Mr. Hockey: My algorithm ‘penalizes’ a team for playing back to back games and then penalizes them even more if they are playing on the road. If they are playing at home they take the back to back penalty but get a home team bonus. But home teams are only playing .426 hockey at home this season.

    Overall home teams win about 55% of the time and road teams win about 45%. This number has been remarkably consistant over the past 3 seasons.

    Kevin: Los Angeles’s playoff hope are all but done. They would need to get on a lengthy winning streak to get back in the race and I don’t see that in them. Too many question marks for that. San Jose is still in the thick of the race though but have to get going again after after poor outings.

    As for tonights predictions, the one short fall with my algorithm is that it doesn’t take into account things like injuries and Los Angeles playing without Demitra just isn’t the same team. But Washington beat Ottawa tonight so anything is possible.

Comments are closed.