Game Predictions – 3/13/2006

We have 5 games on the schedule today, a couple of them with huge playoff significance. In the east we have the free falling Tampa visiting the hot Montreal Canadiens. Montreal has moved 1 point ahead of the Lightning into 7th spot int he east and would desperately like to put some separation between the two teams forcing Tampa to deal with the Thrashers, Islanders and Leafs who are looking to grab a playoff spot. The number 1 thing the Thrashers, Islanders and Leafs is for the game to no go into overtime and become a 3 poitn game.

In the west we have a game of similar importance. Tenth place San Jose plays 6th place Los Angeles who they trail by 8 points but San Jose has 3 games in hand. A San Jose win tonight will put the Kings within reach just by winning those 3 extra games. This is a big game for both the Skarks and Kings.

In other games Vancouver who isjust 2 points ahead of 9th place Edmonton looks to improve its playoff chances against the Dallas Stars. Calgary and Colorado are playing for playoff position as well. Calgary trails Colorado by a single point for top spot in the northwest division and the 3rd seed in the western conference playoffs. The final game, S. Louis at Columbus, no one really cares about.

Here is my predictions success rate since I began using the new prediction algorthm.

Strong: 22 of 35 – 62.9%
Good: 25 of 41 – 61.0%
Some: 78 of 135 – 57.8%
Overall: 125 of 211 – 59.2%

Since last week Strong and Some success rates rose marginally while the Good sucess rate fell from 69%. The overall success rate fell from 59.5%. It was a tough week to make predictions being the trade deadline and all but also because there were so many close games. Over the last week there were 11 games to go to a shootout and another 8 games that ended in overtime.

Home Team Road Team Predicted Winner Confidence
Montreal Tampa Bay Montreal Some
St. Louis Columbus St. Louis Some
Dallas Vancouver Dallas Some
Calgary Colorado Calgary Some
San Jose Los Angeles San Jose Some

This article has 13 Comments

  1. You have the same fraction (78 of 135) for both some and overall. Overall should be 125 of 211 with the same win%.

    All of the games on the schedule today could go either way, which means I had a hard time with my predictions this morning.

    Success Rate (Since David’s new formula on Jan. 27)
    Strong: 13 of 17, 76.5%
    Good: 51 of 92, 55.4%
    Some: 55 of 102, 53.9%
    Overall: 111 of 211, 56.4%

    Strong: None
    Good: Montreal, St. Louis, Dallas
    Some: Colorado, Los Angeles

  2. Ughhh. I fixed that in my program but apparently I fixed it on my notebook and not my main computer. It’s fixed now.

  3. I think I should be your peer-editor. The last 3 errors I have found them all and posted it 1st. It also helps that I wake up really early and check this site first thing in the morning after I wake up.

  4. Good job Kevin…lol Since Dave is doing this all for free and for our enjoyment It’s nice to have somone help him…keep up the good work Dave and dave’s helper Kevin….(good Eye)


  5. I like all your picks today Dave…with the dall/van game going either way…I like Montreal today but I think tampa will come out really hard toady….but you have to go with the hot goalie huet on this just because of that…

  6. Dave I have an Idea I thought I’d throw at you…

    I was wondering how well you would do without your charts or algorythm? I was wondering If you could make your picks on your own first before refering to your charts and then make another list beside yours with your chart picks and that way we will see it you can pick higher percentages without the charts or if the charts make much of a difference…? If you do it, you can still use the internet stats like everyone else to make your decisions… I think it would interesting to see the outcome…I would like to see you try it for the remaining games…

    let me know what you think about this?


  7. Yeah, I was wondering about the same thing. This way, we can compare whether human or computer is better at these things. Of course, his computer is doing better than me, but that doesn’t really help you. Also, David, I was wondering if and when you are going to continue betting on your predictions? I have been doing this since you started and am hovering at the $116 mark.

  8. For me it is all about how much free time I can devote to this. There are a ton of things I would like to do but right now my bill paying job is taking up more of my time. I expect this will continue until the end of March. In April I might be able to get back to keeping track of bets or making human predictions.

    But, with that said, my gut tells me that algorithm/computer generated predictions should be better, especially now when the NHL schedule is so unbalanced. Some teams records are significantly skewed because they have very easy or very difficult schedules. It is diffucult for a human to correctly factor that in. Add in points for overtime losses and shootout losses and it is all too much to process. Philadelphia has 80 points, but are they really better than Montreal with 71? It is also difficult for a human to accurately value things like back to back games, 3 games in 4 nights, home vs road, etc. Pittsburgh vs New Jersey on Saturday is a perfect example.

    That said, there are also factors that an algorithm can’t accurately account for too. Trades, injuries, short term streaks (although I am not sure humans can correctly account for streaks either). The best method would be probably be some combination of the two where the human takes the computer predictions and makes modifications to them. But some clear rules would have to be made as to when it would be appropriate to make such modifications. You might think that you should change the computer prediction from Pittsburgh to New Jersey but there would have been no justifiable reason for that. You could argue that trading Recchi is a justifiable reason but like I said on Saturday, sometimes when a team loses a big name player to trade or injury they play better, at least in the short term, as everyone else tries to step up their game to fill the hole or maybe they see it as an opportunity to pick up some of that players ice time. Pittsburgh beat New Jersey and Philly on the weekend. When Boston traded Thornton they played better for a few games including a big win against Ottawa.

  9. You say that you might want to change the prediction from NJ to Pittsburgh. You didn’t and it worked, as Pittsburgh won 6-3 and 2-0 on back-to-back nights. Pittsburgh & St. Louis are among the teams that have turned it around recently. But the first half, they did so poorly that they are good spoilers now. St. Louis has gone 7-1-3, only trailing Buffalo for the best record since Feb. 7.

  10. Ok Dave I understand…No prob…actuall what you said about combining is good because that’s what I try to do…I look at your chart picks and I usally change 1 to a few picks which I have a gut feeling or there are some issues about the game etc..


  11. I like this site alot especially due to the picks…i always make my own picks no matter what and a lot of times my thoughts are totally different. I have my good days and my bad days like anyone else. I track my picks all season front to back and this season im hovering around the 61% mark which isnt bad. I play Proline and strictly proline. I dont bet small amounts, I go for the jackpots. I tend to lose a game here and there on average nights and thus ruining a good ticket. Thats gambling i guess. Does anyone else here play Proline Pools on average? Thats where the moneys at. David, how many times have i seen you hit all 10 games and more or whatnot on a good night? Ive seen it like twice this season. If you bet that exact outcome scenerio on Proline Pools you would have walked away with a cool $10,000 at minimum two times arleady. I highly recommend playing Pools if people live in Ontario. Great cash…no other game or site or casino can match that payout. Out of aprox. 25 hockey pools Ive seen minimum pay out $350 for all correct and Ive seen max payout $104,000. Average payout on a perfect ticket is about $5000-$10,000. Not bad for guessing all the games in one night all correctly.

  12. I’ve had a perfect night a few times when there has been more than 5 games. I think my best night was a 5 for 5 night. But, I highly doubt I would make money playing proline using your method. I don’t believe that playing parlays is the best way to make money gambling and certainly not with proline since the odds are so poor. I did a check a couple weeks ago and the same 3-team parlay bet at Bowmans had a 50% larger payout.

Comments are closed.