Game Predictor: can we leave the links up for past wks (1 Viewer)

[Hipple Long Ware & Peete]

I like using the game predictor when I am torn between a few players. I'd like to see what the earlier weeks predictions said when compared to what the actual scores were. Thanks

 

[Nightly Mistake]

Why do you like accountability?

 

[Hipple Long Ware & Peete]

Why do you like accountability?

it would be nice to see where if anything it's over or underestimating production...

 

[Captain Hook]

I like using the game predictor when I am torn between a few players. I'd like to see what the earlier weeks predictions said when compared to what the actual scores were. Thanks

Not sure I understand - isn't it always there if you go back to Week 1 content, or Week 2.....any past weeks content is still posted isn't it?

 

[Hipple Long Ware & Peete]

I like using the game predictor when I am torn between a few players. I'd like to see what the earlier weeks predictions said when compared to what the actual scores were. Thanks

Not sure I understand - isn't it always there if you go back to Week 1 content, or Week 2.....any past weeks content is still posted isn't it?

I meant have wks 1 and 2 up side by side with actual totals etc, maybe even run a doddsian algorithm or two on it...

 

[Ignoratio Elenchi]

I think when I did this I meant to post it in this thread, but I put it in the other one instead:

Just pulled in all of the game predictors since 2006 (except for 2006 week 2 which isn't there for some reason), and the actual results of the games from PFR. Just picking winners and losers with no regard for the spread, the game predictor is 672/1040 in that span. That's a win rate of 64.6%.The total score spit out by the game predictor was within 5 points of the actual total score 297 times, a rate of 28.6%. It was within 10 points of the actual total 52.7% of the time. I don't have the spreads and O/Us handy, but I'll try to dig them up.

ETA: The game predictor overestimated the total score 503 times, underestimated it 531 times, and got it exactly right 6 times.

 

[Hipple Long Ware & Peete]

I think when I did this I meant to post it in this thread, but I put it in the other one instead:

Just pulled in all of the game predictors since 2006 (except for 2006 week 2 which isn't there for some reason), and the actual results of the games from PFR. Just picking winners and losers with no regard for the spread, the game predictor is 672/1040 in that span. That's a win rate of 64.6%.The total score spit out by the game predictor was within 5 points of the actual total score 297 times, a rate of 28.6%. It was within 10 points of the actual total 52.7% of the time. I don't have the spreads and O/Us handy, but I'll try to dig them up.

ETA: The game predictor overestimated the total score 503 times, underestimated it 531 times, and got it exactly right 6 times.

Nice. 64% That would make money in vegas overall. Nice job.I'll let someone smarter than me interpret the other #'s.

 

[Ignoratio Elenchi]

Nice. 64% That would make money in vegas overall. Nice job.

Would it? Remember, this is just straight up, not against the spread. I'm not a betting man, but I'm thinking 64% might not be enough to profit when you're just betting moneylines. I really have no idea, though.

 

[videoguy505]

Nice. 64% That would make money in vegas overall. Nice job.

Would it? Remember, this is just straight up, not against the spread. I'm not a betting man, but I'm thinking 64% might not be enough to profit when you're just betting moneylines. I really have no idea, though.

Right. Betting M/L 65% might not be enough.Generally you have to hit 55% ATS to break even.

 

[Hipple Long Ware & Peete]

Nice. 64% That would make money in vegas overall. Nice job.

Would it? Remember, this is just straight up, not against the spread. I'm not a betting man, but I'm thinking 64% might not be enough to profit when you're just betting moneylines. I really have no idea, though.

Right. Betting M/L 65% might not be enough.Generally you have to hit 55% ATS to break even.

I though you said 64% ATS. ML would have way to many complicating factors to know the answer.However, there are no spreads with O/U's. Wasn't that the 64%?Now u see why I asked the math s for help

 

[Ignoratio Elenchi]

I though you said 64% ATS. ML would have way to many complicating factors to know the answer.However, there are no spreads with O/U's. Wasn't that the 64%?Now u see why I asked the math s for help

64% of the time, the game predictor correctly predicted the winner of the game. Straight up, no spreads, no totals, nothing like that - just that 64% of the time, the team that Dodds predicted to win the game actually won the game.Regarding the totals, I don't have the historical over/unders yet. I was just comparing the total score from the game predictor to the actual total score of the game. So for example, in week 1 Dodds predicted the Jaguars would beat the Broncos 19.8-18.4, for a total of 38.2 points. The actual score was 24-17, for a total of 41 points. All I measured was how far off the prediction was from the actual result (in this case, the game predictor was off by 2.8 points). But that doesn't say anything about winning O/U bets - if the O/U on that game was 40 points, then the game predictor would lead you to take the under and you'd have lost. On the other hand if the O/U had been 37 the game predictor would lead you to take the over, and you'd have won. Because I don't know what the actual O/U was for all these games, I can't say how well the game predictor is doing if you were using it to make O/U bets.

 

[TheWick]

I though you said 64% ATS. ML would have way to many complicating factors to know the answer.However, there are no spreads with O/U's. Wasn't that the 64%?Now u see why I asked the math s for help

64% of the time, the game predictor correctly predicted the winner of the game. Straight up, no spreads, no totals, nothing like that - just that 64% of the time, the team that Dodds predicted to win the game actually won the game.Regarding the totals, I don't have the historical over/unders yet. I was just comparing the total score from the game predictor to the actual total score of the game. So for example, in week 1 Dodds predicted the Jaguars would beat the Broncos 19.8-18.4, for a total of 38.2 points. The actual score was 24-17, for a total of 41 points. All I measured was how far off the prediction was from the actual result (in this case, the game predictor was off by 2.8 points). But that doesn't say anything about winning O/U bets - if the O/U on that game was 40 points, then the game predictor would lead you to take the under and you'd have lost. On the other hand if the O/U had been 37 the game predictor would lead you to take the over, and you'd have won. Because I don't know what the actual O/U was for all these games, I can't say how well the game predictor is doing if you were using it to make O/U bets.

About half of the 20 man football pool I run every year of which most just take random guesses hit at about this same rate.