Trying a new way to project fantasy points (1 Viewer)

[joetrow]

First off let me say I work overnights, so today being a day off I decided I would try somthing new, everyone at home is asleep, so why not burn some time. It took me roughly 2 hours to do 6 running backs, I'm sure it would go faster overtime but I can't say this is a time efficent way to do lineups that I will do often.

I went with my roster of RB which is Ronnie Brown, Darren McFadden, Tashard Choice, Mike Bush, Donald Brown, and Leon Washington. I figured out thier attempts per game, yards per attempt, yards per game, rushing TD per game, Rec per game, yards per reception, receiving yards per game, and receiving touchdowns per game. I'll show you Ronnie Browns per game average:

Ronnie Brown: 17 car/game, 5.3 yrds/avg, 89.5 yrds/game, 1 rushing/td/game, 1.5 rec/game, 3.3 yrds/rec, 5 yrds/game, 0 rec/td/game

I then looked at their oppenent this weekends defense stats, San Deigo, and found out how opposing teams normally play them and how they fare. So when I say attempts its the average attempts other teams have ran agianst them this year. Also I figured out how many completed passes went to running backs agianst their D:

SD Defense gives up: 32 attempts/game, 4.3 yrds/avg, 139 yards/game, 1.5 rushing/td/game, 6rec/RB/game, 8.4 yards/rec, 50.4 rec yards/game, 0 rec/td/game

Then I looked up Miami's rushes per game (35). I looked at Miami's rushes per game VS rushing attempts agianst SD D(32). I stuck with 35 here because I think that Miami is a run first team and their wasnt much differnce between the two numbers. Other games though I moved amount of carries around by up to 5. After that I figured out Ronnie Brown's percent of total rushes (49%) and multiplied it by amount of rushes I expect in the game (35) and I came up with 17. So going by this projection Ronnie Brown will have 17 rushes.

I then averaged out Browns per game stats VS SD Defense per game stats:

I got 17 rushes, 4.5 yrds/avg for 76.5 rushing yards and 1.25 touchdowns.

I then calculated receptions the same way as rushing attempts.

I got 2.25 receptions @ 5.9 yrds/reception for 13.3 recieving yards and 0 touchdowns.

I took all the calculations and came up with: 76.5 rushing yards 13.3 recieving yards and 1.25 total touchdowns. In a non-ppr standard Yahoo league that would be 16.48 points this week. Yahoo projects him at 57.4 rushing yards, 19.5 recieving yards, .5td and .2 fumble for 10.73 and I'd almost be willing to bet my projection is closer.

So anyone who made it do the end of my rant what do you think?? Do you think its a fairly accurate way to project points?? Why or why not? There is alot of flaws but I think touchdowns is the thing that will be off by the farthest consistantly. Maybe going touchdowns per carry for Brown VS touchdowns given up per carry by SD D would be a better way to figure out projected TDs? Thanks for any input.

 

[NYCGangGreen]

Wow, my head hurts now. But that is some serious calculations. I don't think you can really estimate the reliability of it until after a small sample size (say 2 weeks).

But I like what you've done. I tend to do a mental version where I look at my guys production....then the team he plays that week and their average #'s allowed and go from there.

I'm curious to see just how your numbers work out. There is always the variable of weather and playcalling depending on the score, but it seems you have a great start to the equation that at least can give a more precise prediction.

 

[Truman]

Don't like it at this stage of the season. 2 games isn't gonna give much of an indication and playing Indy's run D inflates the number of carries while deflating reception totals for RBs.

 

[zoobird]

You really need to learn to use Microsoft Excel if you want to do things this way without going insane and giving up after a few weeks.

I agree with the guy who said two games isn't enough to based your calculations on. In fact, you'd probably do much better using full season projections (or better yet, 'rest of season' projections) from one of the sites that provide them, and then calculating 'per carry' averages from that. If you do choose to use your own projections, at least make sure to "regress them to the mean". For example, if Chris Johnson has a per carry average of 9.5 yards after two games, but a career average of 4.8 YPC, you'd be crazy to use 9.5 in your calculations...you need do some sort of weighted average, where the impact of this season vs. previous seasons gradually decreases as the season progresses.

 

[Da Guru]

I have been trying to project pork bellies for 10 years, still have not got that right. I won`t even attempt RBs.

 

[two_dollars]

I love the effort. Interested to see where this goes. Obviously there will be exceptions. See Frank Gore vs Seattle or Chris Johnson. Maybe weigh studs with a trend towards their annual projections? No formula will ever get it 100%, but anything that gets people closer will be greatly appreciated.

 

[gianmarco]

Question is: So what now? What will you do with that info?

If you're using it to see who starts, then you need to follow the projections exactly and see how it works out. And you're essentially going to be starting your top 2 scorers using this method. Otherwise, what's the point?

It's a nice premise, but this is one of those where it's too much from the math standpoint and we all know it never works out exactly like that. It would be more helpful just trying to determine carries that way and using that to base your WDIS decisions.

 

[puckalicious]

This sounds like what FBGs already does every week with Clayton Gray's Fantasy Points Matchups article.

 

[puckalicious]

The best way to project stats is have the computer (console) play itself in Madden.