What's new
Fantasy Football - Footballguys Forums

Welcome to Our Forums. Once you've registered and logged in, you're primed to talk football, among other topics, with the sharpest and most experienced fantasy players on the internet.

Value-Based Drafting and the Value of D/ST (1 Viewer)

So how do we get better at predicting top D/STs to be rewarded enough to make up for the risk in the long run?
If someone can find a good-looking data set for me, I can run some good numerical analysis and get back to you guys pretty quickly. What I'd be interested in seeing in the data set is: Yards allowed, points allowed, turnovers forced, TD's, Dodds' preseason estimate (or some other top picker), ADP, final fantasy points for the year. What I can then do is create a formula based on Multiple Linear Regression that best estimates a team's fantasy value for the year as a function of last year's points, last year's TD's, last years yards and points allowed, last year's TO's, and preseason expectation (and perhaps ADP). The formula will uncover which metrics are actually useful and which are basically totally random; and I can also tell you the strength of its predictive power. Analysis is easy for me; finding data is not.
Yeah, you aren't exactly special in that regard. The main challenge in doing statistical analysis on FF is getting the data set into a workable format. Given all the tools that are available (even on this site!) you should maybe do your own work on that front rather than expect others to do your data wrangling.
What are those tools? I haven't found a site that easily takes me back through many years of fantasy results; but I'd love if to hear if you know of one.

 
So how do we get better at predicting top D/STs to be rewarded enough to make up for the risk in the long run?
If someone can find a good-looking data set for me, I can run some good numerical analysis and get back to you guys pretty quickly. What I'd be interested in seeing in the data set is: Yards allowed, points allowed, turnovers forced, TD's, Dodds' preseason estimate (or some other top picker), ADP, final fantasy points for the year. What I can then do is create a formula based on Multiple Linear Regression that best estimates a team's fantasy value for the year as a function of last year's points, last year's TD's, last years yards and points allowed, last year's TO's, and preseason expectation (and perhaps ADP). The formula will uncover which metrics are actually useful and which are basically totally random; and I can also tell you the strength of its predictive power. Analysis is easy for me; finding data is not.
Most of these stats are readily available on NFL.com. I don't know your methods, but it seems that it's been done and isn't particularly groundbreaking according to some of the responses in this thread. There may be other stats that are not being used that can be useful as well as things that aren't so easily quantified.
I've only seen guys do single-variable analysis in any fantasy article I've read (i.e. regressing fantasy points solely on last year's points, ADP, or turnovers alone). This thread included. I just thought multiple variables simultaneously could open up the discussion a bit.

 
So how do we get better at predicting top D/STs to be rewarded enough to make up for the risk in the long run?
If someone can find a good-looking data set for me, I can run some good numerical analysis and get back to you guys pretty quickly. What I'd be interested in seeing in the data set is: Yards allowed, points allowed, turnovers forced, TD's, Dodds' preseason estimate (or some other top picker), ADP, final fantasy points for the year. What I can then do is create a formula based on Multiple Linear Regression that best estimates a team's fantasy value for the year as a function of last year's points, last year's TD's, last years yards and points allowed, last year's TO's, and preseason expectation (and perhaps ADP). The formula will uncover which metrics are actually useful and which are basically totally random; and I can also tell you the strength of its predictive power. Analysis is easy for me; finding data is not.
Most of these stats are readily available on NFL.com. I don't know your methods, but it seems that it's been done and isn't particularly groundbreaking according to some of the responses in this thread. There may be other stats that are not being used that can be useful as well as things that aren't so easily quantified.
I've only seen guys do single-variable analysis in any fantasy article I've read (i.e. regressing fantasy points solely on last year's points, ADP, or turnovers alone). This thread included. I just thought multiple variables simultaneously could open up the discussion a bit.
MathNinja please, If you want to crunch the numbers I will help you gather the data. Tell me exactly what you need and I'll do what I can.

 
So how do we get better at predicting top D/STs to be rewarded enough to make up for the risk in the long run?
If someone can find a good-looking data set for me, I can run some good numerical analysis and get back to you guys pretty quickly. What I'd be interested in seeing in the data set is: Yards allowed, points allowed, turnovers forced, TD's, Dodds' preseason estimate (or some other top picker), ADP, final fantasy points for the year. What I can then do is create a formula based on Multiple Linear Regression that best estimates a team's fantasy value for the year as a function of last year's points, last year's TD's, last years yards and points allowed, last year's TO's, and preseason expectation (and perhaps ADP). The formula will uncover which metrics are actually useful and which are basically totally random; and I can also tell you the strength of its predictive power. Analysis is easy for me; finding data is not.
Yeah, you aren't exactly special in that regard. The main challenge in doing statistical analysis on FF is getting the data set into a workable format. Given all the tools that are available (even on this site!) you should maybe do your own work on that front rather than expect others to do your data wrangling.
What are those tools? I haven't found a site that easily takes me back through many years of fantasy results; but I'd love if to hear if you know of one.
http://subscribers.footballguys.com/apps/histdd.php This covers everything except for DEF and K

http://subscribers.footballguys.com/apps/ddform.php This could also be helpful

http://espn.go.com/nfl/statistics/team/_/stat/total/position/defense Yards Allowed and points allowed and can go back to 2002, turnover data is in a link in the header as is TD data for punts, kickoff, interceptions and fumbles;

http://football.myfantasyleague.com/2013/adp ADP data,change the URL for prior years

As for prior year FF points for defenses, I know various leagues such as FFPC have prior year data available but that scoring system does not count yardage and points allowed. Maybe someone else knows where you can find one that suits your needs.

 
So how do we get better at predicting top D/STs to be rewarded enough to make up for the risk in the long run?
If someone can find a good-looking data set for me, I can run some good numerical analysis and get back to you guys pretty quickly. What I'd be interested in seeing in the data set is: Yards allowed, points allowed, turnovers forced, TD's, Dodds' preseason estimate (or some other top picker), ADP, final fantasy points for the year. What I can then do is create a formula based on Multiple Linear Regression that best estimates a team's fantasy value for the year as a function of last year's points, last year's TD's, last years yards and points allowed, last year's TO's, and preseason expectation (and perhaps ADP). The formula will uncover which metrics are actually useful and which are basically totally random; and I can also tell you the strength of its predictive power. Analysis is easy for me; finding data is not.
Most of these stats are readily available on NFL.com. I don't know your methods, but it seems that it's been done and isn't particularly groundbreaking according to some of the responses in this thread. There may be other stats that are not being used that can be useful as well as things that aren't so easily quantified.
I've only seen guys do single-variable analysis in any fantasy article I've read (i.e. regressing fantasy points solely on last year's points, ADP, or turnovers alone). This thread included. I just thought multiple variables simultaneously could open up the discussion a bit.
MathNinja please, If you want to crunch the numbers I will help you gather the data. Tell me exactly what you need and I'll do what I can.
The ones I listed would all be great. If I had 10 years of that data, I could conjure up a model that analyzes each factor's effect simultaneously.

 
So how do we get better at predicting top D/STs to be rewarded enough to make up for the risk in the long run?
If someone can find a good-looking data set for me, I can run some good numerical analysis and get back to you guys pretty quickly. What I'd be interested in seeing in the data set is: Yards allowed, points allowed, turnovers forced, TD's, Dodds' preseason estimate (or some other top picker), ADP, final fantasy points for the year. What I can then do is create a formula based on Multiple Linear Regression that best estimates a team's fantasy value for the year as a function of last year's points, last year's TD's, last years yards and points allowed, last year's TO's, and preseason expectation (and perhaps ADP). The formula will uncover which metrics are actually useful and which are basically totally random; and I can also tell you the strength of its predictive power. Analysis is easy for me; finding data is not.
Most of these stats are readily available on NFL.com. I don't know your methods, but it seems that it's been done and isn't particularly groundbreaking according to some of the responses in this thread. There may be other stats that are not being used that can be useful as well as things that aren't so easily quantified.
I've only seen guys do single-variable analysis in any fantasy article I've read (i.e. regressing fantasy points solely on last year's points, ADP, or turnovers alone). This thread included. I just thought multiple variables simultaneously could open up the discussion a bit.
MathNinja please, If you want to crunch the numbers I will help you gather the data. Tell me exactly what you need and I'll do what I can.
The ones I listed would all be great. If I had 10 years of that data, I could conjure up a model that analyzes each factor's effect simultaneously.
I will put this into a spreadsheet, but I'm not sure how to format it for you. Do you want 10 separate sheets by year? This might take me a week or two depending on how much time I have, but I'll get it done. We could also setup a google doc or something similar so others can work on the gathering, that may speed things up especially if you write the column and row headers how you want so the data can just be entered.

I'm wondering if it's best to include only the data that resulted from the current pass interference rule changes or not.

I'm also wondering what kind of math background you have, any degrees of note? Or perhaps we're looking at a Good Will Hunting situation? Were you trained by Shaolin Math Ninjas? Just curious.

 
Code:
Team	Year	Points	Sacks	Safety	Int	INT TD	Fum Rec	Fum TD	ST TD	YPG	ADPARI	2003	225	21	0	13	1	10	0	1	344	220ARI	2004	284	38	1	15	1	15	1	0	321.3	229ARI	2005	311	37	0	15	3	11	0	0	295.6	177ARI	2006	314	38	0	16	1	17	2	1	349.4	186ARI	2007	404	36	0	18	6	11	0	1	330.2	183ARI	2008	427	31	0	13	2	17	1	1	331.5	179ARI	2009	375	43	0	21	1	8	0	1	346.4	164ARI	2010	289	33	0	17	3	13	4	2	373.6	182ARI	2011	312	42	0	10	0	9	0	4	355.1	179ARI	2012	250	38	0	22	2	11	1	0	337.8	184ATL	2003	299	36	0	15	3	16	0	1	381.8	121ATL	2004	340	48	0	19	4	13	1	1	325.4	157ATL	2005	351	37	1	16	0	13	2	0	325	116ATL	2006	292	37	0	12	1	14	1	0	332.8	150ATL	2007	259	25	0	16	0	11	1	0	355.5	198ATL	2008	391	34	1	10	2	8	0	1	348.2	238ATL	2009	363	28	0	15	1	13	2	0	348.9	195ATL	2010	414	31	0	22	2	9	1	2	332.4	145ATL	2011	402	33	0	19	2	10	1	0	333.6	150ATL	2012	419	29	0	20	1	11	1	0	365.6	166BAL	2003	391	47	0	24	4	17	0	0	271.3	98BAL	2004	317	39	0	21	5	13	1	2	300.2	70BAL	2005	265	42	0	11	1	15	2	0	284.3	67BAL	2006	353	60	2	28	5	12	1	0	264.1	106BAL	2007	275	32	0	17	1	6	0	3	301.6	63BAL	2008	385	34	3	26	5	8	1	0	261.1	146BAL	2009	391	32	0	22	2	10	1	1	300.5	84BAL	2010	357	27	0	19	3	8	0	1	318.9	104BAL	2011	378	48	0	15	1	11	3	1	288.9	107BAL	2012	398	37	0	13	2	12	0	3	350.9	106BUF	2003	243	38	3	10	2	8	1	0	269.6	164BUF	2004	395	45	1	24	4	15	0	5	264.2	155BUF	2005	271	38	1	17	1	13	0	1	343.5	87BUF	2006	300	40	0	13	2	11	2	1	329.6	175BUF	2007	252	26	2	18	2	12	1	2	362.9	175BUF	2008	336	24	0	10	3	12	0	2	326.1	168BUF	2009	258	32	0	28	2	5	0	0	340.6	191BUF	2010	283	27	0	11	2	11	1	1	361.6	206BUF	2011	372	29	1	20	4	11	2	1	371.1	238BUF	2012	344	36	1	12	1	9	0	3	362.9	160CAR	2003	325	40	1	16	3	10	0	2	295.3	90CAR	2004	355	34	0	26	2	12	1	0	336.4	93CAR	2005	391	45	0	23	2	19	1	0	282.6	105CAR	2006	270	41	1	14	2	8	0	0	296.1	84CAR	2007	267	23	0	14	1	16	1	0	324.8	139CAR	2008	414	37	0	12	0	13	2	0	331.2	180CAR	2009	315	31	2	22	1	15	0	0	315.8	168CAR	2010	196	31	1	17	1	12	0	0	335.9	191CAR	2011	406	31	2	14	0	10	0	1	377.6	238CAR	2012	357	39	1	11	3	12	0	0	333.1	248CHI	2003	283	18	0	15	1	5	0	3	309.2	166CHI	2004	231	35	3	17	5	12	1	1	336.9	175CHI	2005	260	41	0	24	4	10	0	1	281.8	135CHI	2006	427	40	1	24	1	20	2	5	294.1	68CHI	2007	334	41	1	16	1	17	0	6	354.7	58CHI	2008	375	28	1	22	1	10	2	1	334.7	96CHI	2009	327	35	1	13	1	15	0	2	337.8	116CHI	2010	334	34	0	21	1	14	0	3	314.3	141CHI	2011	353	33	1	20	4	11	2	3	350.4	123CHI	2012	375	41	0	24	8	20	1	0	315.6	124CIN	2003	346	30	0	14	0	10	1	1	351.2	220CIN	2004	374	37	0	20	4	16	1	0	335.3	186CIN	2005	421	28	0	31	1	13	0	0	338.7	155CIN	2006	373	35	0	19	1	12	0	0	355.1	149CIN	2007	380	22	1	19	1	15	2	1	348.8	173CIN	2008	204	17	0	12	1	12	2	0	325.5	238CIN	2009	305	34	0	19	2	6	1	1	301.4	154CIN	2010	322	27	0	16	1	10	1	0	332	153CIN	2011	344	45	1	10	1	12	2	1	316.2	123CIN	2012	391	51	0	14	2	16	1	1	319.7	172CLE	2003	254	35	0	15	1	7	0	0	309.9	178CLE	2004	276	32	0	15	1	13	0	1	325.9	196CLE	2005	232	23	0	15	0	8	0	2	316.8	235CLE	2006	238	28	0	18	1	9	1	1	344.8	190CLE	2007	402	28	1	17	1	10	0	3	359.6	234CLE	2008	232	17	0	23	2	8	0	1	356.5	183CLE	2009	245	40	1	10	0	9	0	4	389.3	241CLE	2010	271	29	0	19	2	9	1	0	350.1	194CLE	2011	218	32	0	9	0	11	0	1	332.4	196CLE	2012	302	38	0	17	2	12	0	1	363.8	248DAL	2003	289	32	2	13	2	12	0	3	253.5	149DAL	2004	293	33	1	13	0	9	0	0	330.3	113DAL	2005	325	37	0	15	2	11	0	0	300.9	149DAL	2006	425	34	0	18	3	13	1	1	322.8	137DAL	2007	455	46	0	19	2	10	1	0	307.6	111DAL	2008	362	59	1	8	1	14	0	1	294.3	96DAL	2009	361	42	0	11	1	10	0	2	315.9	133DAL	2010	394	35	0	20	3	10	1	3	351.8	125DAL	2011	369	42	0	15	1	9	0	0	343.2	152DAL	2012	376	34	0	7	1	9	2	1	355.4	167DEN	2003	381	36	1	9	1	11	0	2	277.1	161DEN	2004	381	38	0	12	2	8	0	0	278.7	139DEN	2005	395	28	1	20	3	16	0	0	312.9	149DEN	2006	319	35	0	17	2	13	0	0	326.4	153DEN	2007	320	33	1	14	0	15	2	2	336	113DEN	2008	370	26	0	6	0	7	2	0	374.6	184DEN	2009	326	39	0	17	0	13	2	2	315	241DEN	2010	344	23	0	10	0	8	1	1	390.8	178DEN	2011	309	41	0	9	3	9	0	2	357.8	150DEN	2012	481	52	2	16	5	8	1	2	290.8	160DET	2003	270	28	0	15	1	13	3	1	335	220DET	2004	296	38	1	14	1	10	0	4	337.6	184DET	2005	254	31	1	19	2	12	1	0	322.4	181DET	2006	305	30	1	12	1	18	0	0	345.6	198DET	2007	346	37	1	17	3	18	1	1	377.6	234DET	2008	268	30	1	4	0	16	1	0	404	189DET	2009	262	26	2	9	2	13	1	0	392.1	241DET	2010	362	44	1	14	2	15	1	1	343.6	210DET	2011	474	41	1	21	5	13	2	0	367.6	136DET	2012	372	34	1	11	0	6	0	0	341.1	141GBP	2003	442	34	0	21	2	11	1	0	318.8	124GBP	2004	424	40	0	8	2	7	3	0	346.3	152GBP	2005	298	35	0	10	2	11	0	1	293.1	183GBP	2006	301	46	0	23	4	10	1	0	320.9	187GBP	2007	435	36	0	19	1	9	3	2	313.3	140GBP	2008	419	27	0	22	6	6	1	2	334.3	134GBP	2009	461	37	1	30	3	10	1	0	284.4	134GBP	2010	388	47	0	24	3	8	1	0	309.1	97GBP	2011	560	29	0	31	4	7	1	2	411.6	87GBP	2012	433	47	0	18	1	5	1	1	336.8	121HOU	2003	255	19	0	14	1	8	0	0	380.1	184HOU	2004	309	24	0	22	2	8	3	0	341.1	187HOU	2005	260	37	0	7	0	9	0	2	364	188HOU	2006	267	28	0	11	1	11	2	0	337.5	215HOU	2007	379	31	0	11	0	14	3	4	344.2	234HOU	2008	366	25	0	12	1	10	0	2	336.6	190HOU	2009	388	30	2	14	1	13	1	1	324.9	188HOU	2010	390	30	0	13	0	5	0	0	376.9	184HOU	2011	381	44	0	17	1	10	0	1	285.7	155HOU	2012	416	44	1	15	3	14	0	0	323.2	110IND	2003	447	31	0	15	2	15	1	0	299.3	162IND	2004	522	45	0	19	2	17	1	1	370.6	156IND	2005	439	46	0	18	2	13	2	0	307.1	135IND	2006	427	25	0	15	0	11	1	1	332.2	104IND	2007	450	28	2	22	1	15	1	1	279.7	145IND	2008	377	30	0	15	2	11	2	0	310.9	135IND	2009	416	34	0	16	2	9	0	1	339.2	153IND	2010	435	30	0	10	2	11	1	1	341.6	154IND	2011	243	29	0	8	2	9	1	0	370.9	193IND	2012	357	32	0	12	4	3	0	2	374.2	248JAC	2003	276	24	0	15	0	12	1	0	291.1	170JAC	2004	261	37	2	16	0	11	0	0	320.9	153JAC	2005	361	47	0	19	2	9	0	1	290.9	148JAC	2006	371	35	0	20	1	4	0	1	283.6	127JAC	2007	411	37	0	20	2	9	1	1	313.8	116JAC	2008	302	29	1	13	2	4	1	0	330.9	120JAC	2009	290	14	0	15	0	10	0	0	352.3	196JAC	2010	353	26	0	13	0	5	0	1	371.8	252JAC	2011	243	31	0	17	0	11	3	0	313	238JAC	2012	255	20	0	12	0	11	0	0	380.5	208KCC	2003	484	36	0	25	3	12	0	4	356.7	155KCC	2004	483	41	0	13	1	7	1	2	377.3	137KCC	2005	403	29	0	16	1	15	1	1	328.1	172KCC	2006	331	32	0	15	0	15	0	1	328.9	162KCC	2007	226	37	1	14	0	8	1	0	319.4	180KCC	2008	291	10	0	13	2	16	1	0	393.2	238KCC	2009	294	22	1	15	2	13	1	1	388.2	241KCC	2010	366	39	0	14	3	9	0	1	330.2	149KCC	2011	212	29	0	20	2	6	0	0	333.3	177KCC	2012	211	27	1	7	0	6	1	0	356.5	190MIA	2003	311	44	1	22	2	14	1	0	299.2	82MIA	2004	275	36	1	15	0	10	1	1	305.9	112MIA	2005	318	49	3	14	0	17	1	0	317.4	172MIA	2006	260	47	0	8	2	19	1	0	289.1	148MIA	2007	267	30	0	14	1	8	1	1	342.2	116MIA	2008	345	40	1	18	2	12	0	0	329	238MIA	2009	360	44	0	15	1	6	1	2	349.3	164MIA	2010	273	39	1	11	0	8	1	0	309.3	172MIA	2011	329	41	0	16	1	3	0	0	345.1	180MIA	2012	288	42	0	10	0	6	0	2	356.8	209MIN	2003	416	37	2	28	4	7	0	0	334.8	220MIN	2004	405	39	0	11	1	11	1	1	368.9	145MIN	2005	306	34	0	24	3	11	0	2	323.3	149MIN	2006	282	30	0	21	3	15	2	1	300.2	173MIN	2007	365	38	1	15	6	16	2	1	338.1	143MIN	2008	379	45	3	12	0	13	2	1	292.4	81MIN	2009	470	48	1	11	0	13	1	2	305.5	92MIN	2010	281	31	0	15	1	10	1	1	312.6	97MIN	2011	340	50	1	8	0	15	0	1	358.2	174MIN	2012	379	44	0	10	3	12	0	2	350	180NEP	2003	348	41	1	29	5	12	1	1	291.6	121NEP	2004	437	45	0	20	1	16	3	1	310.8	72NEP	2005	379	33	0	10	2	8	0	0	330.2	81NEP	2006	385	44	1	22	0	13	0	1	294.4	140NEP	2007	589	47	0	19	3	12	3	2	288.3	77NEP	2008	410	31	0	14	0	8	0	1	309	94NEP	2009	427	31	0	18	3	10	0	0	320.2	130NEP	2010	518	36	0	25	4	13	1	3	366.5	146NEP	2011	513	40	0	23	2	11	1	1	411.1	108NEP	2012	557	37	1	20	2	21	3	2	373.2	154NOS	2003	340	32	1	14	2	13	0	0	327.1	155NOS	2004	348	37	1	13	0	20	1	2	383.8	178NOS	2005	235	25	0	10	0	9	0	0	312.1	190NOS	2006	413	38	0	11	0	8	0	1	307.3	108NOS	2007	379	32	1	13	3	10	2	0	348.1	170NOS	2008	463	28	1	15	0	7	0	3	339.5	188NOS	2009	510	35	0	26	5	13	3	1	357.8	139NOS	2010	384	33	1	9	2	16	0	0	306.2	120NOS	2011	547	33	0	9	1	7	2	1	368.4	138NOS	2012	461	30	0	15	4	11	0	0	440.1	186NYG	2003	243	45	0	10	2	12	2	0	332.5	140NYG	2004	303	40	0	14	0	14	2	2	324.2	183NYG	2005	422	41	0	17	1	19	1	2	327.5	158NYG	2006	355	32	1	17	2	11	0	0	342.4	123NYG	2007	373	53	0	15	3	10	2	0	305	156NYG	2008	427	42	3	17	2	5	0	0	292	130NYG	2009	402	32	0	13	2	11	1	1	324.9	93NYG	2010	394	46	0	16	0	23	0	0	310.8	157NYG	2011	394	48	2	20	0	11	1	0	376.4	130NYG	2012	429	33	0	21	1	14	1	1	383.4	140NYJ	2003	283	35	1	11	0	9	0	1	332.4	173NYJ	2004	333	37	0	19	2	14	1	1	304.9	167NYJ	2005	240	30	0	21	2	7	1	1	308.8	137NYJ	2006	316	35	0	16	0	9	1	2	331.6	157NYJ	2007	268	29	0	15	0	6	1	3	331.9	176NYJ	2008	405	41	0	14	3	16	2	1	329.4	186NYJ	2009	348	32	0	17	1	14	1	1	252.3	124NYJ	2010	367	40	2	12	3	18	0	2	291.5	79NYJ	2011	377	35	3	19	2	12	0	1	312.1	98NYJ	2012	281	30	1	11	2	12	1	2	324.4	150OAK	2003	270	25	1	14	2	11	0	3	369	134OAK	2004	320	25	1	9	1	9	0	0	371	158OAK	2005	290	36	0	5	0	14	1	0	330.8	183OAK	2006	168	34	1	18	2	5	2	0	284.8	215OAK	2007	283	27	1	18	2	8	0	0	341.6	162OAK	2008	263	32	2	16	0	8	0	5	360.9	182OAK	2009	197	37	0	8	0	12	0	0	361.9	205OAK	2010	410	47	2	12	2	12	1	3	322.8	193OAK	2011	359	39	1	18	0	8	1	2	387.6	182OAK	2012	290	25	0	11	0	8	0	0	354.5	192PHI	2003	374	38	1	13	1	13	0	2	331.7	82PHI	2004	386	47	0	17	2	11	0	0	319.7	102PHI	2005	310	29	0	17	1	10	1	0	325.4	100PHI	2006	398	40	0	19	4	10	1	0	328.1	140PHI	2007	336	37	0	11	0	8	0	0	311.4	124PHI	2008	416	48	1	15	2	14	3	2	274.3	155PHI	2009	429	44	2	25	2	13	2	2	321.1	104PHI	2010	439	39	0	23	1	11	1	1	327.2	103PHI	2011	396	50	1	15	1	9	3	0	324.9	102PHI	2012	280	30	0	8	0	5	0	1	343.2	121PIT	2003	300	35	0	14	2	11	0	2	298.9	101PIT	2004	372	41	1	19	3	13	2	0	258.4	136PIT	2005	389	47	1	15	0	15	1	2	284	86PIT	2006	353	39	0	20	2	9	0	1	300.3	90PIT	2007	393	36	0	11	2	14	0	1	266.4	103PIT	2008	347	51	1	20	2	9	1	0	237.2	113PIT	2009	368	47	0	12	2	10	1	0	305.3	74PIT	2010	375	48	0	21	3	14	0	1	276.8	108PIT	2011	325	35	2	11	0	4	1	1	271.8	79PIT	2012	336	37	0	10	1	10	0	0	275.8	120SDC	2003	313	30	0	13	1	7	0	0	349.6	172SDC	2004	446	29	1	23	1	10	0	1	335	123SDC	2005	418	46	0	10	2	10	0	0	309.2	166SDC	2006	492	61	0	16	0	12	3	0	301.6	141SDC	2007	412	42	0	30	2	18	3	2	320.2	80SDC	2008	439	28	1	15	2	9	1	1	349.9	80SDC	2009	454	35	1	14	1	10	2	1	327	111SDC	2010	441	47	1	16	3	6	0	0	271.6	152SDC	2011	406	32	0	17	1	4	1	1	346.6	138SDC	2012	350	38	1	14	5	14	2	2	326.4	194SFO	2003	384	42	1	23	1	13	0	1	308	172SFO	2004	259	29	1	9	0	12	2	1	342.6	176SFO	2005	239	28	0	16	3	10	1	1	391.2	235SFO	2006	298	34	1	14	1	13	1	0	344.2	215SFO	2007	219	31	1	12	0	9	1	0	346.2	169SFO	2008	339	30	0	12	1	6	0	1	326	185SFO	2009	330	44	0	18	1	15	2	0	326.4	189SFO	2010	305	36	0	15	3	7	0	1	327.8	115SFO	2011	380	42	1	23	1	15	0	2	308.2	154SFO	2012	397	38	1	14	2	11	1	0	294.4	84SEA	2003	404	40	1	16	0	12	2	1	327.4	168SEA	2004	371	36	0	23	3	12	0	0	351.3	146SEA	2005	452	50	0	16	2	11	1	0	316.8	149SEA	2006	335	41	0	12	1	14	1	1	330.3	117SEA	2007	393	45	1	20	1	14	1	3	321.8	151SEA	2008	294	35	0	9	2	11	1	0	378	122SEA	2009	280	28	0	13	2	10	1	0	356.4	186SEA	2010	310	37	1	12	2	10	0	3	368.6	165SEA	2011	321	33	1	22	4	9	0	0	332.2	208SEA	2012	412	36	0	18	2	13	1	1	306.2	147STL	2003	447	42	1	24	3	22	2	0	315.8	148STL	2004	319	34	0	6	0	9	3	0	334.6	128STL	2005	363	41	1	13	2	14	0	1	350.1	190STL	2006	367	34	0	17	1	15	1	0	335.1	195STL	2007	263	31	1	18	2	9	0	1	341.1	185STL	2008	232	30	0	12	0	14	1	0	371.9	238STL	2009	175	25	0	8	1	11	0	0	372.8	241STL	2010	289	43	1	14	0	12	0	0	336.8	252STL	2011	193	39	2	12	1	6	0	0	358.4	188STL	2012	299	52	1	17	4	4	1	0	342.6	248TBB	2003	301	36	1	20	3	13	1	0	279.1	58TBB	2004	301	45	0	16	1	11	2	1	284.5	110TBB	2005	300	36	1	17	0	13	3	0	277.8	126TBB	2006	211	25	0	11	3	9	0	0	329.4	118TBB	2007	334	33	0	16	1	19	1	1	278.4	156TBB	2008	361	29	0	22	3	8	1	2	306.1	152TBB	2009	244	28	0	19	2	10	0	2	365.6	173TBB	2010	341	26	0	19	3	9	0	1	332.7	203TBB	2011	287	23	1	14	3	10	1	0	394.4	182TBB	2012	389	27	0	18	3	8	0	0	379.9	248TEN	2003	435	38	1	21	3	13	2	1	306.3	132TEN	2004	344	32	0	18	1	12	1	0	357.8	138TEN	2005	299	41	1	9	2	11	2	1	319.4	187TEN	2006	324	26	2	17	2	11	3	3	369.7	198TEN	2007	301	40	0	22	2	12	0	0	291.6	163TEN	2008	375	44	0	20	3	11	0	0	293.6	161TEN	2009	354	32	0	20	4	7	0	0	365.6	111TEN	2010	356	40	2	17	1	8	0	2	367.7	173TEN	2011	325	28	0	11	1	12	0	2	355.1	192TEN	2012	330	39	1	19	4	5	1	3	374.9	202WAS	2003	287	27	1	17	0	13	0	1	338.2	161WAS	2004	240	40	0	18	1	8	0	0	267.6	167WAS	2005	359	35	0	16	1	12	1	2	297.9	135WAS	2006	307	19	1	6	0	6	0	2	355.5	142WAS	2007	334	33	2	14	2	10	0	0	305.2	155WAS	2008	265	24	0	13	0	5	0	1	288.8	171WAS	2009	266	40	0	11	0	6	0	0	319.7	163WAS	2010	302	29	0	14	1	13	1	1	389.2	188WAS	2011	288	41	0	13	1	8	0	0	339.8	195WAS	2012	436	32	0	21	3	9	1	0	377.7	175
 
Last edited by a moderator:
can you put yards allowed in that table too?

what kind of league do you want to base ADP on? 12 non-ppr?

 
Drop said:
can you put yards allowed in that table too?

what kind of league do you want to base ADP on? 12 non-ppr?
Yards allowed was already included (YPG). I added ADP from each season. If a team went undrafted, I assigned the lowest ADP + 1.

 
Drop said:
can you put yards allowed in that table too?

what kind of league do you want to base ADP on? 12 non-ppr?
Yards allowed was already included (YPG). I added ADP from each season. If a team went undrafted, I assigned the lowest ADP + 1.
Ok I must have missed that because of the column headers being off initially. Good work, saved me a lot of time. Now we wait for TheMathNinja

 
So no one came up with anything after I took the time to compile and post the data that people were asking for?

 
So no one came up with anything after I took the time to compile and post the data that people were asking for?
Just saw the data for the first time this morning, but something I'll throw out there is that, at least for the 2012 season, and using the MFL D/ST point system of:

Points Allowed

0 = 5 points
1-6 = 3 points
7-13 = 1 point
14+ = 0 points
Total Yards Allowed
0-199 = 5 points
200-249 = 3 points
250-299 = 1 point
300+ = 0 points

Sack = 1 point
INT = 2 points
Fumble Recovery = 2 points
Safety = 2 points
TD (inc. ST) = 6 points
... the bottom five elements (i.e. "base" D/ST scoring) accounted for 86% of the total D/ST scoring.

If, as has been noted upthread, the correlation between ADP and "base" scoring, or between last year's and this year's "base" scoring, are weak at best, I have a hard time believing that the addition of anything less than very generous scoring with regards to yards / points allowed will change that conclusion one whit.
 
i usually dont pickup a defense at all during the draft.. or a kicker.. we draft mid pre-season so id rather grab 2 players and see how they pan out and wait last minute to grab a kicker and dst.. then from then on i dothe weekly waiver wire defensive play of the week

 
I agree with those who are not expecting any revelations to come from this exercise.

Comparing DST to individual players is riddled with flaws, some of which have been pointed out (DSTs harder to project accurately, DST more resistant to injuries, DST more affected by matchups, greater scarcity and need for depth at other positions).

One I didn't see pointed out is the personnel turnover from season to season. For example, consider the Ravens, which turned over several starters on defense. How could there possibly be significant correlation from last season to this season with more than half of the defensive starters being different? Even if they play the same scheme, it is effectively a different defense.

 
So no one came up with anything after I took the time to compile and post the data that people were asking for?
Just saw the data for the first time this morning, but something I'll throw out there is that, at least for the 2012 season, and using the MFL D/ST point system of:

>

Points Allowed

0 = 5 points
1-6 = 3 points
7-13 = 1 point
14+ = 0 points
Total Yards Allowed
0-199 = 5 points
200-249 = 3 points
250-299 = 1 point
300+ = 0 points

Sack = 1 point
INT = 2 points
Fumble Recovery = 2 points
Safety = 2 points
TD (inc. ST) = 6 points
... the bottom five elements (i.e. "base" D/ST scoring) accounted for 86% of the total D/ST scoring.

If, as has been noted upthread, the correlation between ADP and "base" scoring, or between last year's and this year's "base" scoring, are weak at best, I have a hard time believing that the addition of anything less than very generous scoring with regards to yards / points allowed will change that conclusion one whit.
:goodposting:

I was thinking the same thing and was planning to post something similar.

 
So no one came up with anything after I took the time to compile and post the data that people were asking for?
Just saw the data for the first time this morning, but something I'll throw out there is that, at least for the 2012 season, and using the MFL D/ST point system of:

>

Points Allowed

0 = 5 points
1-6 = 3 points
7-13 = 1 point
14+ = 0 points
Total Yards Allowed
0-199 = 5 points
200-249 = 3 points
250-299 = 1 point
300+ = 0 points

Sack = 1 point
INT = 2 points
Fumble Recovery = 2 points
Safety = 2 points
TD (inc. ST) = 6 points
... the bottom five elements (i.e. "base" D/ST scoring) accounted for 86% of the total D/ST scoring.

If, as has been noted upthread, the correlation between ADP and "base" scoring, or between last year's and this year's "base" scoring, are weak at best, I have a hard time believing that the addition of anything less than very generous scoring with regards to yards / points allowed will change that conclusion one whit.
Doing a quick multivariate on the base scoring, I am getting an r-squared of under 10%.

 
i usually dont pickup a defense at all during the draft.. or a kicker.. we draft mid pre-season so id rather grab 2 players and see how they pan out and wait last minute to grab a kicker and dst.. then from then on i dothe weekly waiver wire defensive play of the week
This is pretty much how I operate. I've not drafted a kicker in maybe six seasons, and the defences that I feel confident will be good that are somewhat highly ranked are almost always all off the board. I'd rather take a flyer skill position player that might end up having marginal value if someone gets injured in pre-season than just take a defence for the sake of it, unless there's some mediocre D that has a real nice early schedule

 
I also think the difference between individual players and D/STs makes any comparison between their VBD numbers highly problematic.

But this thread has made me think differently about how to deal with D/STs in auction drafts, where they are usually super-cheap. If a couple extra dollars can get me my pick of the D/STs, it might actually be worth it.

So I'll be interested in seeing the results of the Math Ninja's analysis.

 
I have an article in place that I think should answer this question. I'll link to it once it's live at FBG (I haven't written much yet, but I have done all of the analysis.)

 
So how do we get better at predicting top D/STs to be rewarded enough to make up for the risk in the long run?
If someone can find a good-looking data set for me, I can run some good numerical analysis and get back to you guys pretty quickly. What I'd be interested in seeing in the data set is: Yards allowed, points allowed, turnovers forced, TD's, Dodds' preseason estimate (or some other top picker), ADP, final fantasy points for the year. What I can then do is create a formula based on Multiple Linear Regression that best estimates a team's fantasy value for the year as a function of last year's points, last year's TD's, last years yards and points allowed, last year's TO's, and preseason expectation (and perhaps ADP). The formula will uncover which metrics are actually useful and which are basically totally random; and I can also tell you the strength of its predictive power. Analysis is easy for me; finding data is not.
Pretty much all the data you will ever need - http://www.pro-football-reference.com/

 
I have an article in place that I think should answer this question. I'll link to it once it's live at FBG (I haven't written much yet, but I have done all of the analysis.)
Nice article. I don't think I can have a committee set in stone in the pre-season. Things like Ryan Lindley happends and I have to abondon ship and chase plays like that. In a league where I didn't draft a defense, I'm starting the season with IND-MINN while dropping IND for Atl week 5.

The schedule goes Oak- Mia/Chi- Cle- Jac- Jets- Car

That works for me if Sea doesn't fall in my lap. I look for pick-6 opportunities, blowouts where my def can rack up sacks, or low scoring vegas lines.

 
Last edited by a moderator:

Users who are viewing this thread

Top