There are some fantasy football players that believe that the lineup you pick can lose you a game just as much as it can win a contest. Having a player that can give you a consistent performance week after week should be considered more valuable than a player who goes off every third week and then takes two weeks off between those fantastic performances. Consistency has a value, and it does not take much of a leap to understand that players that you can rely on for solid games when you need them (such as in your postseason) are a huge advantage.
Baseball has a term called "Quality Starts" for pitchers, which is a statistic that represents how often a starting pitcher will put up a good (not great, just good) performance in a given game. The bar is set neither high nor low (six innings pitched, three earned runs or fewer) so as to gauge a decent performance. The theory behind it is that if your pitcher gives you a Quality Start, your team has a fighting chance to win a given game.
So now we need to translate this to football. What is "quality" for each position? How do we define a "Quality Start" for quarterbacks or running backs or any other position? Looking back at the 2016 season, at first I considered taking the #12 QB for the year (Derek Carr, 321.55 fantasy points) and dividing his total by 16 for a game average, just like I have used as a baseline in previous years. I saw immediately that this approach was flawed, if only for the fact that Carr missed a game in Week 17 last year. To rectify this dilemma, I considered two options - Carr's numbers divided by 15 games (since no one would have started him in Week 17 anyway) and also ranking all the quarterbacks with at least 12 starts by weekly average, then using QB12 based on that sorting method. Both ways gave me roughly the same answer (Option 1, Carr over 15 games, works out to 21.4, while the 12th average score was 20.9), so I chose this year to baseline on the average of those two answers (21.15). While some of these decisions (targeting QB12 based on two different ranking methods) may seem a bit arbitrary, but many second tier quarterbacks (QB8 to QB18) all fall into a similar range between 20-22 points per start, and a similar range of quarterbacks (QB10-19) exists when you rank them by total points for the season (between 328.1 and 307 points). The best reason of all to use the average of these two methods for 2016 is because of Tom Brady, who only played 12 games due to a suspension but was the second-best quarterback on a per-game basis (just behind Andrew Luck). If we do not consider the average method, Brady's strong numbers are not even considered, so a blended average is definitely the best option. Lastly, overall we are targeting QB12 as a bare minimum of quality, as the 12th QB should be the "worst starter" in your fantasy league.
So now we move on to the next question - one of quantifying the quality. At what point do we decide whether or not a quarterback has given us a quality performance? Here is where it gets a bit murky, but looking at the distribution of QB performances by starters over the season and it becomes evident that the using the 12th QB average and adding or subtracting a percentage gives us a good range for a QB Quality Start.
Using the QB Quality Start range, we can also define a bad performance or an excellent performance as either falling below or exceeding the Quality Start range. Table 1 gives us the fantasy points that it takes to fall in each of the three areas:
|QB Start Type||Fantasy Points|
|Bad Start||0 to 15.6|
|Quality Start||15.7 to 26.1|
Table 1: 2016 QB Quality Start and Fantasy Point Ranges
We have one more issue in this study, and that is we need to sift through all the quarterbacks and only look at QBs that started an NFL game. That means if Marcus Mariota started Week 16 last year but gets hurt (and this did happen), Mariota is subject to the study (in Week 16) and Matt Cassel (in Week 16) is not. That's an important distinction, as several QBs have had great games in relief yet they should not get counted, simply because no one would have started them on their fantasy roster that week with them expected to do nothing but hold a clipboard on gameday. Before we decide on a fantasy lineup, we usually only know the starters for each week, not who might come in if there is an injury.
Pouring over the games week by week, we find 512 starting QB games spread across 54 NFL quarterbacks from 2016. That’s a very important number – only 54 starters. I would dare say that it was a remarkable number, because the number in previous years had usually been in the mid-to-upper 50s and sometimes over 60. The total starters for 2015 was 53, similar to 2014 (54) and 2013 (51) - and even fewer in 2012 with just 47 starting quarterbacks, so this trend is something to note. To put the 54, 53, 54, 51 and 47 starters in perspective, consider that 14 teams had the same quarterback start every game last year (and 17 in 2015), with all six NFC playoff teams last season having the same starter all year long. The AFC was slightly more complicated with no playoff team having the same starter for every regular season game, but four clubs had their primary starter under center at least 14 times, with Miami's Ryan Tannehill just missing that mark with 13 and Tom Brady starting all 12 of the games he was eligible. That continues the three-year trend of not even half of the NFL teams needing to use more than one starting quarterback throughout the entire regular season. Only four teams needed more than two starters in 2016, with three of the four having consistent quarterback problems (Chicago, Cleveland and the New York Jets). The fourth team, New England, was again complex due to Brady's suspension and the need for two different starters to cover the first four games. Do keep all of this in mind for 2017, as the last five seasons have been unusually healthy years for the quarterback position. This is either a trend in protecting the quarterback more in the NFL or better health by the players overall.
Table 2 shows us the breakdown of all 512 starts and how many of each type of start resulted for each:
|Quarterback||Team||Excellent Start||Quality Start||Bad Start||Total starts|
|Robert Griffin III||CLE||0||2||3||5|
Table 2: 2016 QB Start Types By Player Sorted By Total Starts
That's a lot of info to digest, so let me help. First, we see that 255 of 512 starts fall in the middle - "Quality Starts" - and that only about 20% of all quarterback starts (99) are considered Excellent Starts. Over the years, the results have been mixed, partly due to quarterback healthiness but also due to the change in the NFL towards more productive quarterbacking. The 2016 results are very similar to 2015 (101 vs. 99 Excellent Starts, and 255 vs 246 Quality Starts). Between 2008 and 2009, the threshold for an Excellent Start score jumped over 10% (20.1 and above in 2008, 22.5 and above in 2009). That resulted in fewer Excellent Starts in 2009 and 2010 and highlighted the importance of elite quarterback play in fantasy leagues. More quarterbacks started to perform well in 2011, upping the number of Excellent Starts significantly (150 in 2011, only 127 in 2010). Five seasons ago saw another quantum leap in fantasy quarterback production, pushing the Excellent Start threshold up once again by almost two more points per week (24.4 and above in 2012, 22.6 in 2011), which dropped the number of Excellent Starts again (126 in 2012, 150 in 2013). Four seasons ago in 2013 it was an exact match to 2012, but the pass-happy NFL has raised the bar again in 2014 by almost another fantasy point. The bar rose again in 2015 to 26.8 or more fantasy points - another 1.5 point climb, and over 30% higher than 2007 and 2008. Last season's numbers were similar to 2015, with a small dip in the Excellent Start line (26.2+, down slightly from 26.8+) but still significantly higher than when I first started tracking these numbers. Clearly the NFL is a pass-happy league, and the elite passers are worth their weight in gold.
A summary of these trends is provided in Table 3:
|Year||Excellent Start Score||Excellent Starts||Quality Starts|
Table 3: 2007-2016 Excellent and Quality QB Starts
Now, to dig deeper, let's look at the numbers distributed in two different ways. First, I need to define a valuable starting quarterback in this system. We want a quarterback that will win more fantasy games than lose them, so we want either "Quality" or "Excellent" starts. Using a simple formula of scoring each type of start, we can define the value of a given starting NFL quarterback. Here is the formula:
STARTING FANTASY QB VALUE = EXCELLENT STARTS - BAD STARTS
We can afford to overlook the "Quality Starts" category because they neither win games nor lose them on average - they are just average quarterback performances. We only really care about how often he helps our team vs. how often he hurts it. Giving a "-1" value to bad starts and "+1" to excellent ones does this for us.
On with the results, sorted by value:
|Quarterback||Team||Excellent Start||Quality Start||Bad Start||Total starts||Net Value|
|Robert Griffin III||CLE||0||2||3||5||-3|
Table 4: 2016 QB Start Types Sorted By Value
This is a lot of information once again, but some names leap out at us. It should come as no surprise that Aaron Rodgers (+8) and Matt Ryan (+7) top the chart, as both were fantasy gold mines last year. Kirk Cousins (+5) rounded out the Top 3, and these signal callers were the only quarterbacks able to have +5 or better Net Value and start all 16 games. After these three Pro Bowl quarterbacks comes a few interesting name, highlighted by two top notch options in Drew Brees and Tom Brady, both with a +4 rating, remarkable for Brady given he only started 12 contests. Tyrod Taylor's appearance at +4 is rather surprising, but just speaks to his up and down season and ability to have a high fantasy floor (just one bad start). None of these six quarterbacks had more than two bad starts, and these leaders represent nearly two-thirds of the Net Value (+22 combined) of the quarterbacks with a positive score.
Lastly I will sift through all the numbers for you and get right to the heart of the matter with our final table. Here we have the results sorted by value for the Top 24 QBs on the 2017 ADP list.
|Quarterback||Team||Excellent Start||Quality Start||Bad Start||Total starts||Net Value||ADP|
Table 5: 2016 Top Drafted QBs Sorted By 2017 Value
Note that I drew the line at QB24, a little higher than in years past, but rookie Deshaun Watson of Houston is QB25 on the ADP list, followed by Alex Smith (-5 Net Value on Table 4) - so it is rather safe to say that you should target this Top 24 list for your fantasy squad.
Judging from last season, it would appear that a few quarterbacks mentioned earlier (Tyrod Taylor and Kirk Cousins) have an ADP well above 80 yet are at or near the top of the chart from 2016. Two others are also worth a mention with Blake Bortles (ADP of 127) and Carson Palmer (ADP of 140) looking like solid values if they fall past Round 10-12 of drafts. On the flip side, Russell Wilson (-2), Derek Carr (-2) and Jameis Winston (-3) are all being well overdrafted if you only consider last year’s performances. That's the biggest warning here - there is no reason to believe all of these numbers as indications of 2017 performance, but having this information available should give you more to think about when deciding who you will have leading your fantasy team this year.
Questions, suggestions and comments are always welcome to firstname.lastname@example.org.