Yesterday, I asked how many wins a team full of recent draft picks and replacement-level NFL players would fare. I don’t think there’s a right answer to the question, but it might be a more important question than you think (and you’ll see why on Monday). But I have at least one way we can try to estimate how many games such a team would win.

Neil once explained how you can project a team’s probability of winning a game based on the Vegas pre-game spread. We can use the SRS to estimate a point spread, and if we know the SRS of our Replacement Team, we can then figure out how many projected wins such a team would have. How do we do that?

First, we need to come up with a mythical schedule. I calculated the average SRS rating (after adjusting for home field) of the best, second best, third best… and sixteenth best opponents for each team in the NFL from 2004 to 2011. The table below shows the “average” schedule for an average team:

Opp Rk | Opp Strength |
---|---|

1 | 12.4 |

2 | 9.3 |

3 | 7.2 |

4 | 5.7 |

5 | 4.4 |

6 | 3.2 |

7 | 2 |

8 | 0.6 |

9 | -0.6 |

10 | -1.7 |

11 | -2.9 |

12 | -4.3 |

13 | -5.7 |

14 | -7.4 |

15 | -9.4 |

16 | -12.6 |

On average, the best team a typical team will face in a season (after factoring in home field) will be 12.4 points above average, while the worst will be 12.6 points below average. Once we know the SRS rating of each opponent, if we have a projected SRS rating for our Replacement team, we can use Neil’s formulate to calculate the expected winning percentage in each game. If you sum the sixteen expected winning percentages, you get the number of expected wins.

I installed a new plug-in that is an interactive calculator. There are some deficiencies when using the calculator for this purpose^{1}, but in general I think this is pretty neat. The calculator is dynamic, so if you change any one value, all the other values will immediately change.

If our hypothetical team is 10 points below average, they should be expected to win 4.1 games. Drop them to -13.7, and they’re projected to win just 3.0 games. You can change any of the input values (SRS rating of replacement team, SRS of any opponent) below:

Obviously figuring out the SRS rating isn’t easy, and maybe this just kicks the can down the road. The Chiefs had an SRS of -14.0 last year, while the Jaguars were at -13.0. But those SRS ratings don’t represent the “true” values of the Chiefs or Jaguars, either. Those SRS ratings include all the bad luck and other things that happened to Kansas City and Jacksonville last year. The -14.0 for Kansas City tells us their rating when they finish the year -24 in the turnover battle, but if we simulated the 2012 season 10,000 times, the Chiefs would have a much better turnover ratio in nearly every season. That would cause their SRS to rise significantly.

On the other hand, the Chiefs (and the Jags) are more talented than a mythical team that consists of three first rounders, three second rounders, three third rounders, and 44 late round picks or veterans minimum-type players.

I think I’m settling in on our mythical replacement team finishing 3-13, but am open to other thoughts. More on why this is important on Monday.

- Neil’s formula is an advanced formula that uses the normal distribution; the interactive calculator does not have such capabilities, but I was able to use a best-fit exponential formula. The formula is not the same, so there will be errors as you change the SRS rating farther from -10, but for most values, it will get you close enough to be reasonable. [↩]

{ 11 comments… add one }

Honestly, I think it would solidly be zero wins. I’d bet that the hypothetical equivalent to the 1988 Cleveland Indians of ‘Major League’ fame would be much worse than -10 or even -20 pts per game. The reason is primarily due to interaction effects among players. A bad secondary can be made to look decent by a good pass rush and vice versa. But if they’re both awful, performance spirals downward very rapidly an non-linearly.

The winless expansion 1976 TB Buccaneers are a good comparison. Their SRS was -20 and they even had the benefit of a veteran draft.

I’m sure *some* teams would win zero games, but some teams would draft well. None of the teams with Cam Newton, Andrew Luck, Robert Griffin, Russell Wilson would win zero games, and the Tannehill/Dalton teams would be better than the ’76 Bucs.

If the team doesn’t get lucky in the draft they’re looking at crap at quarterback, no doubt — a Matt Leinart or David Carr I suppose.

Again, though, it’s really hard to get too close to even an estimate on this type of thing. Some people say 4 or 5, some say 0 or 1, and I don’t think you can prove or disprove much. I think 2 to 4 makes sense. I agree with you on the nonlinear thing, but having a true value of -20 is really tough to imagine. I mean if they team whiffed with their draft picks, sure, but not the average team, IMO.

I’m sure *some* teams would win zero games, but some teams would draft well. None of the teams with Cam Newton, Andrew Luck, Robert Griffin, Russell Wilson would win zero games, and the Tannehill/Dalton teams would be better than the ’76 Bucs

I was thinking along these lines myself, it would kind of depend what they got, what system their coach was willing to run (e.g. for example would Alfred Morris have had such a good year, if he wasn’t playing with RG3 and playing the kind of football the Redskins have played in the last year) and who they were up against in their division (e.g. some teams didn’t go so well against the Spread or Pistol offense looks last year).

Let’s say you happen to get Cam, RG3 or Russell (or Brandon Weeden – with more suitable talent around him to run a system like he ran in college – say they get a reasonable WR or someone who can stretch the field and an RB) nobody is going to have a book on this team (I’m thinking like Kingsbury and Sumlin at A&M this year – you had an idea what was coming from their past, but with a freshman QB and coaches applying their system to new talent you didn’t know exactly what was coming if that makes sense).

I’d say your low point would be 2 wins (from a mathematical point of view you can’t project out a 0), with a high point of 6 (e.g. think the 2011 Panthers, new QB with talent, new system against un-expecting defenses who may not figure your game out for the first half of the season).

Like the plug-in by the way.

We are looking at the team 3 years down the road, however. Even TB finished 5-11 by 1978. I’m still going with the Texans best fitting the description.

Great stuff Chase. I’m beginning to back off my 5 games – that was before proper deliberation. 4 games though still seems reasonable to me.

Look at you with the cliffhangers! I’ll be back monday for sure.

But Cam, RG3, Wilson, et al could not come close to their actual rookie performances without Steve Smith, Santana Moss, Pierre Garcon, Golden Tate, Trent Williams, solid line play, etc., etc… On a pure replacement team, even the very best #1 picks wouldn’t appear very good at all. That’s what I was referring to with ‘interaction effects’. Put RG3 onto the 76 Bucs and I’d bet they still win zero games. I don’t think they come close to winning even a single game. Imagine an o-line getting blown up on every single play, or a defense that a) can’t rush the passer b)can’t stop the run AND c) can’t cover receivers. It would be brutal.

I have to nit-pick just a little bit.

The 76 Bucs DID come close to winning a couple games without RG3. (They were tied in the 4th quarter against Miami. They lost by 3 to the also-expansion Seahawks. And they led 9-7 in the 4th quarter against the Bills.)

I haven’t tried this yet on any of the sports I follow yet, but I think this easiest way to resolve this (possibly) is to sim the 1976 Bucs Season, and then put a count in and figure out how many times they win however many games (e.g. some times they will go 0-16 I should imagine, sometimes they may win the two games you mention, other times they may win more). I’ll assume a HFA of 3 and use the SR end of season SRS figures and see how it goes. If it goes positively I’ll post some numbers up.

By the way I do realise it should have been 0-14 and not 0-16 above. Typo on my part.

I went away and ran the numbers on the above basis and came out with the following based on 10,000 sims for the 1976 Bucs Season (huge caveat off of the top – I wasn’t around in 1976 and have absolutely no reference for the 76 season, I haven’t seen or read about any games – all I’ve looked at is the SRS’s):

0 wins – 1304 times (13.04%)

1 win – 3245 times (32.45%)

2 wins – 3057 times (30.57%)

3 wins – 1689 times (16.89%)

4 wins – 573 times (5.73%)

5 wins – 112 times (1.12%)

6 wins – 18 times (0.18%)

7 wins – 1 time (0.01%)

8 wins – 1 time (0.01%)

I ran it for a second time and there is less than 1% variance at 10,000 sims (e.g. so on some sims the Bucs never win more than 6 games but they always seem to win 6 between 0.01% of the time and about 0.5% of the time which is possible but still highly unlikely). Assuming practically the 7 or 8 wins never happens (I am assuming that they were that bad relatively) I guess the question is how much better is your expansion/replacement team than the 76 Bucs? (My gut feeling is they should have an SRS of around -13.0 or -14.0 e.g. a 2012 Jacksonville or Kansas City taking the whole season into account not what they had at the start of it).

Improving the 76 Bucs SRS to -15.0 (but assuming the opponents remain constant) the most likely number of wins now increases from 1 win at 32.45% to 2 wins at 28.85% (marginally ahead of 3 wins at 27.19% – still about 1% variance so it’s marginal which is more likely).

I guess the next question is, put Brady or Manning on this replacement team (with the draft picks, of course), and how many games do they win?