Dr. Bob’s Week 2 NFL Survivor Strategy
Welcome back to Dr. Bob’s Weekly Survivor Strategy blog. Congratulations to everyone who tailed last week, as both Denver and Jacksonville pulled out the win to advance. However, considering that every favorite of >2 points covered (except for New England – who we cashed a Strong Opinion against), that is not that impressive. All 10 of our $1,000 entries remain, carrying a combined equity value of ~$10,300 (+3%).
Even though our picks advanced, an objective analysis shows that our picks on Denver and Jacksonville did not provide as much EE as we expected, since they were picked by others at a higher rate than we forecasted (38% pick rate for Denver as opposed to 34% forecasted, 8% pick rate for Jacksonville compared to 2.5% forecasted). As much as I would like to think our picks move the Survivor contest market[1], this was mostly due to over-projecting the pick % for both the Eagles (14.5% projected pick % vs 9.7% actual) and especially the Commanders (14.7% projected vs 6.1% actual). More entries than expected avoided picking these teams, likely due to their status as the largest favorites in the Circa Survivor shortened weeks (Thanksgiving and Christmas).
We know some readers were questioning if we account for the double Thanksgiving/Christmas weeks in our analysis (we do), and if so, how. In this issue of the Dr. Bob Survivor Strategy Blog, we will discuss our thought process for the shortened weeks, as well as preview our EE-maximizing and contrarian picks, and our 90%-optimal 10-team portfolio selections.
Back to the topic of what I like to call Circa Survivor’s “special” weeks (Thanksgiving and Christmas). For those that are unaware, both weeks 13 and 17 are split in two, with weekday (Th/Fri) games counting separately from the rest of the weekend games. Since the shortened weeks can have as few as 3 games (and therefore there are only 6 teams to choose from), smart players begin strategizing for these weeks well in advance.
Last week, we talked about our 3 keys to NFL Survivor analysis: Projected Win Percent, Future Value, and Projected Pick Percent. The “special” weeks factor into the projections we make for both Future Value and Projected Pick Percent. Teams with high win probability in future games have high future values, but we also add an extra factor for teams that have a high probability of winning in weeks where not many other teams have high probability of winning. In “special” weeks with only 3 games, the projected favorites of Philly (vs. Chicago on Thanksgiving) and Washington (vs. Dallas on Christmas) have two of the highest future values of any team. We take these future values into account when projecting expected Pick %, but last week that was apparently not factored in enough.
Our model improves week-over-week as we give it more data, and we are accounting for a higher-than-expected number of holdouts going forward. However, I would like to make the case that most Survivor contestants are overweighting the value of these “special” week favorites. There are two main reasons we believe saving these teams may be sub-optimal.
Our first reason is overcrowding. If everyone (or a large portion of the field) is going to save these teams for those weeks specifically, it no longer becomes optimal to pick those teams in the shortened weeks. A team with a 75% win probability will not be worth picking if 75% of the field is also selecting them. Sometimes, the optimal EE selection in the shortened weeks is actually the smallest underdog!
The second reason, changing expectations, is somewhat self-explanatory. In our future value calculations, we discount the value of winning further into the future, because circumstances change over the course of a season. Projections for the current week (which most entries use to forecast out the entire rest of the season) have less predictive power as games are played, injuries happen, and we learn more about schemes and personnel. By the end of the year, we expect our team ratings to have changed on average by 3.5 points, and it’s not just our projections – market ratings also change on average 3.3 points. Some people are playing their entries like Washington is going to be a sure win for Thanksgiving, but we are one Jayden Daniels injury away from it basically becoming a toss-up.
The point we are making here should not be mistaken with the popular survivor strategy of “All I need to do is win this week.” Circa Survivor “special” weeks need to be accounted for in your season-long strategy (after all, you do need to go 20-0 to get a piece of the prize). You will need at least one viable option available for each week. That being said, it often makes sense to pick a team that can win today rather than save a team that might be able to win in the future.
Onto our picks for this week.
The EE-optimal Selection: Buffalo Bills (@NYJ)
(Win Prob: 73%, Pick Prob: 10%)
Based on last week’s blog, I’m sure most of you thought we were going to go with the Ravens this week. However, we are lower on the Ravens this week than the market. The Bills are playing an in-conference road game, and will drive more people to choose other teams, leaving value to those that are willing to risk the familiarity inherent in a conference matchup.
We believe the Bills provide more value now and it’s better to save the Ravens to be used down the line. Both the Bills and the Ravens are coming off an intense Sunday night game and have high Future Value scores, so arguments could be made for saving them both for future weeks (in that case – pick the Cardinals).
Based on our model, a $1,000 pick on the Bills carries about $1,070 (+7%) worth of EE.
The Contrarian Play: Cincinnati (vs. JAX)
(Win Prob: 64%, Pick Prob: 2%)
Last week, our successful contrarian pick was on the Jaguars, but this week it is against them. The Bengals barely squeaked by the Browns week 1, but we believe that was due more to the Browns being undervalued than the Bengals being overvalued. Chase was a non-factor in a slow start for what we believe is a top 10 offense, and we expect him to be more involved in the game plan this week. The Jaguars O-line ranked last by our metrics coming into the season and they will face a much tougher matchup than Carolina in the Cincinnati D-line this week. At only a 2% projected pick rate, the Bengals offer a good amount of contrarian value.
The 90% Optimal Portfolio
Here is our portfolio for this week, with an EE of $10,550 (from $10,300 starting) and a 90% downside EE of $9,400. For details on our selection strategy for our 90% Optimal Portfolio, see our Week 1 Blog Post.
| Team | Picks |
| BUF | 2 |
| BAL | 2 |
| ARI | 2 |
| CIN | 1 |
| DET | 1 |
| DAL | 1 |
| LAR | 1 |
Appendix:
Expected Team Pick Probability for Week 2 in Circa Survivor:
| Team | Predicted Pick % | Team | Predicted Pick % |
| BAL | 24.60% | JAX | 0.10% |
| DAL | 13.62% | NE | 0.09% |
| ARI | 12.50% | KC | 0.06% |
| DET | 9.80% | IND | 0.05% |
| BUF | 9.32% | TB | 0.04% |
| LAR | 8.78% | TEN | 0.04% |
| MIN | 6.84% | LV | 0.03% |
| SF | 3.79% | SEA | 0.03% |
| LAC | 2.49% | CAR | 0.03% |
| GB | 2.00% | ATL | 0.03% |
| CIN | 1.95% | CLE | 0.03% |
| PIT | 1.28% | NYJ | 0.03% |
| HOU | 1.21% | NO | 0.03% |
| DEN | 0.61% | NYG | 0.03% |
| MIA | 0.31% | WAS | 0.02% |
| PHI | 0.27% | CHI | 0.01% |
| NO PICK | 0.001% |
[1] To believe we have any effect on the Survivor market, it would be likely we would have influenced the Millions market as well. This was not the case, as our 3 released sides for Week 1 (Cowboys +8, Browns +5, and Raiders +2.5) were some of the least selected. Fun Fact: All Circa Millions selections went ~49.8% (240,373-251,262-9,480 by my calculations) across the entirety of last season, and the 5 most popular Week 1 selections this year all lost (the top 11 selections went 1-10-0). Our Week 1 side selections went 3-0, and we have a proven track record of 57% dating back to 2016 – sign up today for any of our NFL packages to give yourself and edge!