Thursday night. Half your group is on the same 10-point favorite, and the chat has already moved on to next week.
That team should win. It usually does.
Usually is the problem. Survivor punishes you for the one week in five when the sure thing is not. So we went and counted how often those weeks come.
Twenty regular seasons of point spreads and final scores. Sixteen seasons of survivor pick shares. Here is how often heavy favorites lose, and what it cost the crowd when they did.
Key findings on heavy favorites
The chalk wins four of five
By the numbers
- 18.9% — Favorites of 7 or more points who lost outright (298 of 1,576 games) (nflverse games file, 2006-2025 NFL regular seasons)
- 8.0% — Favorites of 14 or more who lost (14 of 174) (nflverse games file, 2006-2025 NFL regular seasons)
- 13.9% — Weeks the single biggest favorite lost (345 weeks) (nflverse games file, 2006-2025 NFL regular seasons)
- 60.0% — Weeks with at least one 7-point favorite losing (207 of 345) (nflverse games file, 2006-2025 NFL regular seasons)
- 79.1% — Win rate for the week's most popular survivor pick (219 of 277 weeks) (Yahoo pick shares via SurvivorGrid, 2010-2025 regular seasons)
- 17.9% — Median share of a week's survivor picks that landed on a losing team (Yahoo pick shares via SurvivorGrid, 2010-2025 regular seasons)
What the numbers say about your NFL survivor pool:
- Heavy favorites, 7 points or more, lost 18.9% of the time. Roughly one in five.
- The loss rate fell as the spread grew: 22.7% at 7 points, 8.0% at 14 or more.
- Road favorites of 7+ lost more often (22.1%) than home favorites of 7+ (17.8%).
- In most weeks, at least one heavy favorite lost somewhere on the slate.
- The crowd's favorite pick won 79.1% of weeks. The team with the highest listed win probability won 84.5%.
- In a typical week, 17.9% of survivor picks landed on a losing team. In 63 of 277 weeks, it was 40% or more.
How we built the sample
Two public datasets, no guesses
This post uses two sources. Both are saved with the post, along with the scripts, so every figure can be rerun.
Spreads and results. The nflverse games file (nflverse/nfldata on GitHub), downloaded October 30, 2026. We used every regular-season game from 2006 through 2025 with a final score and a spread, 5,199 games. Four pick'em games with no favorite were dropped, leaving 5,195. The spread is the file's spread_line field, which nflverse's data dictionary describes as the closing spread, sourced from Pro-Football-Reference. A second public archive, Sports Odds History, matched it exactly on 61% of games, mostly within a half point otherwise. On its lines, favorites of 7 or more lost 19.3%, close to the 18.9% here.
Pick popularity. SurvivorGrid's weekly archive pages for 2010 through 2025, 277 regular-season weeks, fetched October 30, 2026. Each page lists the share of survivor picks on every team. We used the Yahoo pick shares, the only source on the site for all 16 seasons. SurvivorGrid also lists each team's win probability, which it derives from consensus moneylines, and the final result. In late 2025 the listed Yahoo shares sometimes total more than 100% (up to 112%), so we scaled every week to 100%.
Definitions:
- Heavy favorite: favored by 7 points or more.
- Biggest favorite of the week: the largest spread that week. When two games shared it, they split the result.
- Lost: an outright loss. Ties (6 among heavy favorites) count as games the favorite did not win but are not counted as losses.
- Most popular pick: the team with the biggest Yahoo pick share that week.
- Share of picks lost: the Yahoo pick share on teams that lost or tied that week. In the pick data, a tie counts as a pick that did not win (26 team-weeks). It measures picks made that week, not the size of the whole original field.
- Early season: Weeks 1-4. Final week: Week 17 through 2020, Week 18 since 2021.
For team-by-team records, including home, away and underdog splits, TeamRankings keeps NFL win trends data.
How often NFL favorites lose straight up
One game in three goes the other way
Start wide. Across 5,195 games, NFL favorites lost 1,734 times, a 33.4% loss rate. NFL underdogs won one game in three.
That is every favorite, including the 1-point ones nobody in a survivor pool should be near. The useful question is narrower: what about the teams you would actually pick?
How big favorites fare by expected margin
Seven is not ten, ten is not fourteen
The spread is the expected margin of victory. The bigger it is, the rarer the loss, and the drop is not even.
Heavy favorites' losses by point spread, 2006-2025 regular seasons
| Spread | Games | Losses | Loss rate |
|---|---|---|---|
| 7 | 331 | 75 | 22.7% |
| 7.5 to 9.5 | 559 | 128 | 22.9% |
| 10 to 13.5 | 512 | 81 | 15.8% |
| 14 or more | 174 | 14 | 8.0% |
Favorites of 7 to 9.5 points lost more than one game in five. That is not a safe pick. That is a pick with a real hole in it.
Get to 10 or more and the loss rate falls to 15.8%. At 14 or more it is 8.0%. Those teams lost 14 times in 20 seasons. Rare is not never.
Home favorites vs road favorites
Travel adds a little risk
Among heavy favorites, home teams lost 17.8% of the time (209 of 1,174). Road teams lost 22.1% (89 of 402).
Across all favorites, the gap was smaller: 32.6% for home favorites, 34.8% for road favorites.
Road heavy favorites carried smaller lines on average (9.1 points against 10.0), which explains part of the gap. But in every spread band above, road favorites lost as often as home ones or more often.
In theory the spread already prices in where the game is played. History says the road version slipped a bit more often anyway.
Loss rate for favorites of 7+ points
| Loss rate for favorites of 7+ points | Value |
|---|---|
| Home favorites | 17.8% |
| Road favorites | 22.1% |
Source: nflverse games file, 2006-2025 NFL regular seasons
NFL upsets that hit survivor pools hardest
The weeks everyone remembers
NFL upsets of heavy favorites are not rare events. In 207 of 345 weeks, at least one favorite of 7 or more lost. The biggest NFL upsets in our sample came against teams favored by more than two touchdowns.
Biggest upsets by point spread, 2006-2025 regular seasons
| Season, week | Favorite | Spread | Result |
|---|---|---|---|
| 2019, Week 17 | Patriots | 17.5 | Dolphins won 27-24 in New England |
| 2020, Week 15 | Rams | 17.5 | Jets won 23-20 in Los Angeles |
| 2018, Week 3 | Vikings | 16.5 | Bills won 27-6 in Minnesota |
| 2009, Week 13 | Steelers | 14.5 | Raiders won 27-24 in Pittsburgh |
| 2020, Week 15 | Steelers | 14.5 | Bengals won 27-17 in Cincinnati |
| 2021, Week 9 | Bills | 14.5 | Jaguars won 9-6 in Jacksonville |
The spread does not tell you how many entries went down. The pick shares do. The worst single week in our 16-season survivor sample was 2015 Week 2, when 93.7% of Yahoo picks landed on teams that lost. The Saints, picked by 49.8%, fell 26-19 to the Buccaneers.
Other weeks that cleared 80%: 2019 Week 10 (89.6%), 2013 Week 10 (88.9%) and 2024 Week 3 (82.9%). When a week like that wipes out a field, rules like revivals & tiebreakers decide whether the contest keeps going and who takes it.
Early-season upsets
Early favorites slip more often
In Weeks 1-4, favorites of every size lost 36.8% of the time (461 of 1,254). From Week 5 on, it was 32.3%.
Heavy favorites held steady, though. Favorites of 7 or more lost 19.0% in Weeks 1-4 and 18.9% after. The early chaos lived in the closer games.
That is a case for an early heavy favorite, not against one. The risk is about the same as in November.
Late-season rest weeks
The last week was calmer than its reputation
The last week of the regular season has a reputation: clinched teams rest starters and favorites fall. The numbers did not show it.
In the final week, favorites of 7 or more lost 14.9% of the time (20 of 134). In every other week, they lost 19.3%. Favorites of every size lost 26.2% in the final week, against 33.8% the rest of the season.
Treat that as a small sample: 134 games, where a gap of four or five points is within what chance alone can produce.
One likely reason, which we did not measure: the spread is set knowing who is expected to play. The danger is news that breaks after you pick, so check it before picks lock at kickoff.
Pick popularity vs survival rate
The crowd's pick is not the safest pick
Popular survivor pool picks did win more often than unpopular ones. The table shows how teams fared by the share of Yahoo picks they drew, next to the average win probability SurvivorGrid listed for them.
Survivor pick share vs results, 2010-2025 regular seasons (teams with at least 1% of picks)
| Share of picks | Team-weeks | Won | Listed win probability |
|---|---|---|---|
| 40% or more | 105 | 82.9% | 82.4% |
| 20% to 39.9% | 311 | 78.5% | 77.7% |
| 10% to 19.9% | 431 | 74.5% | 75.2% |
| 5% to 9.9% | 475 | 74.1% | 72.2% |
| 1% to 4.9% | 1,271 | 67.3% | 65.0% |
In each band, actual win rates landed within about two points of the listed probability. The crowd was not fooled. It just was not always on the best team.
The week's most popular pick drew 37.8% of picks on average and won 79.1% of the time. The team with the highest listed win probability won 84.5%, yet drew only 25.8% of picks on average. They were the same team in 115 of 277 weeks.
Part of that is smart survivor pool strategy: entries save the strongest team for a later week, or already used it. Part of it is the crowd chasing a story. Either way, the most popular pick lost 58 times in 277 weeks.
Most popular picks that lost, share of Yahoo picks
| Most popular picks that lost, share of Yahoo picks | Value |
|---|---|
| Titans, 2013 Week 10 | 77.4% |
| Giants, 2011 Week 5 | 70.6% |
| Vikings, 2018 Week 3 | 64.2% |
| Buccaneers, 2025 Week 14 | 57.0% |
| Lions, 2013 Week 12 | 56.2% |
Source: Yahoo pick shares via SurvivorGrid, 2010-2025 regular seasons
Watch out
When the most popular pick loses, the field shrinks fast. That can help you if you were elsewhere. It only hurts you if you were on it, so the question is never just who wins, it is who wins that the crowd is not on.
Use the data in your NFL survivor picks
Pick the winner, then pick the spot
Lists of possible NFL upsets this week show up everywhere by Thursday. History gives you a steadier frame for your NFL survivor picks:
- Price the risk. A 7-point favorite lost 22.7% of the time. A 14-point favorite, 8.0%. Know which one you are holding.
- Lean home when spreads are close. Heavy road favorites lost 22.1%, home ones 17.8%, and road teams lost as often or more often in every spread band.
- Do not fear September heavies. Favorites of 7+ lost at the same rate early as later.
- Do not assume the final week is a trap. Big favorites did not lose more often in the final week, though that sample is small. Check that the line still reflects who is playing.
- Check the crowd. The most popular pick won 79.1%. If the safest team is less popular, it may be the better spot.
That is the core of an NFL survivor strategy: survive the week, but think about where the rest of the field is standing. Our NFL survivor pool tips cover saving teams and reading the field in more detail.
Every good NFL survivor pool strategy starts with the same question: how likely is this team to win? Now you have the history behind the answer.
Ready to put it to work? In the Splash Survivor game, you pick one team to win each week. A right pick moves you on, a wrong one spends a life, and by default you cannot use the same team twice. Picks stay hidden until kickoff, and pick distribution reports show where the field went once games start.
Commissioners can set 1 to 5 lives, multi-pick weeks, revivals and buybacks, so one ugly upset does not have to end the season for your group. Your competition is friends and other fans, not the house. Check where you can play before you set one up.
Know the risk. Watch the crowd. Live to see another week.
Frequently asked questions
Quick answers on favorites & survivor picks
Favorites of 7 or more points lost 18.9% of the time in the 2006-2025 regular seasons, 298 of 1,576 games. The rate drops as the spread grows: 22.7% at exactly 7, 15.8% at 10 to 13.5, and 8.0% at 14 or more. In 207 of 345 weeks, at least one heavy favorite lost.
Not automatically. From 2006 to 2025, the week's biggest favorite lost in 13.9% of 345 weeks. It is usually the safest pick that week, but using it now means you cannot use it later. Weigh this week's risk against the teams you will need in harder weeks and against where the crowd is going.
Yes, slightly. From 2006 to 2025, favorites of 7 or more lost 17.8% of the time at home (209 of 1,174) and 22.1% on the road (89 of 402). Across all favorites, home teams lost 32.6% and road teams 34.8%. At similar spreads, the home favorite has been the steadier pick.
Two 17.5-point favorites lost in our 2006-2025 sample. The Dolphins beat the Patriots 27-24 in New England in Week 17 of 2019, and the Jets beat the Rams 23-20 in Los Angeles in Week 15 of 2020. The Bills' 27-6 win over 16.5-point favorite Minnesota in 2018 Week 3 came next.
In Yahoo pick data from 2010 to 2025, the median week saw 17.9% of survivor picks land on a losing team, and the average was 24.2%. In 63 of 277 weeks it was 40% or more. These are shares of that week's picks, not of every entry that started the season.

