The Fantasy Architect: How to Reduce Randomness in Your League
No two fantasy leagues are the same. There are lots of variables, each with many different options, that can be tweaked to create a unique experience. Your old home league settings probably look a lot different than a highly competitive league that you started recently.
In this offseason series, we’ll take a look at various designs for your league, and how those designs affect the fantasy manager experience. Of course, there is no ‘one answer’: every fantasy manager wants something different. This series is intended, in part, to help you explore what’s out there.
All data, unless otherwise specified, is from nflfastR.
Luck in Fantasy
Anyone who has played fantasy football – or pretty much anything else – recognizes that luck plays a significant role in who wins and who loses. Players have wide ranges of outcomes, and it’s not that uncommon for a star to lay an egg while a 2nd FLEX option drops 20 points out of nowhere.
This sort of luck can be incredibly frustrating, especially when you’re on the losing side. The randomness it introduces brings some measure of fun – I wouldn’t want to play in a league without any luck – but reducing it a bit can’t hurt. If you’re reading about fantasy football in early July (and are part of the #FootClan), you’re probably an above-average fantasy manager. Higher levels of luck are usually bad news for good managers, since worse managers get ‘lucky’ and beat them.
There are a lot of ways to quantify luck, and we’ll be working with a simple one. Imagine a standard 12-team league, where the top 6 teams at the end of the season make the playoffs (the #1 and #2 seeds get a bye week). A quick simulation tells us that 7+ wins will usually get someone into the top half of the league and thus the playoffs.
Now, we can measure the ‘skill’ of a manager by the probability that they will win a specific matchup. Using the Binomial distribution, we can calculate the probability of getting 7+ wins (and thus making the playoffs), and plot this vs. the probability of winning in a given week:
Note that a perfectly average team, with a 50% chance of winning each game, will make the playoffs 50% of the time. This makes sense: 6 of the 12, or exactly half, of the teams make the playoffs, so an average team should have a 50/50 chance. From here, the playoff probability rises with the probability of winning each game. The highest point on the chart tells us that a 57% chance of winning each game yields a 70% chance of making the playoffs.
We’re going to focus on one of the interior points; specifically, a 53% probability of winning each week means a 58.5% chance of making the playoffs at the end of the season. I think this 53% win rate is a good number for a reasonably skilled team, and the rest of this article will focus on making changes to the league settings and seeing how it impacts that playoff probability (58.5%).
One of the easiest tweaks to make is deciding between Standard, Half PPR and Full PPR scoring. The PPR stands for ‘points per reception’, so these scoring systems allocate 0, 0.5 and 1.0 points respectively for each reception a player has. This adds another layer to fantasy analysis: possession receivers, or even pass-catching RBs, get a significant boost. The Fantasy Footballers preferred method is Half PPR, which is what I’ve been using so far.
Intuitively, we would expect Half/Full PPR systems to represent less luck, because there are more variables to analyze. To measure this, we will – you guessed it – simulate a bunch of different lineups to see the impact of changing scoring systems.
To start, we have to figure out how good a reasonably skilled fantasy manager (53% win rate) is in terms of an actual QB/2RB/3WR/TE/2FLEX lineup. I ran the numbers, and a manager who is 3.3% better than average at each position has a win probability of 53%, which is what we defined as a skilled manager. This 3.3% simply means that the skilled manager’s drafting, waiver wire selections and start/sit decisions lead to players that score 3.3% more than the average manager at each position.
Now that we have our base simulation, we can run back the same numbers, with the same 3.3% bump, but with Full PPR and Standard scoring. Remember, with Half PPR scoring, a skilled fantasy manager will make the playoffs 58.5% of the time. The results: in Standard leagues, the skilled manager will make the playoffs 57.7% of the time, and in Full PPR leagues 59.1% of the time.
It looks like moving up the ladder from Standard to Half to Full PPR increases the playoff chances by about 0.6% each step. Is that a smaller improvement than you were hoping for? Let’s try something else.
There’s no real consensus on how big a starting lineup should be, or even the ideal composition of the positions. One would imagine that having a larger lineup – requiring managers to start more players – decreases the luck factor. This has two main channels: skilled managers will have knowledge about a larger population of players, and more starting spots dampens the volatility of the overall lineup score.
We’ll run the same simulation as before, but gradually increase the number of players that must be started in the FLEX spot. The result: each additional FLEX increases the playoff probability of the skilled manager by about 0.5%. Specifically, a skilled manager in a 4-FLEX league has a 1% higher chance of making the postseason than a skilled manager in a 2-FLEX league.
Again, this is a little bit disappointing. FLEX players just don’t score enough to really move the needle, especially as you require more and more FLEX players to be started.
Home Field Advantage
This is something that usually isn’t popular in fantasy, although we’ve used it in my home league and have been happy with the results. The idea is that the home team gets a small, constant number of points added to their score, and is meant to reflect the actual home field advantages that real NFL teams enjoy. We use it for the playoffs, where the higher seeded team gets +7 points.
Now imagine a league where home field advantage is installed for the regular season, and each team has an equal amount of home and away games. Sadly, regardless of the size of the home field advantage, it doesn’t improve the playoff probability for a skilled manager; it even slightly decreases it. This follows similar reasoning to what I found last year with consistency. The bump that a skilled manager gets as the home team is smaller than the hit they take as the away team. Since the skilled manager is already projected to win, they have more to lose when they are away than they have to gain when they are home.
Of course, if we always gave the skilled manager the home field advantage, it would help: each additional point increases their playoff chances by 3.5%, a massive number. This is hard to imagine in practice, though. Is the team with the better record each week awarded the home field advantage? That would make it much harder to climb out of an early season hole. Let’s try one more thing.
Smaller Playoff Field
So far, we’ve assumed that 6 of the 12 teams makes the playoffs which, per simulation, means a team must muster 7+ wins (on average) to crack the postseason. What if fewer teams made the playoffs? It would be harder for a bad team to get lucky and sneak in as the 4th seed compared to the 6th seed.
If we only allow the top 4 teams to make the playoffs, my simulation estimates that a team must get (about) 8+ wins. A skilled manager (with 53% probability of winning each game) has a 37% chance of making it in this case, compared to an average team with a 29% chance. This is a 37% / 29% – 1 = 27.5% proportional increase in playoff chances for a skilled team, compared to 58.5% / 50% – 1 = 17% increase with 6 teams in the playoffs.
Similarly, allowing 8 teams into the playoffs – which is what my home league does, despite my protests – means a skilled team has a 78% chance of making it against 70% for an average team. This is a 78% / 70% – 1 = 11% increase for a skilled manager.
This is our strongest lever, then: increasing/decreasing the amount of teams in the postseason has a massive effect on the (proportional) advantage that a skilled team has.
We all know that luck is a big part of fantasy, but hopefully this article convinced you that it’s difficult to remove luck from the equation. Increasing lineup size or moving to Half/Full PPR certainly increases the chances for skilled players, but not by much. Home field advantage doesn’t help at all. The only intervention that has a strong effect is, unsurprisingly, how many teams can make the playoffs. This is also, of course, the policy that would have the biggest change on league experience.
Hopefully you found something interesting, and learned a bit more about luck and how to address it in your league. While it’s nice when the best team wins, don’t forget that luck, crazy comebacks, bad beats and even tilting add something incalculable to fantasy football.
Questions? Message me on Twitter.
I feel like TDs are harder to predict and are more random week to week. Would increasing points for rec/rush yards and/or decreasing points for TDs affect the luck factor?
victory points is the answer to solving randomness
Interesting, like in Catan?
good article. i played in about 50 Dynasty leagues last season (I have a lot of free time, lol) & the luck factor is highlighted by the fact that in those 50 leagues, only 4 #1 seeds actually won the ship. I had a 13-1 team lose in the 1st round. I had a 14-0 team lose in the 2nd round. I had a 7-7 team win the ship. Luck is huge & you just have to roll with it. I’ve found the key to help offset bad luck is depth, depth, depth. Having many decent options on your bench throughout the season to be able to start is huge in mitigating luck, especially towards the end of the season when so many players are injured.
Thanks Tom! 50 leagues sounds awesome haha, and some interesting data…an 8% win rate for #1 seeds is wild. The depth piece also definitely makes sense.