Editor’s Note: This article is submitted by Member Corner writer, Nick Canzanese (known in the DLF forum as ‘Zounder’). We look forward to seeing much more of Nick’s work in the future.
Most owners believe that “Win Now” trades, or moves that sacrifice overall value in the name of present value, are a great way to control win/loss outcomes in the short term to make a playoff and championship push. Is that really the case? How likely is a trade of that nature to be the difference between a win and a loss in any given matchup? Adam Harstad of Footballguys has written on this topic, and it has inspired me to investigate as well.
Margin of Victory
My dynasty league has five years’ worth of weekly scores and uses a regular lineup and scoring system. (0.5 PPR, QB/RB/RB/WR/WR/WR/TE/FLEX/DEF/K) I feel this league is as good a representation of a normal dynasty league as I could possibly find, with a big sample size of 420 regular-season games (or 840 individual weekly scores). I recorded every individual weekly score and the results of every head-to-head matchup, including margin of victory, so we can see just what we’re up against in trying use Win Now trades to turn what would otherwise be a loss into a win.
This tells us that 50% of head-to-head matchups resulted in a margin of victory of 20.59 points or less. Only three non-quarterbacks hit that point per game threshold (David Johnson, Le’Veon Bell, Ezekiel Elliott.) That means that on average, having an extra top three running back added to a roster wouldn’t have made a difference in 50% of all losses!
Furthermore, we also have to remember that when making a “Win Now” trade, you are not adding a player to your lineup, you are replacing a player in your lineup. The actual expected point increase from a Win Now trade is much less than it appears, unless you weren’t starting anyone in that lineup spot.
Additionally, it’s not likely to add a megastar to your lineup via trade. A ten point per game increase is more likely, and that would turn a loss into a win only 27% of the time. A five point per game increase would turn a loss into a win only 15% of the time.
The reason the average Margin of Victory is so large (and the reason the chances of turning a loss into a win is so small) is because of high variance in week-to-week fantasy scores. Owners should account for this when considering a Win Now trade.
Win Probabilities
Let’s get more specific: How much does adding a specified amount of expected point per game increase win probability? To answer that, I created a distribution of all 820 weekly scoring results. We need to know the shape of the distribution because adding points per game to a lineup will matter more for a middling team than an elite one.
It’s doubtful the sample distribution from my league is representative of the true distribution of weekly point totals, so I used statistical software to find the best fit distribution (generalized logarithmic) and then calculated probabilities based on that distribution instead. This distribution is much more likely to be representative of other leagues. The results are as follows:
For example, a team that is expected to score 110 points per game has a win probability of 47.34% any given week in the regular season. Adding five points per game to the expected score would increase their weekly win probability by 8.50%. It’s important to note that most teams in the sample averaged between 100 and 130 points per game through a whole season, but I included the extremes to give you the full scope. Here’s what I took from this data:
-These numbers are lower than one would like. You will need to play many games for an expected win increase of five percent to 15% to manifest itself. For example, making a move to increase your win probability by ten percent is likely to net you at least one more win throughout the regular season. If you’re making the move with only a few weeks left at the end of the season or at your league’s trade deadline, it’s not likely to ever have an impact. Timing matters when making these kinds of trades. The earlier you pull the trigger, the more chances you get for the trade to have an impact. (There are obvious exceptions. For example, if you’re unable to fill a starting lineup due to injuries and you’re pushing for a playoff spot, it’s probably worth the risk, no matter how many games are left.)
-Consider that making a Win Now trade doesn’t necessarily guarantee that you will net those extra points per game. You may even lose points! So first, you have to hope that the trade actually gives you the points you think it will. Then, you have to hope the trade turns enough losses into wins. Compounding those probabilities gives an even bleaker scenario. Owners are likely not to get what they want out of a Win Now trade, especially those made later in the season.
-There appears to be a sweet spot in terms of getting the most bang for your buck. The hump of this data distribution is close to the mean of 112.7 points per game, so you get the largest increase in win probability per point added by getting over that hump (passing the 112.7 expected point threshold.)
-These win probabilities are only accurate for the regular season, since that’s how the sample was collected. (The sample size for the playoffs was too small.) The average point total is likely much higher in the playoffs, so that will change the position of the “hump” in the data distribution. Thus, teams with higher expected point totals are getting more bang for their buck than the chart indicates. It’s worth noting, though, there aren’t that many opportunities for the move to pay off during the playoffs, since there are only two or three games.
-This chart would look different if this were a PPR league, if there were more/less owners, more/less starting spots, etc. But I think it would lead to the same general takeaway. Regardless of what league you play in, this is actionable information.
We can also look at this from a reverse perspective: How much does losing a specified amount of expected points per game decrease win probability?
For example, a team that is expected to score 115 points per game has a win probability of 55.84% any given week in the regular season. Subtracting five points per game from the expected score would decrease their weekly win probability by 8.5%.
Teams scoring 125+ points or more can afford to reduce their weekly expected points and it’s not likely to impact any outcomes, especially if done late in the season. Now, I wouldn’t take this idea to the extreme and sell your half of starting lineup for rookie picks after you clinch a playoff spot. But most contending owners automatically hold expiring assets because they feel they have to. I think this data shows that you can sometimes afford to make your team better in the long term without risking short term outcomes.
I did this twice last season. After week one, I traded for a recently-injured Keenan Allen and gave up a still-healthy RB1 Doug Martin to a contending owner. My next best running backs were Ryan Mathews and Duke Johnson. Other owners were stunned that I would make such a move despite my team being a preseason favorite, having won the championship the year prior. Remember when I mentioned that Win Now trades aren’t even necessarily guaranteed to net you more points in the short term? Martin got hurt, that owner missed the playoffs and my team’s outcomes for the rest of the season were not impacted at all.
Then in midseason, with my team struggling in the middle of the standings and needing a big late season push, instead of making a Win Now trade, I did the opposite and traded Brandon Marshall for two second rounders to another contender. That owner also traded a first rounder and Kevin White for DeMarco Murray. My team was left for dead and that owner’s team become a playoff lock. Yet, I finished second while the other owner finished fourth, out of payout range.
Conclusion
Fantasy football is random. By making Win Now trades, many owners dream that the control they have over weekly outcomes is far greater than it really is. Exploit this belief by letting the other teams in your league make the Win Now moves, especially if it’s late in the season. And don’t afraid to be the one on the other end, reaping long term value at the expense of the misplaced confidence of your competitors, even if you’re a playoff contender. Do what’s best for your team in the long run and let the wins follow.
- 2018 Summer Sleeper: New York Jets - July 29, 2018
- 2018 Summer Sleeper: Pittsburgh Steelers - July 24, 2018
- 2018 Rookie Profile: Michael Gallup, WR Colorado State - April 21, 2018
[quote]Furthermore, we also have to remember that when making a “Win Now” trade, you are not adding a player to your lineup, you are replacing a player in your lineup. The actual expected point increase from a Win Now trade is much less than it appears, unless you weren’t starting anyone in that lineup spot.[/quote]
BEST thing you said. Took me awhile in my leagues but I now see those teams that are constantly dealing future picks to try to win now and the opportunity to profit off of it by trading away veterans and spare parts for youth and picks.
I think your data regarding the average weekly win is a little skewed. When looking at all games played one has to take into account that there are a number of teams in the league that are rebuilding and not competitive. This will lead to higher average margin of victory than if everyone was in win now mode. I think a lot could be gained by running this analysis on playoff games only, as these are all teams that are in the running and competitive today. logically the average margin of victory would be much smaller at that point and these win now trades would be much more likely to play an impact.
Don’t get me wrong, I love to be on the long term value side of these trades, however I think the data is skewed by using all games played versus the only playoff games or games between two teams that are actively competing to win the league that year.
Thank you for reading! Your concern is exactly why I included median margin of victory as well as the average, and that’s why I used the median value when analyzing the results. The median is less sensitive to the outliers you described. I would love to look at playoff margins given a larger sample size though!
A lot of win now trades are made by teams who expect to make the playoffs but want to increase their odds once they get there. Fine, making a trade will increase my win probability by just 7.5% against a middle of the line team in the regular season, but against a playoff team you’d imagine the margins would tighten and it would be a more impactful move.
Even still, I see this article as a decent argument for the win now trade. Improving your win percentage anywhere from 7 to 25% seems desirable to me at the right cost. Good reminder to only target that sort of trade when you are already very good though.
Thank you for reading! I agree that a 7%-25% increase might be worth it at the start of the season, when you have plenty of opportunities for the move to matter. I’d be hesitant in paying future value with, say, only 4 weeks left in the season to increase my win probability by that amount, given the lack of opportunity to turn a loss into a win.
I too would like to see what the outcome would be if the analysis was applied only to those teams that made the playoffs each year. I have long been a proponent of the win now trade, albeit without analytical data to support, but only to make my team as good or better than the other elite teams.
I agree with what you are saying wholeheartedly. A few years ago I dominated a league to the tune of more than 300 total points scored over the 2nd highest scoring team in the league for the season. My end result was a 7-6 record and missing the playoffs. Due to pure luck of the draw on week to week match-ups I ended up facing the highest scoring team of the week on 5 weeks of the season.
Obviously, this was a very isolated result. But it really opened my eyes to how random fantasy football can be. And because of my experience won’t give up a ton of value on a trade just because I am in win-now mode anymore. I still like trading for star players, and I especially like owning players that I like watching play, but I’m not doing a bad deal just to gain a couple PPG. History tells me that I can chase points all over the place and still end up on the outside looking in, so I’d rather make sure I set myself up for the long run than chase right now. 🙂
“Lets say you 70% chance to win a game, and your win-now move make it 80%.
And you need to win 3 games to win the ship.
Without trading up your chances are 0.7^3= 34%
With the trade up your chance are 0.8^3 = 51%
That’s 67% more chances to win the league, IT IS a big deal.”
This is a strong point, but the last line is misleading. Better to think of it as increasing your odds by 17%.
To see why, imagine I have a team with 0.01% chance of winning the title. I make a win-now trade to increase my chances to 1%. That’s “9900% more chances to win the league.”
First of all, it’s a good point to think on, i have always wondered about it.
In 2014, i had a great team, in a standard scoring league with 6 Points per passing TD.
I bought CJA when he was starting his crazy stretch as a win-now move ( that turned out to be a nice buy but suppose not).
CJA was great all playoffs, but Aaron Rodgers my super-stud had a terrible game in Buffalo and i lose.
This is a great illustration of your point.
You’re right, trading for win-now is in most case hardly helpful.
Doesn’t mean that’s a bad idea ?
Let’s say you have a second round pick, you say that using him to buy an old veteran RB for playoffs will only give you something like 10% more chance to win a game, which sounds like a waste.
So let’s say you don’t do it, and use it to draft a young rookie or a some young player. What are the odds this trade helps you to win a championship ? It’s very hard to compute, but i’ wagger that not very much higher than buying a veteran when you have a true chance.
It’s not about how useless this 2nd round pick is, one single piece will never have that a great impact on your championship in average, and the young guy you’ll buy will proabably improve your PPG less than the win now guy.
So the true question is “To win a championship, assuming you have an etablished team, is it a better idea to use it on a win-now when you have a good chance or on a long-term prospect ?” .
Just throwing percentages without showing the alternative doesn’t help.
Moreover, you sounds like +10% isn’t a big deal. It is.
Lets say you 70% chance to win a game, and your win-now move make it 80%.
And you need to win 3 games to win the ship.
Without trading up your chances are 0.7^3= 34%
With the trade up your chance are 0.8^3 = 51%
That’s 67% more chances to win the league, IT IS a big deal.
Thanks for reading! If you feel that way about a 2nd round pick (and you may be right) then it’s not really a “Win now” trade as I defined it. In your example, you’re looking at acquiring a player that will give you more career value than you would expect from the pick. There’s nothing “Win now” about that; you’re simply increase the value of your roster. I would peg a “Win now” trade to be something like taking on a player that you expect to have career value X, while giving away a pick you expect to give a career value > X.
“Lets say you 70% chance to win a game, and your win-now move make it 80%.
And you need to win 3 games to win the ship.
Without trading up your chances are 0.7^3= 34%
With the trade up your chance are 0.8^3 = 51%
That’s 67% more chances to win the league, IT IS a big deal.”
This is a strong point, but the last line is misleading. Better to think of it as increasing your odds by 17 percentage points.
To see why, imagine I have a team with 0.01% chance of winning the title. I make a win-now trade to increase my chances to 1%. I now have 9900% better odds of winning, but the EV of the move was only 0.0099*title.
