September 3, 2026

Is streaming the QB position a good fantasy strategy?  Is it even viable?   

My keen interest is always to see what the data says. Let’s not go by anecdotes and opinions. We let the data talk.

To investigate whether QB streaming is a good idea or not, I designed simulations to try and settle the question.

All the analysis is going to boil down to 1 central question:

• How many QBs can each of your league-mates roster, before streaming QB becomes a worse idea than drafting / holding?

And we want the answer as it depends on league size.

Along the way, this study reveals 2 other key learnings:

1. What is the most accurate QB baseline to use, in fantasy points, for the purpose of Value Based Drafting calculations, if you consider streaming QBs as the “replacement player”?

2. What’s the “QB rank” you could expect to achieve, for your league size?  E.g. at the end of the season, can you say “Streaming got me as many points as ‘the QB3’ !” ?

TL;DR So what’s the Answer? 

There’s a graph at the bottom displaying instructions for how to deal with a draft situation. You should look at that graph for an approach to understanding “Is it a good idea to draft, or should I stream?” depending on how many QBs have been taken so far.

For a higher level answer…. After averaging everything out, the answer appears that streaming QBs should usually be a viable strategy for league sizes of 10 teams and under. At league size 12, it’s close, maybe a wash. (Assuming 22 “taken” QBs.) Overall, you can’t judge the strategy based on any single year: Some years were good, some years were bad for streaming.  And yes, it really depends on league size and how many QBs each of your opponents holds.   

Methodology

Note: the analysis of QBs is different from Kickers and D/STs, since QBs are typically drafted and held.  I.e. we have to address holding as your opponents’ primary strategy.

Concept: Imagine you’re scraping for the leftovers, each week. If your opponents already hold the top-scoring 20 QBs, what do you do? You look for the QB with the highest projection that week.

Procedure of simulation:

• Each run assumes a certain number of “taken QBs”— meaning how many QBs your opponents are holding.

• Your opponents start in week 1 owning all the top QBs, based on their ADPs that year. And they will hold them, for at least the first month— so you can’t get them.

• You add/drop QBs, to get the best streaming prospect each week. That means you pick among the remaining QBs— the leftovers that your opponents are not holding— and you select the one with the highest projection for each specific week. (Based on Subvertadown weekly projections.) The outcome fantasy scores get recorded.

• After the first month, some of your opponents wake up to realize their QB isn’t the best they can get. Their strategy now changes: they’ll grab the best scoring QBs (the ones with the highest fantasy point averages so far). This is to simulate that the “best” QBs are not available to you. (Unless you already have him on your roster. The simulation doesn’t let your opponents steal a good pick-up from you.)

• All the weeks of the season are carried out this way. Your opponents keep away the highest scorers. You stream. The results all get recorded, to eventually measure what REALLY worked.

• The simulation repeats for a different “number of held QBs” in the league. We want to represent tiny leagues up to larger ones where 25+ QBs might be held.

The existence of QB projections is the essential ingredient for simulating: you need weekly rankings.  I believe my own projections are more than a good-enough metric-- it’s quietly been overperforming when I do my accuracy assessments each year.  However, the main weakness is that I don’t capture weekly QB injuries/absences, when I simulate past years; I can’t easily fix that, but I think this analysis gets us to quite a reasonable answer anyway.

Further Methodology Detail: How I simulated Opponent QB Holding, by Allowing Them to Add/Drop: Making Streaming Harder

I’m going to explain the assumption in the “Methodology” section. Why don’t we just assume your opponents hold the same QB all season?

The dose of reality is that many of your opponents drafted a dud, and they can do better. During the season, they’re also going to add/drop, to get the best QB from the waiver wire. You're competing for them.

The real question is “When”. When do they decide to drop their drafted QB?

I optimized the simulation to make it as hard as possible to stream. Short-cut to the answer: It turns out that 6 weeks is some kind of optimum.  If we assume they drop their QB before 6 weeks, there’s a chance you can grab their stud who just had an unlucky start. But if we do the opposite— assume they hold longer than 6 weeks— then our simulation lets them hold duds for too long. Since my goal was to make streaming look as challenging as possible, I settled on the 6-weeks-to-swap result. This made Holding look as good as possible.

Further Methodology Detail: How I simulated Own QB Streaming, by Allowing “Short-term Foresight” while Streaming

Okay, this is the last detail to explain. (It’s just if you’re interested!)

It’s possible with streaming that you pick up a great QB who might be good to hold for a while.

It would be shame to drop that breakout QB, just because a single week’s forecast looks bad.

Therefore, I allowed a small 30% weighting factor to look ahead at later weeks. “Yeah, it’s a bad week to stream, but the week after that looks great, so let me hold him short-term”.

The First Result: Comparing Average Points Scored By Streaming QBs, Versus to Points Scored from Holding QB.

The question is: “If our league-mates held the same QBs all season, how many points could we earn from streaming the QBs remaining to us each week?” 

The black curve answers that, assuming the top-projected QB is chosen each week as our streamer.  

The x-axis here is “how many QBs are taken each week.” 

Somebody could mis-interpret simulations to claim that Streaming QB is always a fantastic idea.

It’s easy to be misled by seeing the BLACK line (streaming) above the GRAY line (holding).

But there’s a fallacy.

Your opponents almost never hold just 1 single QB. Probably they hold 2 QBs on average— or somewhere in the range of 1.5 - 3.0 QBs.

So if you have 10 opponents, it’s not accurate to assume that you get to stream at position “10”.

You likely need to consider the right side of the chart— like at position “20”.

The Quantitative Number (Fantasy Points vs. #Taken QBs) is Useful for Baselines in Value-Based Drafting

In addition to being interesting, the black curve above is actually useful for Value Based Drafting.

This means it can be used to help you evaluate the relative positional value (of QBs compared to other positions) to optimize your fantasy draft.

If you use my draft tool TapThatDraft , then it includes an option to “Stream QB” that evaluates QB value based on this result.

Second Conclusion: What equivalent QB Rank you can achieve?” (compared to the other QBs your league took)

You can also see— just by eyeballing the results in the above chart— that streaming the QB20+ can apparently be better than holding the QB10+. 

You can apply: If you think your league isn’t going to take more than 20 QBs from the draft, then you can pass on the "QB10" in your draft. Because your streaming results might be better.

It would be nice to represent it in an easy diagram, to consider all rank possibilities.

To begin, let’s consider the idea of EQUIVALENT RANK.

To get an intuition for this, as an example, imagine that 20 QBs are taken.  The question will be: If you streamed the “20th best QB each week”, over the last decade, how many Held QBs did you effectively beat that way?

This “equivalent rank” can be visualized.

Here’s a graph showing how well streaming “RANKED” over all seasons.

As an example, notice that streaming the 20th Best QB often produced “QB9 numbers”:

In other words, “What Rank QB do you achieve by Streaming, when ‘N’ are already taken?”.

A lower rank is better; A higher rank is worse. 

Getting to a Rule of Thumb: “Should I draft a QB earlier, or just wait and stream?”

This final analysis gets at the heart of how and when to DRAFT.

At some points during your draft, you might definitely pass on QB because you expect you can do better streaming QB.

But at other times, it might be a difficult decision, because streaming is NOT expected to surpass the points of drafting and holding.

So now let’s construct a graph to guide us.

We want to answer “if I draft the QB available right now, would it do better than streaming?”

Based on these 2 information inputs:

1. “QB rank # R is available to me RIGHT NOW”, and

2. “My opponents usually roster X #QBs-per-team”.

And here’s what it looks like— and note this is based on the same information as above, it has just been transformed (taking the ratio of equivalent rank to current rank). :

You can use this graph as a look-up table: Draw a line like I did (I chose 2.0), and if a certain QB falls to you, ask if they fall above that line.

You can also imagine a vertical line to indicate your worst-case scenario (like say, at QB 26)— as if you didn’t draft any QB at all.

Here are some observations:

• If your league mates typically hold more than 3 QBs per team, then it almost NEVER makes sense to depend on streaming. You should try to hold at least 1 QB, to improve your odds at higher scoring.

• As you can see, the ratio of Taken / Streamer-Rank is usually just above 2.0, when fewer than 21 QBs are taken. 

  • This is important because, if your league holds exactly 2 QBs per team, then statistically you don’t need to worry about drafting a late QB, as long as you know that fewer than 21 will be taken.

• Meanwhile, the exception to the above statement: You can also see that QBs ranked #1 - #4 might be worth drafting, if you find yourself with that opportunity.

As a closing statement, I want to point out that this analysis assumes YOU hold only 1 single QB at a time.  Obviously there’s a lot of potential from holding 2 QBs at a time, if that’s possible.

TL;DR

For me, the best summary is the final picture, on the right side. Streaming QBs can be a valid strategy, as long as your league-size isn't more than 10-12 teams (and 12 teams is ambiguous). I've shown how the VBD-baseline should be altered, depending on your league behavior. I've shown the equivalent rank of a streamer can often reach "taken QBs divided by 2".

Thanks for reading, and I'm happy to discuss, take feedback, or answer questions!

/Subvertadown