how much speed could you get putting in the same amount of power from an upgraded wheelset (from stock)

It depends on half a zillion factors, most of them immeasurable. For the most part, you won't get faster as measured on the clock, but you'll enjoy your ride more.

What's your stock wheelset? I think it's mostly bearing related until you get up above 15mph or so. When I'm riding on some good hubs, I gotta feather the brakes to stay behind lotsa people I encounter on my commute when they start to coast, and I'm not drafting them.

I have noted, with my Cateye Strada Wireless computer, that my speed has, indeed, increased since I upgraded. My Trek 7.5 FX came with a 24-spoke X2 rear wheel. And a 20-spoke radial front. I hated them! So...

I built up 3X 32-spoke wheels using Mavic A719 rims, DT DB spokes, and Shimano Ultegra hubs. With the same energy expenditure as before, the increase was right around a highly visible 25%. I stand duly impressed.

I went from a 34MM deep rim to a 46MM deep rim and lost 185 grams of mass. My butt meter sync'd with my bike computer seems to show about .5-1.0 MPH increase in average speed on the same 2-3 hour rides I have done for years. I have noticed that with the same amount of effort I can keep the bike 1-2 MPH faster on the flats than I could in the past at the same fitness levels. It also could just be a big fantasy I have created to justy spending money on a new wheels. Either way I am happy.

Unless the bearings on your old wheels are horribly misadjusted and tight, the speed difference from new hubs will be infitessimal. All of the endorsements above can be explained by improved rim and spoke aerodynaimics (except Panthers007's who is obviously "pulling your chain"), not by improved hub bearings.

Above 15 mph or so, maybe 12, yes.

But, I can coast from 10 mph on the flats for a lot longer on 36 hole Campy Record with markedly non-aero box section rims than I can on REALLY cheap 32 hole hubs with Mavic cxp22 rims. At higher speeds, I'd guess the lower spoke counts and aero profile rims would make the second wheelset the faster choice.

The coasting test is easier to perform for me, so it's the only one I've done. I'm willing to accept higher speed aero-oriented test results that others have recorded.

Maybe someday we'll find out what wheels thecyclist currently rolls and what kind of riding he or she does...

I'm not sure how you guys are answering his question. He might as well asked "My bike is broken, help me fix it pls". A question like the OP's doesn't deserve an answer.

Complete and utter rubbish.

I think he was kidding, or at least I hope so.

It is more likely that the width of the rim changed, made the tire flatter and ended up skewing your cyclocomputer.

I'm pulling no chains, Hill. The Trek-supplied Bontrager wheels were MISERABLE! On a certain stretch of road, where I measure cadence in a certain gear, I would be going 11.5mph on my Cateye. With the new wheels and Ultegra hubs - meticulously adjusted/greased - I was going a hair over 14mph. Go figure. I was blown away!

Anyone want some Bontrager wheels?

<EDIT> I should add that I also put on some much better rubber - Rivendell Ruffy Tuffy tires at 90psi.

Empirical unbiased testing has shown that the most you can hope to gain is about 0.3 mph with very deep section aero wheels. Several years ago Bicycling magazine conducted downhill coasting tests comparing several aerodynamic wheels with a conventional wheel an the results were in the 0.1 - 0.4 mph gain.
Uphill tests would, of course, favor lighter weight wheels but it would be hard to eliminate the differences in riders.

Al

Rofl, that's pretty comical. A downhill roll is empirical and unbiased? Give me a break. There are REAL legit studies done in a wind tunnel on the # of watts saved/speed on most of the popular oem wheelsets out there.

Bicycling magazine, don't make me laugh. Want to see a real test? https://www.rouesartisanales.com/article-15505311.html

Yeah, I remember that article in Bicycling many years ago. I'm not a physicist or mechanical engineer, but I knew right away how poorly conceived it was. We want to know effect on acceleration and hill climbing. Rolling resistance is just a small factor.

The graph above, however, is still biased, because the Y access doesn't start at zero. If it did, the slope would be less.

Testing a wheel without a bike and rider attached is even more comical. Bike aerodynamics are dominated by the rider's position. You can spend a whole lot of money on wheels and it just won't make much difference.

And you're wrong.

Who cares what the rider is doing. We're interested in how many watts a wheelset eats up. Granted, it might be more useful with a frame attached to it, but are you goign to start testing every permutation of that wheelset on every production frame?

I don't think it matters.

No it wouldn't.

I can re-do that graph with a 0 there and the slope would be the exact same. Think about it. This graph isn't about rolling resistance, it's about aerodynamics.

Yes, but when you want to know how much the wheels make a difference, you want to eliminate whatever difference the bike and rider make. I think a wheel testing machine would make sense if they simulate real world conditions such as wind.

But really, I doubt the aerodynamics of a rim make much difference. The only relevant thing in a rim is its weight. Same for spokes. These things will matter if you're a champion racer. That's about it.

Oops, you're right about slope, operator, but it would still look different at a glance.

That really is good information but all of the wheels tested are compared at a constant 50 km/hr. How many people here will be riding at a constant 50 km/hr? And how many can translate wattage saved to speed gained as compared to a conventional wheel at realistic speeds? The "Bicycling Magazine" tests were conducted several years back when that magazine was more technical than it is now. Their downhill tests were conducted in a controlled environment with several professional riders riding all of the wheels through rotation so that all wheels were ridden the same number of times by all riders. I no longer have a copy of the article but remember that the speeds varied from zero at the top of the hill to, I think, about 35 mph. In other words, real world conditions. The speed gains were all less than 0.5 mph.

And i'm telling you, the test results from a "controlled" downhill test is as good as a realistic guess. https://yarchive.net/bike/rolling_resistance.html , do a search for the first occurence of the word "downhill".

Speed is not a good metric to compare performance, wattage is. Thus the existence of power meters. Which is the point i'm trying to make.

There are many factors at play - and they all have to be placed into the equation. From the temperature of the road surface, to the number of pebbles run over, to the 5lbs of Lobster Thermidor Rider Joe consumed.

Regards Bicycling Magazine, it's not fit to line a birdcage with.

That's an interesting discussion about rolling resistance and tires but not relevant to this thread. What I've tried to do is answer the OP's question which is about "speed" and an "upgraded wheelset".