my felt roadie is under 21# with my aero bars, blowout bag and lights...

i'm not saying anything,
but you know what i'm saying...

I'd say to buy a new bike and train on your old one!

Depends on where you ride and race. If you live and race in Colorado, then yes, buying a new bike with a primary focus on weight will make a big difference. If, on the other hand, you're racing and training in Florida, then weight is honestly not a big deal at all. As long as the other mechanical components are working and aerodynamics suit you, then there's no need to upgrade OR buy a new bike. I see from your profile that you're in NC. That's sort of a toss-up. I just moved to Virginia and it's pancake flat here, so a lightweight bike is of no benefit (I shouldn't say "no" benefit -- it's still nice for overpasses and quick acceleration) but when I lived in California I was LOVING my sub-20-lb carbon bike on the 10-15% hills.

If you have any particularly heavy components (i.e. stock wheels, old clips, steel seatpost) it may be worth upgrading, but don't bother upgrading components like the shifters/deraileurs - the cost of a lighter one does not justify the couple of grams that you'll actually save. If you do decide that weight is a primary concern, though, there are lots of mid-range carbon aero wheels or carbon aero framesets (no components - you'd transfer your current stuff over) on the market right now. Either of those would be a pretty good bang for your buck.

Wheels are the first upgrade as it's not only lighter but rotational weight (have at that comment techno weenies...I know you are out there) with better wheels will make you quicker. That being said, you paid 400 for it ten years ago and are still at the sport...bikes have changed so much in ten years. Seriously, why not get a new bike at this point?

This is why:
http://www.youtube.com/watch?v=dVwKE9yDqVo

:)

jono

Sorry, I don't mean to sound like a jerk, but that's a totally different phenomenon. Angular momentum would affect the handling of the bike, certainly, but not the acceleration or sustained speed. In terms of being able to accelerate from a stop or maintain steady-state speed, the wheel weight does make a difference, but that's actually just plain linear inertia.

I'm pretty sure I've even seen the whole "cut rotational weight" thing debunked. One outstanding resource for this type of thing is Analytic Cycling. They have a couple models (http://www.analyticcycling.com/WheelsClimb_Page.html or just go to their main page and click "Wheels & Aero & Weight" on the right) that lets you calculate scenarios and vary the parameters. Something I like to do is set the "standard" and "test" riders exactly equal and then vary one parameter at a time.

For example, setting them equal and changing the inertia value for the wheels from something extremely high (0.15) to zero results in truly negligible time savings (around 0.02 sec) in a scenario such as a triathlon where you can usually (ideally) hit a steady state speed and maintain it much more so than, say, a criterium where you're constantly accelerating for corners and breakaways and what-have-you. So basically, don't buy wheels for their lower rotational inertia. Further, changing the weight, regardless of whether it is on the bike or the wheels or the RIDER, results in the same time savings. So buying wheels for lighter weight is fine and good, but it's no DIFFERENT from saving the same weight elsewhere.

So, back to the original point -- saving weight is saving weight. If you save 10 lbs going from your 26# bike to the absolute lightest road bikes out there, and then decided to climb a 5km, 5% grade, you could save somewhere between 90 seconds and 2 minutes. Of course, if you raced on a pancake flat course for an hour with your fancy-pants new bike, you'd save a second or two max.

On the other hand, improvements in aerodynamics, fit, and general mechanical quality will be much more significant for the triathlon-minded rider. Fiddling with the numbers on the model, improving the rider's aerodynamics could result in 4-5 minutes over a 40km course.

Basically, it's really tempting to lift your bike with one finger and think it's going to make you super fast, but the reality is that weight is just one part of the equation and when you look at the actual numbers, it's not even that significant, especially for the triathlete.

All that being said, though, if you feel like you need a new bike and you can afford one, go get it. I can tell you day and night about how the physics of bike weight work, but when you go get on your bike if you feel like it's heavy and it's holding you back, then you won't be as happy. A new bike will improve the "total package" of weight, aero and mechanical, which is what you need to do if you're looking for serious time savings.

Finally, if you're really curious, start taking stuff apart and weight it separately. Check out WeightWeenies.starbike.com or CompetitiveCyclist.com for weight resources to compare to. Wheels are easy, so start there. Then maybe look at your pedals, saddle, seatpost, chain...

Consider something else, too. How much do you weigh? When thinking about reducing bike weight don't disregard the total 'vehicle' weight. It is easy to get seduced by the thought of dropping five or six pounds from a bike. For example, your current bike is 26 and you mentioned a 17 pound bike. Yes, that is a 9lb reduction, which is a considerable 34% reduction. But if you also add in how much the combined rider+bike weight is, the difference diminishes drastically (like that alliteration?).

For example

1. You weigh 185lbs + 26lb current bike = 211lbs
2. You weigh 185lbs + 17lb new bike = 202lbs
3. You still have the nine pounds of reduced weight, but it is a real vehicle weight reduction of 'only' 4.5% which isn't as glamorous as it originally seemed when you thought, "Wow, 9lbs or 35% lighter."

I'm not saying you shouldn't be concerned with weight. Just ask youself if 1) that 4.5% weight reduction is the most important criterion for a new bike, or 2) if you could lose total vehicle weight in other areas. Looks down at post-IM gut :(