This is generally how I put it when I address the "to convert or not to convert" question.

Mathematically, it doesn't NOT make any difference in the decision in regards to the after-tax result between converting (and pay tax NOW) and not converting (pay tax later) IF:

1.  There is NO change in tax rate; 

2.  Same rate of return for the Roth IRA and IRA accounts; AND

3.  The inflation rate (let's just say it's used as the discount rate for the relevant present value computation) is identical to the rate of return (the growth rate of the investment).

Regarding the impact of each variable:

*  Tax rate hikes make Roth conversion favorable.

*  High inflation (inflation rate exceeds the growth rate) makes Roth conversion unfavorable - you pay tax with "cheaper" dollar if you do NOT convert and pay tax later.

The examples about CA and NY folks doing the conversion and moving to NV or FL, categorically, fall in the consideration under "Item 1", i.e. Tax Rate Change". 

Likewise, a move from a low tax state to CA or NY would practically result in a "tax hike".

That said, I would say @BobKamman has missed a very important consideration in the real world - the issue of rational versus behavioral decision making.

@BobKamman said:

Take any amount.

Choose any tax rate, and assume it's the same at the time of conversion and at the time of distribution. 

Choose any rate of return, and assume the traditional IRA is invested the same way as the Roth IRA.

Choose any number of years. 

Show just one example of where the Roth account after taxes, is greater than the traditional IRA after taxes.  

You can slice and dice the comparison, mathematically and rationally, and conclude there is NO difference.  I'd give @BobKamman  that.        

But I'll also give @BobKamman not ONE but two examples.

Let's take an IRA of $100K at a modest 30% tax rate thru the period.

You would be comparing the growth of $100K in the IRA (at the same rate) and $70K (after tax) in the Roth IRA, and yes, over the years, after paying 30% tax from the IRA withdrawals, it would all be mathematically equivalent to the (nontaxable) withdrawals from the Roth IRA.  Mathematically speaking.  I have no problem there.

However, in the real world, the $100K IRA would be converted to a $100K Roth IRA, and the converter would pay the $30K tax as part of the tax filing for the conversion year.  In the mathematical comparison, it would then be necessary to compare (1) the after tax withdrawals from the IRA PLUS the growth of the $30K to (2) the nontaxable withdrawals from the Roth IRA account that started with $100K tax free growth.  Mathematically, they should be the same also, but this is where rational versus behavioral economics comes in play.  In the real world, the converter would say "ouch!" and pay the $30K in tax as part of the tax filing in the conversion year, while the NON-converter would spend the $30K in upgrading the kitchen, a diamond ring, a wedding for the daughter or in my case a few cases of Kavalan Px Sherry Single Cask 6, for my own consumption or for my friends.

At the end of the comparison period, the converter would have the SAME amount in the ROTH IRA account - tax free - as the non-converter's IRA account because in Year one both accounts started with $100K.

Of course the rational decision maker would (justifiably) argue the value of the utility of the kitchen remodeling, the daughter's wedding, half of the corvette, or (in my case!!) the few cases of Kavalan Px Sherry Single....

Example two, by the same token, and an example that parallels the "To convert or not to convert, that is the" Question.  Twin Brothers B1 and B2 have argued all day about which way to go - universal life insurance or Term Life Insurance. 

B1 says, Universal Life is better because there is this investment element.  B2 says, Term Life is better because if you invest the cost difference prudently and beat the rates provided by the blood-sucking insurance companies in the Universal Life (easy to do because you can find the same hedging investment programs and SAVE the charges charged by the blood-sucking insurance companies), you'd be so much for ahead.

@BobKamman wolud likely agree with B2.  I regrettably did 30 years ago.

But in reality, after 30 years when B1 and B2 met again and made the comparison (both bald by then), B1, the dumb brother, has a pot of investments in his Universal Life Policy.  B2 had a gut from the bourbon and beers that he had paid over the years with the cost difference.

Rational vs behavioral...   Dumb Brother always beats the Smart Brother....