A basic problem with "the rule called 'regression to the mean'" is that it assumes what it purports to show, viz. that the mean doesn't change.
For example, since 1928, the S&P 500 has averaged around 400. Anyone expecting it to "regress to the mean"? I didn't think so. Why not? It's because the mean isn't a constant - the S&P has an upward bias.
On what basis are we to assume that the market rate of return doesn't also change over time?
The business insider article showed that 20 year time frames are too short a time to compute an average for "market" returns - even assuming that there is some constant long term average. The graph there showed that the average return over 20 year periods ranged from 6% to 18%. Hardly a constant average.
So now we have two questions. The original: Is there a single long term average return for the market? We now add: how long a time frame do we need to use to get even a decent approximation of that "constant" average.
Rather than explain things, "regression to the mean" shows how people take something on faith (consistency of long term returns) to circularly prove that over the long term, a certain average rate of return can be expected.
People like to assume constancy. That's the point of the seeking alpha piece. Market doing well? That will continue; not. Market return averaged X% in the past? That will continue. Maybe, maybe not.
Just for the record, "mean regression" is a short term prediction - it says that when the last value of a random variable is sufficiently far from its mean, the next value it assumes will more likely than not be closer to the mean.
http://mathworld.wolfram.com/ReversiontotheMean.html
If a random variable truly has a constant mean, then over time you'll approach that mean, not because of mean regression, but because of the law of large numbers.
http://mathworld.wolfram.com/LawofLargeNumbers.html

Hi Guys,
Change happens; it is a certainty.
"Richard Russell, the famous Dow Theorist, once noted that over a shorter time frame almost anything can happen in the financial markets, but over much longer, meaningful periods of time (referring to years), the surest rule in the stock market is the rule called "regression to the mean.""
This quote was extracted from this regression advocating article:
https://seekingalpha.com/article/2315705-regression-to-the-mean-and-why-investors-should-not-ignore-its-importance
I am in complete agreement with the main theme of this article. There is a compelling and irresistible market pull towards a regression-to-the-mean. In any single year, almost any extreme is possible; anything can and does happen even if probabilities are low. But as the timeframe expands, the marketplace adjusts to deliver more predictable average returns. Over time, sanity rules.
I have zero confidence that I can predict tomorrow's market returns, but I am very comfortable about staying in the market for the loooong run.
Best Wishes

A factor that affects returns about as much as expense ratio is turnover. Here's an old but readable column citing research from a decade ago supporting the thesis that trading costs amount to at least as much as costs included in the ER:
https://www.advisorperspectives.com/pdfs/newsltr27-2.pdf
I tend to lean toward managed funds, but at the same time I also keep an eye on expenses, both documented (ER) and implicit (trading costs via turnover). FWIW, with the exception of a few placeholder positions I have (in case I ever want the funds), and one major bond fund holding, all of my funds have less than 50% annual turnover. Of those, all but one fund has under 40% turnover.
M*'s portfolio analysis claims that my weighted average ER is about a half percent below a "similarly weighted hypothetical portfolio". So while my expenses are not rock bottom, I'm satisfied with my average expense ratio too.
One minor gripe concerning the cited AAII article - it uses RMDs as withdrawal amounts. The main problem with this is that people's need for cash in retirement is fairly stable, as opposed to RMDs. RMD calculations require you to take 1 / (remaining lifetime), which grows exponentially as you get older and your expected lifetime shrinks toward zero.

I agree that mixing in some bonds other than Treasuries would be more representative of real investor results. Beyond that, it's not clear what that mix should be. Bogle, of all people, thinks that even the total bond market index (US Aggregate) is still weighted too heavily toward US government bonds.
(See the section "The bond index portfolio is not representative of the vast majority of fixed income investors" in this 6 page pdf.)
Aside from junk, I would expect the general shape of the long term curve for US bonds (in whatever mix) to be similar to the image above, though the return should be somewhat higher (corresponding to the higher risk of some of the bonds in the mix).

FYI: Brexit, threats from North Korea, political upheaval in the U.S., and other recent geopolitical events have been putting investors on edge and creating spikes in market volatility. Add elevated market valuations to the mix, and it might seem like a recipe for stock market mayhem
Regards,
Ted
https://www.fa-mag.com/news/shifting-to-international-equities-34952.html?print
M* Snapshot BBALX:
http://performance.morningstar.com/fund/performance-return.action?t=BBALX&region=usa&culture=en_US
Lipper Snapshot BBALX:
http://www.marketwatch.com/investing/fund/bbalx
BBALX Ranks #1 In The (WA) Fund Category By U.S. News & World Report:
https://money.usnews.com/funds/mutual-funds/world-allocation/northern-global-tactical-asset-allc-fund/bbalx

Thanks for linking this article @Ted,
Well worth the read:

The second, specifically in the fintech world, is the blockchain. Many technologists tell me the block chain is the most important innovations in human history. They equate it in terms of importance to fire, the wheel and the internet.
The typical financial advisor has either never heard of the blockchain or is unable to describe it, meaning that advisors today are as clueless as their clients. And that’s not a healthy situation for the client. The blockchain is the underlying technology that was used to develop bitcoin, the world’s first digital currency.

Since XT's inception it has trailed QQQ:

YTD DSENX outperformed S&P 500 by 2%. It is not bad at all and I would be satisfied if it continues that way.

1. The objective is to get as much into the Roth as possible while paying taxes on $10K.
Using Bee's example, the current combined value of all five accounts is $61,830. If you recharacterize all but the one that did best (PRIDX), you wind up with $13,106 in the Roth and $48.724 recharacterized back to the traditional IRA. If you recharacterize all but, say, PRWCX, you wind up with "just" $11,206 in the Roth and $50,624 recharacterized back to the trad IRA.
Either way you pay taxes on $10K, but leaving PRIDX as the converted fund results in the largest percentage of your portfolio converted into the Roth for the same amount of tax.
That's not to say that you're stuck with these investments. You could choose to exchange PRIDX inside the Roth for PRWCX and/or exchange PRWCX inside the traditional to PRIDX, or anything else. Perhaps I should have said you pick the "subaccount" with the highest value to keep as a Roth, and recharacterize the others.
2. Bee mentioned that an objective is to keep income low enough to qualify for ACA subsidies. Getting $1500 out of a taxable account to pay for the Roth conversion could result in recognizing capital gains, and thus increase income. That extra income might disqualify Bee from any ACA subsidy. On the other hand, getting the $1500 out of a Roth account would not increase income, so the ACA subsidy would remain safe.
3. Say you have two funds each worth $1500, the amount of the tax. You expect the first to double in the next four years, the other one to grow by 50%. All else being equal, you'd rather keep the first investment and sell off the second, since you'd have more money ($3000 vs. $2250) at the end of four years. Of course there are other considerations, notably asset allocation and risk. In glossing over those considerations, I may have underestimated their importance.

DSENX has performed well during its nearly 4 years of existence.
I use the S&P 500 as a benchmark for this fund. Not an ideal benchmark, but for me, close enough. I've noticed that, over the past year or so, DSENX is barely outperforming the S&P 500.
I'm not going to make a sell decision based on the last year's performance, but I'm wondering how much longer DSENX's outperformance can last. Any ideas?

If you want to convert $10K, then you just pick the fund that did the best. (Technically you have until October 15, 2018 to do the recharacterizations, so you could watch and wait a long time to see which one does the best.)
You'll owe $1500 in taxes. As you said, you should have a good sense of how to pay for this. You may be suggesting that this come out of a Roth. Generally (unless one is trying to keep one's income down), that's not a good idea. Once money is in a Roth all growth is tax-free. It doesn't get any better than that. Money from anywhere else is less valuable.
Even if you are going to pull the tax money from a Roth, you likely want to pull it from the fund(s) that you expect to do worse going forward.

msf,
Three questions:
1. Why pick the fund that did the best to convert the $10K?
2. I understand why it is generally not a good idea to pay the tax from the Roth. But what do you mean by "unless one is trying to keep one's income down"?
3. Why pull the tax from the fund(s) that you expect to do worse going forward?
Mona

December 19, 2007: "Birinyi thinks the bull market that started in 2002 is still very much intact."
https://google.com/search?tbs=cdr%3A1%2Ccd_min%3A12%2F31%2F2006%2Ccd_max%3A12%2F31%2F2008&q=Birinyi+thinks+the+bull+market+that+started+in+2002+is+still+very+much+intact.+&oq=Birinyi+thinks+the+bull+market+that+started+in+2002+is+still+very+much+intact.+&gs_l=psy-ab.3...3654.16044.0.16913.2.2.0.0.0.0.149.285.0j2.2.0.dummy_maps_web_fallback...0...1.1.64.psy-ab..0.0.0....0.LdndDfWFUvQ

@msf,
Thanks for your comments. I see your point.
I have also read that the best time to convert to a Roth would have been at a market low. "Convert to Roth low when potentially more under priced shares can be sold "cheaply". Ultimately it is the after tax return investors should be striving for. Since all Roth returns after inherently "after tax" returns, I strive to maximizing as much of the 15% tax bracket by considering adding to my taxable income up to the 15% tax bracket maximum.
I should also say that I am trying to stay eligible for ACA Insurance subsidies which can be at odds with Roth conversions and maximizing tax brackets.
Much like "after tax returns on investment", I believe even income (whether earned or unearned) can be looked at through the lense of "after tax income".

Unfortunately common ETF myopia. Describes a problem (how to get diversification with a bond ladder), and then says: "Enter ... ETFs". Seemingly ignorant of the fact that defined maturity bond funds (DMFs) have existed since 1985.
Vanguard has a much more analytic piece on DMFs here:
https://personal.vanguard.com/pdf/ICRDMB.pdf
Call it confirmation bias, but I think it's spot on. They observe that the protection against rising rates that individual bonds (or DMFs) supposedly offer is essentially illusory. That's because as rates rise, bond prices drop. You may choose to think "well, I'm not selling, so I haven't lost value", but you've paid an opportunity cost (not getting the higher yield of new bonds) equal to the price drop. Vanguard says that the "return of principal ... benefit is more emotional in nature."
They note that even laddering bonds (or DMFs) has increased volatility relative to a "regular" fund (or ETF) because of the saw-tooth nature of laddered bond duration. It falls until a bond matures, then spikes up with the replacement bond, and then repeats this pattern.
Where DMFs excel IMHO is in immunizing a bond portfolio. If you've got a particular need some time down the road (e.g. college expenses), you can invest in a DMF that matures just when you need that money. ISTM that here zero coupons work best, because then there's no reinvestment risk (the possibility that you won't get the same rate of return on reinvested interest). Only American Century offers zero DMFs, and it isn't issuing any new ones.
Vanguard also identifies what it considers a drawback of ETF DMFs vs. OE DMFs: "Investors purchase DMF exchange-traded funds (ETFs) at market price, but receive net asset value (NAV) at maturity. The difference in these amounts represents a sometimes unrecognized cost that investors may incur."
Aside from that, the higher costs of OEFs, and the usual ETF overhead of brokerage fees, market spread, etc., OEF and ETF DMFs seem to have pretty much the same benefits and drawbacks.

You seem to be combining two things: (1) selecting which fund to convert (i.e. keep as a Roth), and (2) how to pay the tax on the conversion. These can be taken separately.
If you want to convert $10K, then you just pick the fund that did the best. (Technically you have until October 15, 2018 to do the recharacterizations, so you could watch and wait a long time to see which one does the best.)
You'll owe $1500 in taxes. As you said, you should have a good sense of how to pay for this. You may be suggesting that this come out of a Roth. Generally (unless one is trying to keep one's income down), that's not a good idea. Once money is in a Roth all growth is tax-free. It doesn't get any better than that. Money from anywhere else is less valuable.
Even if you are going to pull the tax money from a Roth, you likely want to pull it from the fund(s) that you expect to do worse going forward. Those may not be the same funds as just did well this year. We all know that "past performance does not guarantee future results", YMMV, etc.

One of the best Roth conversion strategies I've come across requires making many smaller "Roth conversion sub accounts' in January and then selecting the best performing (between January - April of the following year) "sub-account" to convert. "Sub-accounts" not converted are then re-characterized and returned to deferred IRA status.
Here we are in October, when, historically speaking, the market performs its best over the next six months (Oct-Apr). Might this not be a good time to consider this same strategy?
Here's one scenario. Lets say back in January I divided my IRA ($50,000 account) into (5) $10,0000 sub-accounts (that were designated for Roth conversion). Each invested differently. I will use T. Rowe Price funds for this example since I am familiar with their performance. Each sub-account holds just one fund, so (5) funds each valued at $10,000 as of January 2017.
January 2017 - September 2017 has been very good YTD for these funds. So much so that a these $10,000 conversion would be valued at the following for a few funds I track ( I own each as well):
PRWCX = $11,206, PRHSX = $12,286, PRIDX = $13,106, PRMTX = $12,666, and PRNHX = $12,566

On a $10,000 conversion, at a 15% tax, $1500 would be due in April 2018. As you can see, 4 out of the 5 funds could fully fund that tax bill with the tax free growth since January 2017 when each conversion was set up.
In fact, had you not set up these conversion accounts in January you could do it now and pay the additional taxes on the October 2017 value.
The 15% taxes for the 4 funds that made the cut (PRWCX didn't since it has grown less than the tax liability) would be $1500 on the original $10,000 plus the taxes on the gains since January.
So if you converted in October you would owe (taxes in parenthesis) for each of these funds:
PRHSX = $12,286 ($1843)
PRIDX = $13,106 ($1966)
PRMTX = $12,666 ($1900)
PRNHX = $12,566 ($1885)
The gains so far this year (January - October 2017) would cover the tax liability which... is a very good thing. One could merely go to cash or if history is in you favor continue to stay invested until April 2018 (tax filing for the conversion).
As important as having a Roth conversion strategy one should also have a good sense as to the tax liability on these conversions as well as how to pay for these conversions (taxes).

FYI: Ric Edelman, founder of Edelman Financial Services, and one of the driving forces behind the creation of the $1.35 billion iShares Exponential Technologies ETF (XT), has had an interest in technology as well as its implications for advisors and investors. His most recent book, The Truth About Your Future, is all about how rapid technological advances will radically change how we live, invest and retire. ETF.com chatted with him about how fintech is helping bring to fruition the future he anticipates.
Regards,
Ted
http://www.etf.com/sections/features-and-news/fintech-will-change-advisory-game?nopaging=1
MFO's Link To amazon.Com: Ric Edelman: "The Truth About your Future"
https://www.amazon.com/Truth-About-Your-Future-Money/dp/1501163809/ref=sr_1_1?ie=UTF8&qid=1506935763&sr=8-1&keywords=the+truth+about+your+future
IShares Fact Sheet XT:
https://www.ishares.com/us/products/272532/XT?cid=ppc:ishares_us:google:tickers&gclid=Cj0KCQjwx8fOBRD7ARIsAPVq-NkxaNRvf4UP9fJvwQ8ya_n8PHh-IZj-j0uKaem2DNGg8tALlkAiKKoaAkfPEALw_wcB&gclsrc=aw.ds
M* Snapshot XT:
http://www.morningstar.com/etfs/XNAS/XT/quote.html

While the intent of that business insider chart is clear, the data are less so.
What stock market (US or global, S&P 500 or Wilshire 5000 or ...), and what bond market (ten year T-bonds, corporates, or ...)? If one uses returns of the S&P 500 and 10 year treasuries, the numbers don't match. They're not far off, but they show that at least this reader doesn't know quite what data were used.
What sort of rebalancing if any was done over the rolling 1, 5, 10, and 20 year periods? Nothing is said about this either.
Here's the data I used. It goes all the way back to 1928.
http://pages.stern.nyu.edu/~adamodar/New_Home_Page/datafile/histretSP.html
Those older figures are important, because while returns since 1950 (a few years after WWII ended) have been "okay", earlier periods showed that the market can be even worse. Not that losing 2% each year for five years (cumulative loss of 9.6%) is that great.
It used to be we were told that "the stock market" (whatever that meant) never had a loss (ignoring inflation) over any ten year period since WWII (maybe that's what you were thinking of). But then 1999 - 2008 came along. And right after that, another rolling ten year nominal loss, 2000 - 2009.
There's a problem with putting too much faith in historical data. Things change. One can either ignore that, or adjust expectations accordingly.
We've had a 35-40 year bull market in bonds (with 10 year treasury yields now bouncing around 2%), that followed a 35-40 year bear market in bonds (that started in 1941 with yields around 2%). That's a looong cycle that should be incorporated into projections. There's no way that bonds can boost a portfolio's returns - they yield virtually nothing, and if the yield goes up, the total return could turn negative.
From the end of 1940 to the end of 1956 there were only two years where the total return of 10 year treasuries broke 4% (none 5% or more). Which gets us back to the question of whether these figures include rebalancing, and more generally, why even bother with bonds now? Cash isn't yielding that much less, and if interest rates rise, cash should track bonds. (In the 1940-1956 time frame, 3 mo treasuries returned 1.6% annualized, not much worse than bonds at 2.6%, and with virtually no volatility.)
The only cherry picking I'm doing here is with the end date of 1956. Bond yields bottomed out around 1941 at 2%, just as they seem to have done now (more or less). The graph below is ten year treasury yields over time.

Thanks for starting the thread @slick...it helps me think.
Presently my portfolio consists of 62% equities and 36% (Bonds and other...Preferred, Convertibles) and 2% cash.
My allocation funds make up 21.2% of my portfolio (VWINX, PRWCX, BRUFX) and they give me a benchmark to compare the overall portfolio against:
VWINX is 38/59/3 (Equities/Bonds (other)/Cash)...YTD = 6.0%
PRWCX is 62/24/14...YTD = 9.93%
BRUFX is 57/33/10...YTD = 9.06%
"BeeX" (my combined portfolio) is 62/36/2...YTD = 11.82%
Drivers of YTD performance have included:
Funds...YTD
PRIDX...27.16%
PRMTX...28.85
POAGX...19.51%
VWO...24.69%
Collectively these funds make up 50% of my equity holdings and provided 7.22% of the overall portfolio performance.

@slick: Fantastic post!! I've always appreciated @Old_Skeet's discussion of his portfolio and sleeve system and this is right up there.
I have 22 funds and 16 stocks plus the cash in funds and that I hold in my IRA. I won't comment on the stocks except to say I should stick to funds and I'm glad I'm headed in that direction. Like you and others, I check performance but I tend to focus on slightly different things. I view the returns as a reflection of my asset allocation decisions and prefer category rankings when reviewing funds because the results can differ. For instance, 12 of my 22 funds have greater than 20% returns with 3 of those over 30%. Another 7 had returns greater than 10% and only 3 were between 0-10%. However, only 7 of my 22 funds are top decile in their category, 2 more are top quartile, 7 more are top half and 6 are doing pretty lousy with 5 of the 6 in the bottom decile of their category.
My 3 largest positions are 7-10% positions in my portfolio and their rankings are GPIIX at 40, POAGX at 24 and GPEIX at 90 (ugh!). If I screen out large cap emerging markets funds GP isn't doing much better but I'm a believer and the longer term record is still good so I'm not even close to thinking about giving up.
I'm overweight healthcare at 16.6% of my portfolio and that's been working really well especially with my 3 healthcare funds (HQL, SBIO and PRHSX) ranked top 1%, 8% and 32%, respectively. Half of my healthcare exposure is within other funds or a couple of the stocks but being overweight has helped a lot this year.
I'm also overweight emerging markets at 15.7% of my portfolio but while that should make me a big winner this year, all 3 specific EM funds I own (GPEIX, SFGIX, MEASX) are having tough years. Like with healthcare I'm getting exposure from other funds as well, some of which are having great years, but I feel like EM has been a disappointment.
I've had no bonds for years, often to my hindsight's regret, but I count cash in my IRA as an investment decision. Together with the cash in funds I own it's 15% of my portfolio and has been since the beginning of the year in rough terms. Overall, my returns on the 85% invested are very close to the S&P but considering my overweights to healthcare and EM should both be helping as well as getting a currency benefit from a 50/50 split between domestic and international investments, I would have hoped to be having a better year. Which brings me back to stocks...

Hi Guys,
When investing in equities and/or bonds, often an optimistic outlook is needed to help an investor to stay the course. But that optimism should be grounded in realistic expectations. The target return numbers quoted from the Boston-based group of investors fail that simple test by a rather wide margin.
Here is a Link to a very informative chart that summarizes historical equity and bond returns since the end of WW II:
http://www.businessinsider.com/range-of-annualized-stock-bond-returns-2014-11
These data provide significant benchmarks for planning purposes. While history is never exactly repeated, it does provide a terrific point of departure for making projections.
Unless circumstances demonstrate an excessive change in economic/political conditions, returns from any 5-year or longer timeframe should not be dramatically different from the historical database. These data should be used to guide expectations.
Daydreaming might provide some comfort for some folks, but the marketplace does not deliver for these daydreamers unless they are lucky. In the investing world both luck and skill determine the outcomes. Daydreaming is not a factor. It can do substantial harm to both the portfolio and to an individual's self esteem. You gotta know the data.
Best Wishes