Page 24 of 58

Re: Stock scream room

Posted: Sat May 25, 2019 2:00 am
by D1984
ochotona wrote:
Fri May 24, 2019 12:14 pm
Sell in May Go Away reference:

https://stocktradersalmanac.com/Strategy.aspx

Seems well-paired to an HBPP philosophy. Rules-based, simple, infrequent actions, low cost. The ETF would be DIA, seems to work better with Dow than S&P 500. Backtested to 1950... I can't verify with PV.com

I just tested Sell in May on DIA and put it into the HBPP... for 2005-2019 data (admittedly a short period of time... GLD) you can run 61% Sell in May DIA and 13% all other assets, and get the same MaxDD but 94 extra basis points, which is pretty nice.

IT BEATS THE GOLDEN BUTTERFLY!

It's the kind of thing I could tell my daughter to implement for my wife and I when we've got drool coming out of our mouths.

Below is DIA vs DIA six best months... no other assets involved.

Image
Is there any reason to trust that "sell in May" will always continue to work, though? It worked great post 1950....but sucked before then; see:


https://www.cxoadvisory.com/calendar-ef ... -long-run/

https://www.forbes.com/sites/rickferri/ ... away-myth/

http://www.equityclock.com/2017/05/11/s ... y-12-2017/

So what exactly changed around 1949-50 to make it work? Will whatever it is continue into the future?

Re: Stock scream room

Posted: Sat May 25, 2019 8:05 am
by pmward
Yeah just because something back tests doesn't mean it's not anything more than just random luck. In this case, a couple of the largest bear market drops happened to be between May and October, and a couple of the biggest rally's happened to be in the other months. These couple cases severely skew the data. If you take out those couple random outlier years you would come to a different conclusion. And, as last year showed, this isn't always the case. Sell in May and go away would have really burned someone last year.

Re: Stock scream room

Posted: Sun May 26, 2019 3:53 am
by D1984
pmward wrote:
Sat May 25, 2019 8:05 am
Yeah just because something back tests doesn't mean it's not anything more than just random luck. In this case, a couple of the largest bear market drops happened to be between May and October, and a couple of the biggest rallies happened to be in the other months. These couple cases severely skew the data. If you take out those couple random outlier years you would come to a different conclusion. And, as last year showed, this isn't always the case. Sell in May and go away would have really burned someone last year.
Ocho, pmward,

The "sell in May" effect does seem to be merely a spurious correlation/statistical anomaly in that it wouldn't have worked so well before 1950; also, even when it did work (i.e. in the 1950 and later era up to the present) you would still have had higher returns--albeit with higher volatility and higher maxDD--simply by buying and holding through all months rather than by being in cash/T-bills half the year. Time in the market beat timing the market.

Just out of curiosity I went and looked back at S&P 500 monthly returns back to 1927. It appears that most of the worst months were typically in September and October; this applies in the present era as well as pre-1950. If you had simply sold at market close on August 31st and bought back again at market close on October 31st you would have avoided most of the worst of the crashes in 1929, 1930, 1931, 1932 (and would've sidestepped losses in 1933 to turn it from an excellent year into a truly stunning year and would've saved yourself almost a 4% loss in 1934 as well); you would've also avoided the worst two months in the 1937 crash and would've saved yourself a little over 6% in late 1941 and just over 10% in 1946.

It's not all kittens and puppies and rainbows though....this strategy would've hurt you in 1939 (market was up 15.47% and 0.31% respectively in September and October that year) and 1973, 1974, and 1975 (would have made 1973 worse by about 4.4%; 1974 worse by about 3.4%, and 1975 would've had a lower return by about 5.3%...although in compensation you would've gained a extra 3.77% in 1977, 9% in 1978, and almost 6% in 1979).

You would have given up a big gain of 12.90% in 1982 (when in autumn of that year the market was rocketing off the mid-August bottom) but would've missed a loss of 2.15% in September 1987 and the big 20% one-day crash of October 1987 (and missed that month's loss of 21.54% for the whole month). You would've evaded around a 5.2% loss in the 1990 bear market by being out those two months in 1990.

This strategy did very well in the 2000-02 bear market and the 2008 crash; in 2011 it would've cost you a tiny bit; you'd have had a 2.10% positive return by buying and holding but would've had a 0.98% negative return (less than one percent) by staying out those two months. In 2015 it would've cost you about 5.75% and turned a 1.37% gain into a 4.37% loss; but in 2019 it would've yielded around a 0.52% positive return (vs a negative 4.5% return for buy and hold the whole year) for the year by avoiding the sharp decline in October.

Will this strategy continue to outperform in the future? Who knows? YMMV.

EDITED TO ADD: All of the above assumes that for the two months you are out of the market (September and October) you are invested in plain cash earning 0% interest (i.e. earning nothing). If you assume instead that you would've invested in t-bills, a savings account, or 6-month or 1-year Treasuries during those two months then the returns would be even higher (and the losses even less) for the "stay out in September and October" strategy mentioned above.

Re: Stock scream room

Posted: Sun May 26, 2019 8:45 am
by ochotona
1984, what was your data source? Sep / Oct seem really to be bad months overall, I agree. Maybe one could use some indicators like MACD to determine whether taking action during Sep / Oct is justified, and for short periods of time like this, I'm thinking maybe buying puts from time to time would work.

Re: Stock scream room

Posted: Sun May 26, 2019 9:28 am
by D1984
ochotona wrote:
Sun May 26, 2019 8:45 am
1984, what was your data source? Sep / Oct seem really to be bad months overall, I agree. Maybe one could use some indicators like MACD to determine whether taking action during Sep / Oct is justified, and for short periods of time like this, I'm thinking maybe buying puts from time to time would work.
Late 1920s to early 1970s -

Image

Early 1970s to mid-2017 -

Image

If you want to go to before 1928 then Stooq and Shiller have returns back to the late 1800s; going back to 1900 you would've missed some times when September and October were good during those years (like 1915) but also would've missed crappy months in 1902, the Panic of 1907, and 1917

Of course after 1976 VFINX is available and PV lets you backtest it to 1985 and see monthly returns

Re: Stock scream room

Posted: Sun May 26, 2019 2:28 pm
by ochotona
Aw perfect, thanks! My objective is to find out when the really oversized declines occur... the ones we remember for decades and decades. I'd love to side-step those if there is meaningful seasonality + anticipatory technical indicators.

Yup... September October, bad moon rising for some reason.

Re: Stock scream room

Posted: Sun May 26, 2019 6:37 pm
by pmward
ochotona wrote:
Sun May 26, 2019 2:28 pm
Aw perfect, thanks! My objective is to find out when the really oversized declines occur... the ones we remember for decades and decades. I'd love to side-step those if there is meaningful seasonality + anticipatory technical indicators.

Yup... September October, bad moon rising for some reason.
Coming from someone with experience in technical analysis, good luck on that one. That's the holy grail that everyone wishes they had. I don't think what you're looking for is realistic, it's along the lines of finding the fountain of youth. If it were possible to quantify these things, people already would have figured it out long ago, and the minute the cat got out the bag it would no longer be valid. The only thing that is clear is the past. It's easy to ride a confirmed trend. It's almost impossible to call a top, or any change of trend for that matter, in advance, or even in the early stages. There is no way to predict the future. All you can do if you want to trade is look at probabilities, hedge your bets, manage your risk, and pray. If you want to trade it's not about finding times to get all in or all out, it's about finding areas you can place small bets, and limiting your downside risk if you're wrong. You're much better off not trying to place all in bets on picking tops and bottoms. It takes a super human strong stomach to be able to even attempt this.

Re: Stock scream room

Posted: Sun May 26, 2019 10:50 pm
by ochotona
I'm going to paper trade this one

Re: Stock scream room

Posted: Mon May 27, 2019 4:15 am
by D1984
ochotona wrote:
Sun May 26, 2019 10:50 pm
I'm going to paper trade this one
What are you going to use to determine if you should be out or in during the two "bad months" of September and October? MACD? NH/NL? Relative Strength vs cash?

Re: Stock scream room

Posted: Mon May 27, 2019 7:15 am
by ochotona
D1984 wrote:
Mon May 27, 2019 4:15 am
ochotona wrote:
Sun May 26, 2019 10:50 pm
I'm going to paper trade this one
What are you going to use to determine if you should be out or in during the two "bad months" of September and October? MACD? NH/NL? Relative Strength vs cash?
MACD, also the Recognia AI pattern recognition service offered to Schwab clients looks tailor made for this kind of thing. I admit there are many ways to interpret MACD.

Re: Stock scream room

Posted: Mon May 27, 2019 8:17 am
by pmward
ochotona wrote:
Mon May 27, 2019 7:15 am
D1984 wrote:
Mon May 27, 2019 4:15 am
ochotona wrote:
Sun May 26, 2019 10:50 pm
I'm going to paper trade this one
What are you going to use to determine if you should be out or in during the two "bad months" of September and October? MACD? NH/NL? Relative Strength vs cash?
MACD, also the Recognia AI pattern recognition service offered to Schwab clients looks tailor made for this kind of thing. I admit there are many ways to interpret MACD.
MACD gives you a general feel for the strength and momentum of the trend. But it is absolutely useless as a market timing tool, no matter how you interpret it. It can always change on a dime. MACD is ok as a way to get extra confirmation on a narrative, but it's one of the worst tools to try to make buy/sell signals from. Also, I used to trade patterns, patterns fail all the time. False breakouts/breakdowns are a part of the game. This is why traders use risk management strategies like stop losses, scaling into positions, setting percentage limits on each individual position, hedging with options strategies like straddles, etc to be able to get out fast and limit damage if they are wrong.

The risk management strategies are MORE important than the technical strategy itself. Trading patterns and seasonality with MACD you would still be hard pressed to even get 50% accuracy in your calls and using a 100% all in or out bet (i.e. no risk management at all) means you're going to lose big time on the whole. If you want to trade you need to spend your time learning how to manage risk first, before you ever put on a single position. So how are you going to manage your risk with the random assortment of indicators that all are open to subjective interpretation? The reason most traders fail is because they don't pay attention to risk management. Everything for a trader starts and ends with risk management. I'm really trying to hammer this home for a reason. Those traders that can manage risk succeed, those that can't go bust. This is 100% true in all cases.

I'm not sure what's wrong with your current blend of buy and hold, PP hedging, and trend following? It's a much better strategy for someone without the knowledge in how to properly manage risk or the stomach for trading (which I know you don't from how you reacted to the trend following failure in December; you were not able to stick with the system after one false positive). Trading is more akin to gambling than investing, and that's why a lot of traders implement similar risk management strategies to professional gamblers. I'm trying to protect you from yourself here, haha. You're free to do what you want, but I think this is a mistake. If you want to trade, you need to learn from someone who knows what they are doing (and these days it can be hard to find legit teachers that really know what they are doing instead of just saying they do on the internet).

Re: Stock scream room

Posted: Mon May 27, 2019 2:43 pm
by ochotona
I digitized SEP and OCT from 1928 - 2018. If you step away from these two months and go to cash, you gain 14 basis points on average each time you do it, over the entire 91 years. But you avoid some of the worst large drawdowns ever. SEP and OCT have 11 monthly drawdowns greater than 10% for this period. Seems pretty easy to just buy some puts, sell some, whatever makes you feel better. Or do nothing. It's like living in tornado country.

Re: Stock scream room

Posted: Mon May 27, 2019 3:36 pm
by pmward
It's just random noise. Nothing more. The problem with backtesting is that it makes people believe in things that don't really exist.

Re: Stock scream room

Posted: Mon May 27, 2019 3:44 pm
by mathjak107
Well so far templeton was right .. the most expensive words in the English language when it comes to investing is THIS TIME IS DIFFERENT

Re: Stock scream room

Posted: Mon May 27, 2019 3:51 pm
by pmward
mathjak107 wrote:
Mon May 27, 2019 3:44 pm
Well so far templeton was right .. the most expensive words in the English language when it comes to investing is THIS TIME IS DIFFERENT
The context he is speaking of with that comment is much different than seasonality and trying to call market tops.

Re: Stock scream room

Posted: Mon May 27, 2019 3:59 pm
by dualstow
pmward wrote:
Mon May 27, 2019 3:36 pm
It's just random noise. Nothing more. The problem with backtesting is that it makes people believe in things that don't really exist.
It’s the financial version of ‘The Bible Code’.
https://www.goodreads.com/book/show/521 ... Bible_Code

Re: Stock scream room

Posted: Mon May 27, 2019 5:07 pm
by ochotona
I'll be the first to admit that 1200 months is a small dataset.

Re: Stock scream room

Posted: Mon May 27, 2019 6:59 pm
by pmward
It's not even the size of the dataset that matters, some things are simply random anomaly. The data in question, even over many years, is still also skewed heavily by a couple of really bad years, and those selloffs could have just as easily happened in November or December if the news that caused those selloffs would have come a few weeks later. It is also not something that repeats every year; it's a coin flip. I mean, even last year, we rallied all summer, and sold off in the winter during the "best time of the year for stocks". If you torture data enough you can make it tell you anything you want. I'm saying this because I don't want to see you lose a ton of money implementing a bad strategy. I'm also honestly not aware of anybody that really takes the "sell in May and go away" thing seriously. I follow a lot of institutional traders and investors, and I've never seen a single one use that as a basis for any of their moves.

Re: Stock scream room

Posted: Tue May 28, 2019 12:58 pm
by Cortopassi
pmward, I like your writing and insights.

Wondering what your thoughts are regarding technical indicators becoming self fulfilling, since it seems so many people/algorithms use them? It *seems* there are plenty of bloggers (and shysters) who use these indicators very well, or to sell products, newsletter, subscriptions. Their "unique" indicator that gives clear buys and sells all the time.

That is, until you buy their subscription.

TA was yet another method I tried years back, and it simply failed for me. It was an attempt again to get the emotion out of trading and do it robotically, but you've got it right, you have to manage risk first. I did not.

Re: Stock scream room

Posted: Tue May 28, 2019 1:35 pm
by pmward
Oh geez, I could write a book. I know I've mentioned it here in a few places, but I used to be a swing trader, mostly in the realms of individual small cap growth stocks and the general market ETF's (SPY, QQQ, IWM, GLD, TLT, USO, and the Spyder sector ETF's were my favorites). I did pretty well with it and really only stopped because I found that the stress and time involved were not worth it; the juice was not worth the squeeze. And when I discovered the PP and MPT I realized there was a way I could protect myself without having to actively monitor all my positions. Buy and hold with no downside protection has never made any sense to me, and that's what drew me to trading and then in turn to the PP.

So with that experience I can tell you that there is no such thing as a foolproof system. Technical analysis and quantitative analysis do have value; much more value than they get credit for on the popular forums like Reddit, Bogleheads, and even here. It's not so much that TA or QA fail investors, its that investors expectations of TA and QA are way way way too high. There is no foolproof system. There is no magic indicator. All you can do is "analyze" the markets. What TA specifically does is measure market sentiment by looking at some combination of price, volume, trend, and momentum. That's it. It makes no judgements or predictions, that's the art that needs to learned in how to interpret that data; it's simply a way of analyzing data.

When using TA or QA what you are doing is looking at probabilities. TA and QA are backward looking not forward looking. They can describe what happened, but it's up to you to decipher what the probability of what will happen next. But, since nobody can see the future nobody can ever be certain. It's just a game of probabilities. There are always going to be indicators or data points that are bullish and bearish at any period of time. So you don't want to stick to any one data point. I might see a bullish chart pattern, but volume could be low. Large caps could be breaking out, but small caps could be breaking down. Price may be moving up, but momentum may be trending down. You have to look at it from all angles, and use all of that information to make a judgement call. Then, when you have a high probability move, when you place on a position, you need to use risk management to make sure that you don't lose. I used to have a rule where I would never risk more than 1% of my portfolio on any one trade. So, that means if I put on a position that was 10% of my portfolio I could withstand a 10% loss on that position. If I put on a 20% position, I could withstand a 5% loss on that position. If I put on a 100% position, I could only withstand a 1% loss. I also held myself to a maximum of 5% total capitol at risk at any one time, so 5 positions with a maximum loss of 1% of my portfolio each was my max (I usually had 1-2 positions on most of the time, really only in strongly trending markets would I be fully invested). I would also regularly scale into and out of positions to further manage risk. On top of all of that I would also raise my stop loss as my positions went up to lock in gains (there's nothing worse than giving back a large gain, I learned this lesson early on the hard way as one of my first trades was a stock that broke out for a quick 50% gain in the course of a couple days, and I got greedy and wound up closing it out like 3 weeks later for a 13% loss).

Contrary to popular belief, most good traders lose more than they win. BUT they manage their risk so that they cut their losses short and let their winners run. Most traders won't go into a position without at least 3-5x potential gain vs potential loss. Risk management is the most important piece because it means you don't have to be right often, so long as you catch a big winner every now and again. It also limits your total potential downside losses. Trading can work very well, but it has a cost in that it is very time consuming, it takes a long time to learn and a lifetime to master, and it can be very stressful as well. You really get to see a lot of who you are and how you react to fear and greed when you're playing the day to day tape.

So yes, technicals do provide a lot of value, more than they get credit for, but they are not magic nor are they foolproof. Playing with TA or QA without having any form of risk management is playing with fire, and this is where most people get burned. They go all in on TA or QA and pay no mind to the most important part, managing their risk.

Also, things are always changing. One indicator or data point might work really well for awhile, then magically it stops working. TA and QA are strategies that need to be constantly tweaked and adapted as things are always changing. Which is why I'm always skeptical of backtest data conclusions. People find something in a backtest that looks good, then they look for a reason to justify it. It should be the other way around. They should instead for a theory and then look back to see if the data doesn't prove it wrong. Also, backtesting is better at proving what *doesn't* work as opposed to what does work.

Finally, for people trying to sell you their wares, as with anything you have to be careful and skeptical of everything. There are some legit resources out there that will teach you some stuff, but there are sure a lot of false prophets out there as well. In general, it's better to be taught how to fish as opposed to paying for fish. The better TA educators out there are the ones that teach you how to perform analysis, not the ones that try to lock you into their proprietary signals. And any teacher that teaches TA or QA without covering risk management strategies are someone I would also avoid. The good teachers always start with risk management and move out from there.

One last thing I will say is that even though I don't "trade" using technicals anymore, I still use technicals all the time to analyze the assets I'm invested in with my golden butterfly. I use what I see here to help guide my decisions of when to rebalance, how much to rebalance, if I should cut my cash position down a few percent to buy another asset on a dip, what specific assets I should put money in on each paycheck, etc. Small tweaks that add a bit of alpha here or there. I also view the small cap portion as a 20% VP and if I ever have a very strong conviction in a certain narrative come up I would not hesitate to use some of that allocation to play it.

Re: Stock scream room

Posted: Wed May 29, 2019 9:11 am
by pmward
Speaking of technicals... you all can thank me for the S&P losing support today. I'm vesting for a very large chunk of company stock next Monday, so of course we need to sell off right beforehand, haha. Ugly break of major support today though. I'm really hoping for a head fake false breakdown and a quick recovery back above 2800 before the weekly close. If not, the next support level is all the way down at the 2625-2675ish range so we could test those levels in the coming weeks. This is the first real technical damage of the recent pullback, we just opened the door to a possible correction if stocks don't catch a bid here and bounce back up. On the plus side... TLT is breaking out above resistance today. It's been such a strong run that I wouldn't be surprised to see a pullback or sideways consolidation here. If we keep going up the next resistance level is about 132.50. I would be very surprised to see it continue to go up to those levels uncontested though, as the market is a bit extended in it's current run. But if stocks continue to sell, anything is possible. Either way, my theory of a blowoff top to new all-time highs in bonds over the next year or two to end the decades long bull run is looking on target.

Re: Stock scream room

Posted: Sat Jun 01, 2019 10:20 am
by jacksonM
Too bad Budd's not here any more. He always got upset when the PP didn't immediately perform as expected in the short term so today he might have been happy. With stocks taking a big hit I think my SEP-IRA which is mostly LTT's and gold might have had its biggest one day gain ever.

So will this be a continuing story for a while like it was when I first started the PP? Stay tuned.

Re: Stock scream room

Posted: Sat Jun 01, 2019 10:28 am
by Kbg
I run a leveraged PP and, ya. Big pop. TMF is up huge.

Re: Stock scream room

Posted: Sat Jun 01, 2019 11:45 am
by pmward
Yeah I didn't think TLT was going to test the resistance level at 132.50 without a pullback first. But there seems to be nothing that can stop it, it's taken off like a rocketship. If it gets above that the next big supply level is 138.50, and all time highs are just a hop, skip and a jump from there. I'm not sure how much further this rally can go, but man it has some serious steam behind it, so who knows?

Re: Stock scream room

Posted: Mon Jun 10, 2019 11:29 am
by dualstow
I am very happy for PointedStick as I believe is all in stocks. As long as I live, I don't think I could ever go back to a 100% allocation. I only did that when I was starting out in investing, before my dad said, "You really should have some bonds."

S&P is closing in on its all-time high.