Skip to main content
No. of Recommendations: 8
Hi guys,

I hadn't looked at them for years, until now, but I'm curious whether you folks ever look at SIPro's built-in screens. I just noticed that some of them seem to be doing great and one is on a real tear. O'Shaughessy's Tiny Titans is up some 14,000% since 2010 and looks like it's running straight up the chart so far this year.

Passing Companies: 37
Average 12m gain: 258.0%
Median 12m gain: 294.8%
Top 12 12m Avg gain: 398.6%
Minimum 12m gain: 88.4% (remaining 36 all >100%)

Avg gain YTD: 108.8% -- Beating me by 100% :)

I don't know how much trouble it would be, but would it be worth back-testing for the heck of it?

Thanks for considering,

Dan
Print the post Back To Top
No. of Recommendations: 8
I found the Tiny Titans was theoretically nice but a little too tiny with too much slippage. In the past I’ve used a slightly modified version which I labeled Mid Titans. Be aware the AAII screens don’t reflect real trading, they include all passing stocks no matter how many or how few and no consideration of slippage.

From 20200103 till last Friday for monthly run with my backtester:
ci.EXCHANGE <> 'O' \
ci.COUNTRY = 'United States' \
PSD.PRICE * PSD.AVM_03M/21 > 450 \
PSD.MKTCAP >= 50 \
PSD.MKTCAP <= 225 \
ISQ.EPS_Q1 > 1.03 * ISQ.EPS_Q5 \
PSD.PRICE > 0.85 * PSD.PRICEH_52W \
MLT.PSPS > 0 \
MLT.PSPS <= 1 \
order by (psd.RS_26W) DESC
Top 10 Hold till drop 15

CAGR Fri 72.1% Mon using (O,(HL)/2,C)/3 66.5%, Friction 0.35% on all trades, GSD 34.8% CAGR/GSD 2.1
UlcrIndx 4.2% MxDD 9.3% Annual Turnover 795%

Not my results, I’ve been mostly in less gain, less excitement much less effort ETF allocations.
But can't completely quit, still run a couple of screens

RAM
Print the post Back To Top
No. of Recommendations: 0
please convert to English or gtr
Print the post Back To Top
No. of Recommendations: 0
Thank you, RAM, much appreciated!

I just have one question: "Top 10": by ... by Returns, P/S or ...?

Thanks again.

Dan
Print the post Back To Top
No. of Recommendations: 3
Musselmant: please convert to English or gtr
The definitions are directly from the AAII Stock Investor Pro Field Definitions they pretty much match the GTR1 glossery if you leave off the XXX. Prefix.
Example Psd.AVM_03M = Data Category Price and Share Statistics, AVM_03M The average cumulative daily trading volume for each of the last three months (roughly the last 90 days, on a rolling basis).

Raptor:
The order by RS_26W sorts the relative strength returns from highest 26 week return and takes the highest 26 week relative strength but does not sell and buy a new one until that stock drops out of the top 15.

But although I used a version very close to this for a couple of years, I would not recommend it!
Backtests show very good results the last few years however the returns from 1997 to present have a CAGR around 35% but GSD is 28%. This is using only one start date every 21 days unlike GTR1 which tests all 21 start dates. However the test includes 1444 Buys with 2500 monthly Holds.
The biggest problem is the smaller cap stocks have less consistently valid and recent data and in some cases slippage was much less predictable compared to larger cap stocks.
Print the post Back To Top
No. of Recommendations: 0
Tiny microcap stocks...
... at snapshot prices that may have been unachievable...

not something anyone would (should) commit more than 10% of a portfolio to (1 or 2% of a portfolio per stock at top 10 or top 5)...

that said, not a bad idea for an allocation.
Print the post Back To Top
No. of Recommendations: 11
MidTitans uses top 10 by 26 week total return.

The AAII screen results do not include trading costs (friction) and some screens like Tiny Titans pick illiquid stocks with high trading costs.

backtests for 54 AAII screens, April 1, 2015
https://boards.fool.com/backtests-for-54-aaii-screens-316942...

Tiny Titans 12 month CAGR ranged from 114% to 218% in gtr1 (depending on the day of the month trades were done). YTD TR range was 42% to 76%.
http://gtr1.net/2013/?~TinyTitans_20150401_numerous:h21f0.4:...

a gtr1 version of MidTitans posted above:
http://gtr1.net/2013/?~MidTitans_20210617_RAMcN:h21f0.35::st...
Print the post Back To Top
No. of Recommendations: 3
Borisnand, thanks for the GTR1 links!

I'm alway curious if people would have the gumption to continue following a screen after a string of bad relative performances. For instance, here's the Mid-Titan CAGR for the past coupla years compared to the S&P 500:

      Mid-Titans  S&P 500  Mid- vs S&P
Year CAGR CAGR
2014 7.69 13.67 -5.98%
2015 10.62 1.31 9.31%
2016 8.79 11.85 -3.06%
2017 0.26 22.04 -21.78%
2018 13.14 -4.33 17.47%
2019 1.67 31.33 -29.66%
2020 3.32 18.89 -15.57%
2021 35.96 13.13 22.83%

The question is, if you started this screen in 2014, and then had 4 of 6 losing years, the last 3-of-4 of which were quite large losses to the S&P, would you have stuck around for the 2021 gains. My guess is maybe no.

Personal anecdote: I started doing YieldEarnYear a REALLY long time ago (2006?), and finally, after years and years of under-performance, I gave up on it last Spring. And then promptly lost out on its spectacular relative out-performance since last summer. I think I stuck with it much longer than most - really gave it the benefit of the doubt.

Tails
Print the post Back To Top
No. of Recommendations: 4
I gave up on it last Spring. And then promptly lost out on its spectacular relative out-performance since last summer.

There's only so many directions one can take a portfolio in at once, and your faith in YEY over the years highlights the problem with any specific allocation / approach decision, backtested: how long is relative underperformance reasonably tolerable? The range of expected change decision triggers is probably >6 months to <3 years. You can be commended for giving YEY that long a chance to come back. And we have no idea whether that spectacular relative outperformance signals a market shift to its Factors coming back into favor, or a temporary CV-fuel/MLP related spike.
Print the post Back To Top
No. of Recommendations: 8
I gave up on it last Spring. And then promptly lost out on its spectacular relative out-performance since last summer.

My condolences.

Classic YLDEARNYEAR 1-5 and 1-10 have both returned about 120% (not annualized) in the last 14 months.
About 62% for SPY in the same period.
The difference since the start of December has been particularly pronounced and steady, beating the S&P for seven months running.

But the overall performance lag since the start of 2014 is so significant that even 50-60% is just a blip.
From then till last month, -1.9%/year versus +14.2%/year for SPY.

I have a hunch that earnings and dividends might mean something again some day.
But I haven't put much money into that theory. (some for earnings, none for dividends)

Jim
Print the post Back To Top
No. of Recommendations: 1
It takes a lot of time to validate a (relatively) low Sharpe ratio trading model, and this is
without taking into account the non-stationarity of the market.

For example you may look at Irene Aldridge's High Frequency Trading book, page 60, Table 5.2:
with monthly performance data, a strategy with Sharpe ratio 1 needs almost 3 years of actual trading performance data to be validated, even if one has daily's performance data the situation does not seem to improve considerably (just a couple of months less):


Claimed Annualized Sharpe Ratio vs Minimum Number of months required (monthly performance data):

0.5 131
1.0 34
1.5 16
2.0 9
2.5 7
3.0 5

The signal to noise ratio is really weak most of the time ...

Stefano
Print the post Back To Top