I’ve written an article for Seeking Alpha that I think P123 users might enjoy. It also includes links to a public high-performing universe, screener, and ranking system that are quite different from those I’ve seen from P123 users. Here’s the SA link:
LLCs. The complete list is AB AHGP ALDW AM AMID AMJ AMLP AMZA ANGI APLP APO APU ARCX ARES ARLP ATAX ATLS AZURQ BBEPQ BEP BIP BKEP BPL BPY BSM BWP BX CAFD CBA CCLP CELP CEM CEN CEQP CG CHKR CINR CLMT CNNX CNXC CODI CPLP CPPL^17 CQP CTR CVRR DCP DKL DLNG DM DMLP DPG DSE ECT EEP EFC EMES EMLP EMO ENBL ENFR ENLC ENLK EPD EQGP EQM ETE ETP EVA EVEP FEI FELP FEN FGP FIF FIG FMO FPL FUN GEL GER GLOP GLP GMLP GMZ GPP HCLP HEP HMLP IEP JMF JMLP JMP JPEP^17 KED KKR KMF KNOP KYE KYN LGCY LINEQ^17 LMRK MCEP MEMP MEP MIE MLPA MLPG MLPI MLPJ MLPS MLPX MLPY MMLP MMP MPLX NAP NBLX NDP NEN NEP NGL NML NMM NNUTU NRP NS NSH NTG OAK OCIP OKS OZM PAA PAGP PBFX PER POPE PSXP PTXP RHNO RMP SDLP SDR SDT SEP SGU SHLX SMLP SMM SPH SPP SRF SRLP SRV STON SUN SXCP SXE SXL SZC TCP TEGP TEP TGP TLLP TLP TNH TOO TPYP TPZ TTP TYG UAN USAC USDP VLP VNOM VNRSQ VTTI WES WGP WLKP WMLP WNRL WPT WPZ YMLI YMLP ZMLP
What do you mean? I just wanted to include Canadian companies that are on one of the major US exchanges.
Those two stocks were appearing in my results with increases of 30,000% or so, which really made my rolling backtest results far rosier than they should have been. That’s because their initial prices were extremely low (a few pennies). So I excluded them to make my results more realistic.
I notice little difference in performance between your custom universe and the Russell 3000. This is good, IMO.
I wonder about the 3 year growth rate formula whereby you are taking the absolute difference between the compound growth rate and 7.5%. This seems very arbitrary, but I do understand that you are looking for sustainable growth and that you want to penalize both negative and rapid growth. Why not discard the arbitrary 7.5% and instead look at idiosyncratic growth? Given that the underlying factor is sustainable growth, I personally found that substituting the following worked very nearly as well:
** relative growth rate, 3 year, industry:
abs(Sales3YCGr%Ind - Sales3YCGr%)
Anyway, I dropped fiscal momentum and accruals from my main ranking system in favor of other metrics (I have a personal cap on the number of factors I use), but seeing them perform well here makes me want to revisit these factors. Part of the issue was I couldn’t decide on a single metric for either fiscal momentum or good accountancy. Fiscal momentum is difficult because there are so many ways to cut the data, and – as you point out in your metrics – there is such thing as too much of a good thing. I also have trouble with quantifying good accountancy – most accruals measures are reflected in cash measures of profit. I also find it peculiar that many accrual metrics penalize increases in net operating asset value. Should not long-term growth in book value be seen as a good thing?
Using the relative growth rate is a great suggestion. I’ll have to try that. I just didn’t think of it.
Regarding net operating assets–you want those to be low compared to total assets. NOA is not the same as book value: NOA takes book value and subtracts cash and equivalents and adds debt. I like companies with modestly increasing total assets (again, around 7.5% is good), but I like companies with decreasing NOA, and low NOA-to-asset ratios. See this paper: www.haas.berkeley.edu/groups/finance/HHTZ-032904-jae.pdf.
NetIncBXorNonC(3,TTM) is the TTM value 3 quarters ago, not 3 years ago. Is that what you intended? From my interpretation of the thread and what you described, you want to compare the latest NetIncBXorNonC with the value from 3 years ago.
When you use “TTM” the index value indexes quarters ago, not years ago. This tripped me up for a while.
So if you’d wanted the TTM value from 3 years ago, you would use:
NetIncBXorNonC(12,TTM)
If you meant “TTM NetIncBXorNonC 3 quarters ago”, my apologies. But then I’d have to ask out of curiosity - why look at 3 qtr ago TTM numbers?
Great thanks yuval. I actually found that exact site and file independently while waiting for your reply. Looks like the list has added some tickers since even when you posted it. Here’s a fresh list (copy/paste friendly):
I noticed in your list there are some expired tickers: “LINEQ^17”, “JPEP^17”. How did you discover those? Looks like they are proper removed from this new list I posted (current as of Mar 30 2017)
This topic is starting to really peak my interest. I’m playing around with the Abs(Sales%ChgTTM - 5) function from the previous version of this strategy. I am playing with:
if sales%TTM >0, subtract x and flip the sign
if sales%TTM <0, subtract 2x and flip the sign
Lower is better
Conceptually, it would start to drive a bigger wedge between companies that are close to 0% but on either side of it. Using an x of 5 and 10 doesn’t seem to make any difference just quickly ranking with all fundamentals but would make me feel more at ease using the factor. Any thoughts?
Also using sales% alone confirms the original stance that the extremes of this don’t give you the expected return.