Everyone can do machine learning using ChatGPT

General
18

Wycliffes,

I found the Benjamini-Hochberg Procedure interesting and it was new to me. It is actually pretty simple Here is an explanation for those interested in the statistics or who want a more in-depth understanding of the paper: Benjamini-Hochberg Procedure

Much appreciated,

Jim

19

This paper relies on Chen & Zimmerman’s “Open Source Cross-Sectional Asset Pricing,” which can be found here: Open Source Cross-Sectional Asset Pricing by Andrew Y. Chen, Tom Zimmermann :: SSRN. They have a website: https://www.openassetpricing.com/. I was unable to find in the website a description of each of the factors, but I’m assuming it’s there somewhere, even if its couched in Compustat codes. There are links to GitHub and to Google Drive. If someone can help me, I would be very grateful.

20

There is an Excel file:
CrossSection/OldSignalDocumentation.xlsx at master · OpenSourceAP/CrossSection · GitHub

Might be helpful but I haven’t looked it over closely.

21

I have gathered some information from various studies with an overview and a brief explanation of each study in the link provided above. I have also used Chat GTP to extract a summary of most of the studies. In the English explanation, I have only used…“=GOOGLETRANSLATE(N4;“nb”;“en”)”

22

Yes, the “Description” column in the CZ tab of that file is just what I was looking for. Thank you!

23

TL;DR: The authors of this paper using Bootstrapping to get p-values along with the Benjamini-Hochberg Procedure (BH) to get a false discovery rate (FDR) correctly state that they get a different result from a paper using Bayesian analysis: Is There a Replication Crisis in Finance?

Both papers are ultimately about False Discovery Rates (FDR) so there is a true disparity in the results of the papers, I believe.

The latter paper has been sited frequently in the forum. I first became aware of it when Yvual mentioned it and he has mentioned it since as well as blogging about it I think:. E.g, in this thread: Why such a large discrepancy between backtesting and real-life results?. it is an interesting paper and I have also discussed it in the forum. I don’t mean to imply that Yuval or I endorse or use anything within the paper or what we might use if we do use some of it. Probably we find some similar lessons as well as some different lessons from the paper.

So which paper is right? Lots of good factors or surprisingly few?

One could probably explore this on their own with P123 data. The BH is not that hard to do. The bootstrapped p-value could be a little challenging. I am not sure that there is a Python library that does this (although there is a library in Scikit-Learn that gives confidence intervals using bootstrapping), but you can find multiple examples of code to do this with a google search. I believe R does this and a bootstrapped p-value could be obtained using an iterative approach over at JASP.

Here is a simple video by Stat Quest explaining getting a p-value using Bootstrapping: Using Bootstrapping to Calculate p-values!!! For someone on the forum who has sent me links about Stat Quest (you know who you are or if you have ever watched the Stat Quest videos) I cannot help but say: Bam!

Some Bayesian analysis can be found at JASP, PyMC3 with Python or over at R using the BEST program.

Jim

24

I’ve made a quick start in converting the descriptions to P123 factors (or what I understood from it). Took the 6 (edit: 26 now, will do more later) with the highest t-value , hope to add them all later.

https://www.portfolio123.com/app/ranking-system/409467

Results are not bad.

1 Like
25

Thanks, Victor. That’s nice and helpful.

26

Thanks Victor. Very helpful.

1 Like
27

I think the topic of this thread may have changed and I may be hijacking it a little. That’s cool and I like people changing the topics. My apologies if this is off topic now.

But the paper does talk about bootstrapped p-values which I find, personally, to be one of the most interesting parts of the paper.

AND I did get the code from ChatGPT which is consistent with the original post. The print out the below code (and a random factor) was P-value: 0.71199999999999996625

FWIW for those interested……

Code:

import numpy as np
import pandas as pd
observed_statistic = 0

Load data from file

df = pd.read_csv(‘~/Desktop/bootstrap.csv’)
data=df[[‘factor’]]

Define the function to calculate the test statistic (e.g. mean)

def test_statistic(data):
return np.mean(data)

Define the number of bootstraps to perform

n_bootstraps = 1000

Initialize an array to store the bootstrap replicates

bootstrapped_statistics = np.empty(n_bootstraps)

Perform bootstrapping

for i in range(n_bootstraps):
# Resample data with replacement
resampled_data = np.random.choice(data.values.flatten(), size=len(data), replace=True)
# Calculate test statistic for resampled data
bootstrapped_statistics[i] = test_statistic(resampled_data)

Calculate the p-value

#observed_statistic = test_statistic(data.values.flatten())
p_value = np.sum(bootstrapped_statistics <= observed_statistic) / n_bootstraps

Print the p-value

#print("P-value: ", p_value)
print("P-value: ", “%0.20f” % p_value)

28

Thank you!!!

I see that you have assigned different weights to the different nodes. How did you determine the weight of each node?

I took this list from a study, converted it, and sorted it.

Acronym Economic category R t(R) SD SR maxDD Min 0,05 0,95 Max Start N
STreversal Other 35,56 13,989 24,316 1,462 50,364 −36,964 −4,485 14,123 79,534 10805 1,098
IO ShortInterest Other 34,32 4,803 45,799 0,749 70,314 −58,515 −17,895 24,4 46,499 29189 493
RD q Profitability 17,385 4,26 23,083 0,753 36,337 −14,516 −6,727 12,271 45,744 32539 384
IndRetBig Momentum 16,529 11,663 13,557 1,219 64,652 −24,055 −4,223 7,183 34,887 10805 1,098
ProbInformedTrading Trading frictions 16,034 4,325 16,159 0,992 27,94 −25,799 −5,687 7,995 15,767 30741 228
Frontier Intangibles 15,599 6,651 17,785 0,877 41,978 −18,694 −6,052 9,343 25,927 23223 690
AccrualsBM Investment 14,974 6,299 17,854 0,839 41,987 −17,393 −5,982 8,985 31,215 23589 677
SmileSlope Trading frictions 14,699 10,479 7,001 2,099 4,775 −4,653 −1,22 4,196 15,482 35124 299
FirmAgeMom Momentum 14,562 7,056 19,543 0,745 85,36 −43,68 −7,27 8,968 32,022 10805 1,076
retConglomerate Momentum 13,997 6,753 13,433 1,042 24,048 −16,426 −4,368 8,037 21,581 27817 504
roaq Profitability 13,804 5,206 19,575 0,705 66,879 −33,62 −7,489 8,434 42,393 24317 654
EPq Value vs, growth 13,652 12,36 8,472 1,611 37,4 −13,883 −2,369 4,906 12,79 22735 706
CFq Value vs, growth 13,434 6,362 16,264 0,826 72,984 −39,397 −5,16 6,67 32,551 22553 712
IntMom Momentum 13,308 5,545 22,958 0,58 88,552 −83,162 −8,599 10,886 20,057 10805 1,098
AnnouncementReturn Momentum 13,184 14,3 6,481 2,034 13,483 −19,375 −1,533 3,363 6,943 26176 593
SP q Value vs, growth 12,761 6,168 15,935 0,801 66,272 −36,687 −4,668 7,546 30,01 22553 712
MS Other 12,29 5,003 16,464 0,746 32,814 −19,743 −4,647 7,197 65,859 27302 539
ChNAnalyst Intangibles 12,088 1,84 37,447 0,323 99,981 −99,843 −10,287 13,574 55,667 27908 390
BMq Value vs, growth 11,959 4,936 18,212 0,657 52,478 −28,28 −5,351 8,163 39,226 23589 678
Mom6mJunk Momentum 11,686 3,346 21,627 0,54 53,689 −36,607 −7,183 8,839 42,169 28853 460
XFIN Investment 11,679 4,836 16,817 0,694 61,192 −36,495 −5,99 8,208 24,596 26511 582
ChTax Intangibles 11,678 10,049 8,87 1,317 21,738 −20,135 −2,648 4,727 18,141 22950 699
MomOffSeason Other 11,535 5,47 20,173 0,572 63,137 −14,896 −5,955 9,224 59,863 10805 1,098
ChangeRoA Profitability 11,065 7,787 10,393 1,065 38,746 −19,908 −2,886 4,576 17,454 24684 642
OperProfRDLagAT q Profitability 10,99 4,371 17,601 0,624 67,867 −45,839 −7,098 7,137 19,732 26329 588
MomSeasonShort Other 10,98 6,629 15,844 0,693 76,656 −54,927 −5,166 6,675 43,246 10805 1,098
AssetGrowth Investment 10,953 7,36 12,317 0,889 35,141 −9,133 −3,978 6,394 28,474 19206 822
NetEquityFinance Investment 10,836 5,378 14,033 0,772 53,302 −26,395 −5,692 6,355 18,058 26511 582
RD Profitability 10,822 5,567 16,206 0,668 45,119 −15,353 −5,082 8,037 50,778 18840 834
BM Value vs, growth 10,817 5,02 16,62 0,651 48,562 −25,557 −5,14 7,991 40,615 22493 714
ChangeRoE Profitability 10,813 6,858 11,533 0,938 32,843 −16,991 −3,18 4,249 47,698 24684 642
MomVol Momentum 10,757 3,481 29,56 0,364 99,313 −68,026 −11,043 11,546 37,009 10805 1,098
EarningsStreak Other 10,655 10,064 6,367 1,673 14,726 −14,726 −1,678 3,672 5,792 31016 434
AssetLiquidityMarket Other 10,564 7,569 10,621 0,995 34,046 −15,988 −3,277 5,62 17,601 23070 695
NOA Investment 10,314 7,101 11,062 0,932 40,426 −12,749 −3,967 5,737 27,47 23042 696
PS q Other 10,286 5,751 10,88 0,945 47,416 −13,94 −4,492 5,216 12,435 30712 444
ResidualMomentum Momentum 10,279 8,113 12,058 0,852 43,496 −29,36 −4,184 5,683 17,812 11139 1,087
cfpq Value vs, growth 9,781 5,407 12,565 0,778 47,498 −29,978 −4,275 5,664 14,11 26603 579
Price Other 9,713 3,106 29,918 0,325 79,409 −24,709 −8,14 13,004 100,833 10805 1,098
ShortInterest Trading frictions 9,553 5,975 11,068 0,863 20,305 −15,606 −4,439 5,62 16,318 26723 575
REV6 Momentum 9,534 4,131 15,368 0,62 64,114 −34,365 −6,128 6,317 14,066 28033 532
PS Other 9,52 2,824 23,601 0,403 61,673 −37,699 −8,8 10,718 38,949 26329 588
Mom12mOffSeason Other 9,507 3,443 26,415 0,36 97,148 −87,633 −10,624 9,943 29,996 10805 1,098
Mom12m Momentum 9,296 3,13 28,413 0,327 99,532 −88,699 −11,678 10,508 29,484 10805 1,098
dNoa Investment 9,251 8,419 8,404 1,101 17,888 −14,126 −2,654 4,623 16,915 22858 702
MomSeason Other 9,173 6,861 12,789 0,717 58,009 −27,835 −4,105 6,268 24,178 10805 1,098
DolVol Trading frictions 9,164 4,109 21,331 0,43 46,791 −16,811 −6,052 7,922 83,559 10805 1,098
CBOperProfLagAT q Profitability 9,091 5,741 11,151 0,815 39,919 −28,827 −4,442 4,766 14,235 26114 595
AgeIPO Intangibles 8,995 3,316 17,173 0,524 54,993 −23,749 −7,144 7,796 21,436 29586 481
FEPS Other 8,843 2,848 20,811 0,425 80,805 −31,77 −7,216 10,261 31,558 27817 539
AssetLiquidityMarketQuart Other 8,732 5,001 13,115 0,666 47,901 −14,251 −4,816 6,71 19,52 23620 677
EarningsSurprise Momentum 8,607 9,543 6,844 1,258 32,039 −13,24 −2,326 3,658 9,629 23190 691
OScore Profitability 8,459 3,276 18,076 0,468 65,41 −42,947 −7,634 8,33 20,511 26329 588
GPlag q Profitability 8,4 4,878 12,054 0,697 56,507 −22,77 −4,786 5,319 10,972 26329 588
RIO MB Other 8,37 4,166 15,259 0,549 70,199 −19,43 −5,652 7,779 26,724 23042 692
DelDRC Profitability 8,271 1,582 23,673 0,349 48,281 −42,979 −3,801 5,549 42,596 36738 246
SP Value vs, growth 8,269 5,123 13,457 0,615 58,84 −25,006 −4,67 6,493 20,62 18840 834
CredRatDG Profitability 8,212 3,145 14,537 0,565 56,133 −23,704 −6,132 6,529 12,815 31471 372
BMdec Value vs, growth 8,125 6,002 11,204 0,725 43,023 −16,354 −4,12 5,823 18,12 19206 822
AMq Value vs, growth 8,071 3,587 16,912 0,477 74,096 −40,193 −5,875 8,328 21,328 23589 678
MomSeason06YrPlus Other 8,014 7,009 10,786 0,743 39,311 −21,163 −3,203 5,274 32,475 11718 1,068
RDIPO Intangibles 7,986 3,291 16,034 0,498 54,932 −26,903 −6 7,171 19,231 28156 524
LRreversal Other 7,929 3,658 20,733 0,382 67,396 −22,419 −5,603 7,674 75,696 10805 1,098
BidAskSpread Trading frictions 7,873 2,443 30,825 0,255 84,23 −22,807 −8,812 12,451 102,674 10805 1,098
MomOffSeason06YrPlus Other 7,854 6,034 12,286 0,639 43,147 −31,543 −3,781 5,546 41,583 11688 1,069
AnalystRevision Momentum 7,801 8,734 5,981 1,304 24,575 −13,81 −2,132 3,265 5,551 27850 538
InvGrowth Investment 7,771 7,191 8,945 0,869 34,881 −8,768 −3,457 4,663 16,469 19206 822
RIO Disp Other 7,766 3,534 14,688 0,529 52,278 −16,378 −5,047 7,493 25,716 27817 536
NetDebtFinance Investment 7,729 8,854 6,079 1,271 15,284 −4,989 −2,139 3,568 8,317 26511 582
Cash Value vs, growth 7,705 3,137 17,236 0,447 62,28 −16,82 −6,668 7,818 46,972 26235 591
EntMult Value vs, growth 7,7 5,418 11,848 0,65 49,869 −17,387 −4,551 6,28 19,392 18840 834
NetPayoutYield Value vs, growth 7,672 4,332 14,551 0,527 59,081 −23,302 −5,789 7,424 17,538 19571 810
OrgCapNoAdj Intangibles 7,67 4,989 12,818 0,598 48,007 −17,128 −4,183 6,163 33,081 18840 834
DivYield Value vs, growth 7,628 3,464 21,061 0,362 77,561 −31,775 −8,32 9,507 88,791 10805 1,098
ExchSwitch Trading frictions 7,486 4,924 11,596 0,646 26,117 −13,52 −4,582 6,374 12,93 22980 698
EBM q Value vs, growth 7,464 5,685 9,603 0,777 28,163 −17,715 −3,055 4,308 21,2 24590 642
MomRev Momentum 7,454 3,507 20,322 0,367 95,38 −60,502 −7,601 8,663 35,977 10805 1,097
CustomerMomentum Other 7,437 2,136 22,965 0,324 87,213 −61,346 −6,128 9,473 36,098 28335 522
DivOmit Value vs, growth 7,433 4,146 16,946 0,439 62,135 −26,05 −6,89 7,662 42,036 10805 1,072
iomom cust Momentum 7,264 3,175 13,518 0,537 44,183 −33,24 −4,821 6,553 15,782 31471 419
RevenueSurprise Momentum 7,233 8,584 6,394 1,131 18,352 −12,137 −1,746 2,927 14,765 23190 691
ChInv Investment 7,2 8,028 7,423 0,97 24,401 −6,697 −2,644 4,178 15,967 19206 822
betaVIX Trading frictions 7,167 3,328 12,723 0,563 44,428 −13,073 −4,288 6,959 18,266 31471 419
MaxRet Trading frictions 7,166 2,682 25,561 0,28 84,963 −45,188 −10,793 11,013 48,034 10805 1,098
Activism2 Intangibles 7,032 1,763 16,157 0,435 54,174 −12,828 −6,804 7,906 16,228 33177 197
iomom supp Momentum 7,028 2,953 14,044 0,5 39,746 −22,748 −5,279 6,547 20,423 31471 418
PM q Profitability 7,025 2,947 18,363 0,383 66,098 −47,176 −8,132 7,529 20,543 22553 712
RetNOA q Profitability 6,997 3,255 16,123 0,434 63,183 −36,101 −6,855 7,057 18,418 23680 675
DelBreadth Intangibles 6,88 2,899 14,976 0,459 42,848 −29,964 −5,476 6,541 29,975 29433 478
VolumeTrend Other 6,864 5,387 12,188 0,563 29,105 −25,261 −3,691 5,264 45,626 10805 1,098
ChangeInRecommendation Intangibles 6,745 7,128 4,924 1,37 7,211 −6,466 −1,408 2,642 6,657 34334 325
zerotradeAlt1 Trading frictions 6,741 3,638 17,724 0,38 56,42 −27,511 −6,513 8,349 54,367 10805 1,098
OperProfLag q Profitability 6,731 3,001 16,404 0,41 68,712 −39,694 −6,718 6,686 18,946 24590 642
NetDebtPrice Value vs, growth 6,727 4,133 12,343 0,545 35,617 −15,666 −4,773 5,217 21,375 23223 690
InvestPPEInv Investment 6,598 8,666 6,324 1,043 28,493 −7,032 −2,053 3,457 13,252 19024 828
CapTurnover q Other 6,572 4,065 11,936 0,551 48,146 −17,783 −4,505 5,518 22,782 24317 654
DivYieldST Value vs, growth 6,554 8,976 6,981 0,939 17,29 −11,302 −1,963 3,267 34,66 10805 1,097
RIO Volatility Other 6,54 3,507 17,813 0,367 73,855 −21,933 −7,098 8,195 50,584 10805 1,095
AssetTurnover q Other 6,451 4,68 10,453 0,617 31,334 −12,193 −4,277 5,167 16,28 23223 690
PriceDelayRsq Trading frictions 6,438 3,466 17,737 0,363 43,594 −31,971 −5,739 7,316 64,449 10805 1,094
MomSeason11YrPlus Other 6,435 7,117 8,287 0,777 23,154 −18,298 −2,924 4,337 12,701 13545 1,008
zerotrade Trading frictions 6,432 3,305 18,614 0,346 46,739 −27,052 −7,2 7,963 67,172 10805 1,098
Mom6m Momentum 6,282 2,288 26,267 0,239 99,264 −77,393 −9,823 8,814 32,091 10805 1,098
IdioRisk Trading frictions 6,26 2,66 22,511 0,278 88,132 −39,119 −10,011 9,601 38,075 10805 1,098
tang q Intangibles 6,218 4,622 9,497 0,655 52,453 −9,233 −3,559 4,753 27,608 26023 598
DelFINL Investment 6,152 11,461 4,443 1,385 18,264 −5,483 −1,545 2,52 7,868 19206 822
MomSeason16YrPlus Other 6,026 6,577 8,144 0,74 16,833 −8,477 −2,89 3,969 17,795 15372 948
std turn Trading frictions 5,982 2,895 19,769 0,303 80,934 −45,882 −8,529 8,966 25,169 10805 1,098
CBOperProf Profitability 5,936 3,523 12,886 0,461 53,472 −14,01 −5,3 6,505 17,411 22858 702
ConsRecomm Other 5,891 2,409 12,746 0,462 42,565 −12,732 −5,186 6,761 19,279 34303 326
skew1 Trading frictions 5,853 4,086 7,151 0,818 18,39 −9,273 −2,32 3,658 7,431 35124 299
OPLeverage q Intangibles 5,831 3,274 13,118 0,445 42,393 −15,361 −5,231 6,544 21,029 24317 651
ReturnSkew Trading frictions 5,82 7,715 7,216 0,806 31,453 −18,157 −2,099 3,169 12,179 10805 1,098
DelEqu Investment 5,585 3,969 10,761 0,519 36,342 −8,165 −3,651 5,597 21,525 22858 702
RDcap Intangibles 5,563 2,954 11,985 0,464 39,292 −9,083 −4,353 6,098 20,935 29433 486
sfe Value vs, growth 5,529 1,714 21,573 0,256 88,5 −54,648 −8,49 9,102 21,883 27880 537
AdExp Intangibles 5,515 2,577 16,114 0,342 53,395 −20,097 −5,655 6,895 43,517 20299 680
ShareVol Trading frictions 5,502 3,197 16,395 0,336 77,968 −29,05 −7,302 7,471 28,075 10805 1,089
PayoutYield q Value vs, growth 5,439 4,427 9,049 0,601 30,056 −13,113 −3,384 4,117 16,62 24411 651
depr Other 5,431 3,677 12,357 0,44 51,003 −14,443 −4,405 5,822 30,729 18659 840
ChEQ Intangibles 5,404 4,02 10,282 0,526 29,477 −14,126 −3,669 4,943 21,503 22858 702
IdioVol3F Trading frictions 5,349 2,233 22,907 0,233 91,773 −43,016 −9,771 9,48 39,316 10805 1,098
EarningsForecastDisparity Intangibles 5,295 3,518 9,399 0,563 35,356 −14,629 −3,599 4,209 10,182 29980 468
Leverage q Profitability 5,283 2,342 16,651 0,317 77,886 −41,682 −5,998 7,953 17,376 24317 654
NetPayoutYield q Value vs, growth 5,264 2,148 18,053 0,292 64,716 −30,815 −7,553 8,439 36,091 24411 651
OptionVolume1 Trading frictions 5,24 2,021 12,92 0,406 55,349 −23,404 −4,069 5,505 19,543 35153 298
NumEarnIncrease Momentum 5,225 9,206 4,286 1,219 19,702 −6,137 −1,62 2,415 5,178 23407 684
ForecastDispersion Intangibles 5,198 2,273 15,323 0,339 59,378 −25,571 −6,72 6,97 13,746 27817 539
CBOperProfLagAT Profitability 5,176 3,218 12,249 0,423 42,409 −27,702 −5,954 5,064 13,288 23042 696
hire Intangibles 5,168 5,322 7,234 0,714 27,421 −9,683 −2,733 4,139 11,145 23953 666
zerotradeAlt12 Trading frictions 5,162 3,39 14,564 0,354 46,51 −21,01 −5,145 6,463 58,4 10805 1,098
AM Value vs, growth 5,139 3,016 14,205 0,362 70,638 −28,135 −5,176 7,341 20,865 18840 834
ShareIss1Y Investment 5,085 6,228 7,81 0,651 25,524 −13,625 −2,875 4,232 10,95 10805 1,098
Accruals Investment 5,022 6,651 6,25 0,804 20,468 −8,761 −2,194 3,471 7,165 19206 822
MRreversal Other 4,992 3,137 15,222 0,328 67,114 −13,021 −4,228 5,373 61,602 10805 1,098
EquityDuration Value vs, growth 4,893 2,578 14,394 0,34 66,695 −16,086 −6,031 7,16 20,939 23223 690
OperProf Profitability 4,876 3,143 11,765 0,414 54,613 −25,482 −4,121 4,563 19,835 23223 690
PctAcc Investment 4,77 4,612 7,773 0,614 22,69 −8,329 −2,878 4,172 12,633 23589 678
OPLeverage Intangibles 4,744 3,818 10,357 0,458 39,75 −12,354 −4,142 5,035 20,604 18840 834
ShareIss5Y Investment 4,669 5,268 8,384 0,557 29,015 −8,052 −2,899 3,897 30,07 11535 1,074
DelCOA Investment 4,658 6,201 6,217 0,749 31,089 −6,524 −2,259 3,439 10,706 19206 822
IndIPO Intangibles 4,646 2,673 11,743 0,396 37,676 −18,905 −4,542 5,156 15,404 27544 548
IndMom Momentum 4,638 3,974 11,163 0,415 57,47 −27,742 −3,73 4,725 25,526 10805 1,098
IntanSP Value vs, growth 4,624 2,589 14,347 0,322 62,892 −13,204 −5,376 6,803 23,346 20667 774
Size Other 4,623 3,187 13,875 0,333 52,279 −10,973 −4,258 5,973 53,26 10805 1,098
OperProfRD Profitability 4,599 2,311 15,091 0,305 66,236 −14,319 −6,737 7,788 17,573 23223 690
OrgCap Intangibles 4,573 4,16 9,165 0,499 33,984 −10,615 −3,832 4,604 11,703 18840 834
ReturnSkew3F Trading frictions 4,55 7,829 5,559 0,818 26,474 −13,135 −1,583 2,654 10,253 10805 1,098
BrandInvest Intangibles 4,48 1,674 19,932 0,225 71,063 −24,679 −8,084 8,987 40,012 23953 666
DownRecomm Intangibles 4,477 6,1 3,819 1,172 5,991 −5,991 −1,061 2,077 7,208 34334 325
CF Value vs, growth 4,474 2,562 14,556 0,307 50,586 −36,293 −5,488 5,886 17,254 18840 834
CoskewACX Trading frictions 4,469 3,562 9,36 0,477 35,927 −11,753 −3,452 4,608 19,586 23223 668
betaNet Trading frictions 4,45 3,224 12,874 0,346 62,429 −12,213 −4,669 6,097 29,654 12450 1,044
grcapx3y Investment 4,45 5,477 6,626 0,672 23,556 −6,536 −2,484 3,636 13,852 19935 798
betaCC Trading frictions 4,449 3,292 12,605 0,353 61,686 −11,579 −4,446 5,502 29,33 12450 1,044
RIO Turnover Other 4,433 2,829 14,976 0,296 64,628 −20,468 −6,611 7,251 18,778 10805 1,096
AssetTurnover Other 4,373 4,31 8,397 0,521 41,306 −7,92 −3,595 4,324 10,475 19206 822
WW Q Other 4,345 1,46 21,063 0,206 77,356 −21,674 −7,394 10,386 42,262 25902 601
GP Profitability 4,333 3,549 10,178 0,426 55,799 −15,813 −4,124 5,19 12,305 18840 834
Illiquidity Trading frictions 4,309 3,464 11,897 0,362 38,768 −11,582 −4,826 6,088 33,749 10805 1,098
tang Intangibles 4,304 3,219 11,188 0,385 37,32 −12,065 −4,163 4,874 38,744 18659 840
BetaTailRisk Trading frictions 4,29 2,709 14,942 0,287 44,177 −19,12 −5,867 6,466 37,067 11718 1,068
Tax Profitability 4,278 5,236 6,812 0,628 34,392 −16,421 −2,321 3,227 11,11 18840 834
ResidualMomentum6m Momentum 4,25 4,001 10,132 0,419 37,327 −23,253 −4,004 4,079 13,009 10989 1,092
ChInvIA Investment 4,207 6,234 5,586 0,753 29,42 −6,255 −1,936 2,868 10,375 19206 822
OrderBacklogChg Investment 4,194 2,567 11,378 0,369 44,392 −18,4 −4,45 5,326 13,777 26511 582
pchsaleinv Other 4,176 7,785 4,456 0,937 32,391 −3,709 −1,716 2,496 5,249 19024 828
sgr q Other 4,117 3,15 9,98 0,412 36,258 −20,083 −4,513 4,033 12,006 22917 700
grcapx Investment 4,065 5,625 5,937 0,685 22,295 −5,778 −2,317 3,218 8,255 19571 810
DivInit Value vs, growth 4,054 2,99 12,97 0,313 49,172 −43,755 −3,727 5,101 34,573 10805 1,098
DelNetFin Investment 4,046 6,924 4,837 0,837 21,473 −7,375 −1,84 2,489 4,947 19206 822
UpRecomm Intangibles 4,039 5,409 3,886 1,039 8,024 −6,836 −1,074 2,11 4,534 34334 325
EarnSupBig Momentum 3,925 3,082 9,664 0,406 32,411 −17,367 −4,074 4,276 12,992 23190 691
Beta Trading frictions 3,891 1,384 26,887 0,145 91,975 −25,643 −10,529 12,209 66,2 10805 1,098
Leverage Profitability 3,853 2,299 13,969 0,276 75,307 −33,085 −4,922 7,008 20,942 18840 834
PctTotAcc Investment 3,848 4,09 5,364 0,717 20,51 −7,088 −1,952 2,906 5,433 32353 390
OptionVolume2 Trading frictions 3,847 1,754 10,913 0,353 24,647 −9,058 −2,761 2,686 36,023 35185 297
CompEquIss Investment 3,805 3,068 11,763 0,323 57,38 −19,376 −3,639 3,766 42,241 11354 1,08
DivSeason Value vs, growth 3,706 14,429 2,457 1,508 6,476 −3,414 −0,813 1,424 4,68 10805 1,098
CashProd Intangibles 3,638 2,572 11,832 0,307 58,144 −27,924 −4,558 6,021 15,122 18659 840
BetaBDLeverage Trading frictions 3,587 1,969 12,555 0,286 49,749 −15,154 −5,209 6,087 17,786 26876 570
IntanCFP Value vs, growth 3,585 2,499 11,567 0,31 44,67 −27,887 −4,267 5,622 24,293 20485 780
DebtIssuance Investment 3,577 5,702 4,391 0,815 13,197 −4,961 −1,609 2,145 9,098 26329 588
MomOffSeason16YrPlus Other 3,553 3,358 9,29 0,382 40,307 −9,647 −3,904 4,107 22,418 16071 925
TotalAccruals Investment 3,551 3,563 8,247 0,431 43,768 −7,858 −2,547 3,703 16,382 19206 822
cfp Value vs, growth 3,523 1,763 15,02 0,235 70,311 −32,368 −6,223 6,093 16,724 23589 678
WW Other 3,51 2,124 13,726 0,256 61,703 −16,077 −4,858 6,29 31,135 19024 828
VolSD Trading frictions 3,475 2,355 14,113 0,246 39,969 −31,742 −5,457 6,035 43,474 10805 1,098
Spinoff Other 3,423 2,29 14,297 0,239 62,055 −20,934 −4,777 5,28 54,375 10805 1,098
sinAlgo Other 3,408 2,55 11,166 0,305 56,2 −15,83 −4,584 5,083 35,224 18718 838
VolMkt Trading frictions 3,405 1,807 18,03 0,189 80,691 −31,954 −7,799 8,763 21,094 10805 1,098
EP Value vs, growth 3,393 2,909 9,76 0,348 41,611 −15,944 −3,879 4,803 17,139 18659 840
BetaLiquidityPS Trading frictions 3,365 2,111 11,723 0,287 45,425 −13,297 −4,829 5,695 14,063 24138 649
AssetLiquidityBook Other 3,255 2,535 10,735 0,303 42,858 −13,004 −3,874 4,672 33,696 18687 839
realestate Intangibles 3,246 2,257 10,221 0,318 36,575 −15,268 −4,227 4,959 11,786 25780 606
IntanEP Value vs, growth 3,242 2,781 9,365 0,346 40,047 −13,871 −3,601 4,412 16,625 20667 774
MomOffSeason11YrPlus Other 3,205 2,773 10,598 0,302 27,372 −9,702 −3,796 4,19 51,967 13515 1,009
EarningsConsistency Intangibles 3,118 3,803 6,735 0,463 36,496 −8,2 −3,047 3,239 6,818 19571 810
RDS Intangibles 3,099 2,55 8,377 0,37 29,631 −14,937 −2,957 4,063 11,682 26876 570
GrSaleToGrInv Intangibles 3,082 5,449 4,681 0,658 19,914 −4,589 −2,008 2,263 6,394 19206 822
IntrinsicValue Other 3,068 1,372 14,915 0,206 60,048 −24,942 −5,703 6,73 17,221 27971 534
PatentsRD Other 2,957 2,921 5,637 0,525 20,997 −4,651 −2,093 2,581 10,977 28335 372
AssetLiquidityBookQuart Other 2,953 1,412 15,684 0,188 66,599 −19,197 −7,17 6,565 38,509 23680 675
IntanBM Value vs, growth 2,913 1,705 12,614 0,231 44,075 −18,729 −4,872 6,138 24,469 24317 654
GrAdExp Intangibles 2,849 1,503 14,123 0,202 80,327 −25,982 −4,973 5,514 24,665 23953 666
ConvDebt Intangibles 2,831 4,26 5,561 0,509 20,422 −23,012 −1,979 2,489 5,7 18659 840
CompositeDebtIssuance Investment 2,813 5,327 4,241 0,663 16,535 −5,517 −1,81 2,18 3,866 20667 774
rd sale q Other 2,777 0,672 23,002 0,121 85,224 −17,849 −8,758 9,696 58,856 32904 372
GPlag Profitability 2,761 2,622 8,747 0,316 36,922 −8,688 −4,167 4,034 11,789 19024 828
RoE Profitability 2,745 2,319 9,13 0,301 48,518 −22,086 −3,388 4,249 14,631 22493 714
ChNNCOA Investment 2,716 5,35 4,203 0,646 13,783 −4,704 −1,723 2,279 5,299 19206 822
MeanRankRevGrowth Value vs, growth 2,7 3,063 7,026 0,384 33,692 −7,794 −2,818 3,491 9,186 21032 762
pchgm pchsale Other 2,654 4,191 5,26 0,505 17,225 −7,684 −2,483 2,545 4,86 19024 828
quick Investment 2,651 2,137 10,377 0,255 61,508 −17,187 −4,158 4,341 29,082 18659 840
PriceDelaySlope Trading frictions 2,634 2,641 9,524 0,277 25,608 −11,884 −3,776 4,096 29,902 10805 1,094
CapTurnover Other 2,629 2,368 9,156 0,287 49,854 −10,745 −4,058 4,341 10,646 19389 816
ChForecastAccrual Intangibles 2,607 3,914 4,445 0,587 10,817 −5,838 −1,744 2,224 5,006 27971 534
EBM Value vs, growth 2,546 3,505 5,556 0,458 23,888 −5,635 −2,281 2,856 10,374 22858 702
DelCOL Investment 2,536 3,499 5,998 0,423 32,895 −6,617 −2,508 3,325 6,655 19206 822
PayoutYield Value vs, growth 2,535 2,044 10,188 0,249 55,369 −9,169 −4,373 5,087 16,252 19571 810
saleinv Other 2,481 2,989 6,943 0,357 28,359 −11,825 −3,104 3,137 6,522 18659 840
GrGMToGrSales Intangibles 2,443 3,806 5,312 0,46 19,093 −5,611 −2,385 2,712 5,316 19206 822
GrLTNOA Investment 2,38 3,168 6,218 0,383 21,305 −8,99 −2,77 3,137 8,077 19206 822
CitationsRD Other 2,369 2,182 6,332 0,374 32,343 −8,176 −2,743 3,063 9,089 28335 408
currat Value vs, growth 2,3 2,05 9,385 0,245 50,149 −14,905 −3,68 4,236 26,14 18659 840
FailureProbability Other 2,237 0,756 21,551 0,104 82,379 −41,921 −9,2 8,86 30,412 24868 636
OperProfLag Profitability 2,222 1,491 11,251 0,197 55,295 −27,828 −4,454 4,514 19,342 23407 684
salerec Other 2,156 2,573 7,012 0,308 44,771 −6,142 −2,974 3,257 11,471 18659 840
AnalystValue Intangibles 2,143 1,044 13,697 0,156 67,921 −24,318 −5,662 5,507 23,206 27971 534
IdioVolAHT Trading frictions 2,134 0,857 23,831 0,09 96,593 −43,551 −11,273 9,862 35,077 10805 1,098
rd sale Other 2,112 0,923 18,933 0,112 85,143 −18,465 −7,065 7,436 61,738 19206 822
Investment Investment 2,108 1,918 9,03 0,233 54,311 −12,673 −3,677 4,048 26,882 19571 810
BPEBM Value vs, growth 2,024 2,723 5,685 0,356 33,062 −7,161 −2,355 2,713 14,38 22858 702
ShareRepurchase Investment 2,014 2,575 5,446 0,37 24,889 −8,318 −2,276 2,471 5,631 26511 582
AOP Intangibles 1,984 1,463 9,049 0,219 34,526 −9,994 −3,725 4,58 11,022 27971 534
B, Individual anomalies AbnormalAccruals Investment 1,976 1,868 7,367 0,268 50,373 −5,307 −2,964 3,411 15,244 26511 582
ChAssetTurnover Profitability 1,94 3,857 4,132 0,47 16,633 −6,321 −1,663 1,933 5,648 19571 810
DelLTI Investment 1,935 3,605 4,409 0,439 18,242 −4,58 −1,613 1,942 9,633 19571 810
pchdepr Investment 1,918 3,133 5,087 0,377 23,974 −5,578 −1,833 2,331 11,422 19024 828
EarningsSmoothness Other 1,915 1,442 10,283 0,186 44,334 −10,875 −4,432 4,712 18,51 22312 720
betaRR Trading frictions 1,869 0,831 20,971 0,089 90,344 −21,495 −8,518 9,39 55,245 12450 1,044
fgr5yrLag Intangibles 1,847 0,646 17,732 0,104 68,426 −27,447 −6,775 7,82 18,003 30162 462
ChNWC Profitability 1,814 4,3 3,492 0,52 24,108 −4,351 −1,324 1,851 6,097 19206 822
Activism1 Other 1,713 0,768 9,034 0,19 21,33 −8,014 −3,689 4,234 12,648 33177 197
ChPM Other 1,709 3,067 4,613 0,371 19,056 −4,448 −1,968 2,373 4,668 19206 822
BookLeverage Value vs, growth 1,559 1,203 10,806 0,144 61,967 −13,506 −4,096 4,229 28,613 18840 834
BrandCapital Intangibles 1,52 0,807 14,281 0,106 71,066 −21,931 −6,016 5,929 30,452 20299 690
GrSaleToGrReceivables Other 1,481 2,945 4,163 0,356 21,693 −4,096 −1,686 2,015 5,631 19206 822
ExclExp Intangibles 1,438 1,049 8,307 0,173 20,432 −34,727 −2,163 2,663 7,241 30526 440
DownsideBeta Trading frictions 1,317 0,604 20,859 0,063 78,091 −35,682 −9,21 9,044 36,892 10805 1,098
OperProfRDLagAT Profitability 1,189 0,756 13,065 0,091 72,754 −28,402 −6,199 4,781 12,783 19024 828
Herf Intangibles 1,188 1,508 6,594 0,18 51,465 −7,384 −2,49 2,819 18,758 18659 840
Coskewness Trading frictions 1,131 1,152 9,39 0,12 70,164 −21,274 −3,578 4,21 14,991 10805 1,098
IdioVolCAPM Trading frictions 1,057 0,373 27,1 0,039 99,137 −26,001 −9,6 11,392 82,801 10805 1,098
SurpriseRD Intangibles 1,044 1,412 6,136 0,17 49,918 −10,417 −2,36 2,65 16,462 19084 826
HerfBE Intangibles 1,026 1,138 7,495 0,137 52,971 −11,185 −2,725 2,831 28,812 18993 829
pchquick Investment 1,021 1,752 4,825 0,212 56,445 −11,42 −2,143 2,045 3,456 19206 822
DelayNonAcct Other 0,936 0,696 8,732 0,107 38,014 −9,843 −4,08 4,032 10,909 28702 506
roavol Other 0,892 0,311 20,792 0,043 86,569 −21,953 −8,471 7,88 39,757 25017 631
Tax q Profitability 0,871 0,908 7,389 0,118 65,873 −11,265 −2,793 2,401 32,551 22553 712
KZ Intangibles 0,843 0,618 10,444 0,081 63,554 −17,243 −4,576 4,615 13,567 22858 702
salecash Other 0,828 0,71 9,756 0,085 63,509 −25,1 −4,047 3,899 14,508 18659 840
FRbook Intangibles 0,727 0,667 6,882 0,106 47,642 −10,08 −3,2 3,085 7,885 29644 479
OrderBacklog Intangibles 0,715 0,601 8,369 0,085 53,035 −9,697 −3,716 3,514 15,79 26144 594
AccrualQuality Investment 0,702 0,388 14,464 0,049 65,655 −15,428 −5,99 6,701 31,358 20851 768
EarningsValueRelevance Other 0,683 1,154 4,583 0,149 29,746 −6,322 −2,007 2,089 4,853 22312 720
AccrualQualityJune Investment 0,646 0,358 14,395 0,045 68,988 −14,969 −5,733 6,809 30,395 21032 762
EarningsTimeliness Other 0,559 0,904 4,823 0,116 36,643 −6,167 −1,969 2,269 10,844 21944 732
DelSTI Investment 0,542 0,848 4,538 0,119 40,103 −3,884 −1,682 2,164 8,19 25780 606
PriceDelayTstat Trading frictions 0,464 0,588 7,526 0,062 60,189 −15,351 −3,374 3,116 17,205 10805 1,094
HerfAsset Intangibles 0,454 0,497 7,584 0,06 61,889 −11,207 −2,654 3,121 30,242 18993 829
roic Profitability 0,333 0,161 15,832 0,021 75,953 −36,024 −7,025 5,713 18,188 22858 702
BetaFP Trading frictions 0,271 0,079 32,808 0,008 99,634 −27,002 −13,184 14,528 83,153 10805 1,098
RetNOA Profitability 0,141 0,152 7,065 0,02 47,799 −10,833 −2,981 2,833 17,562 23223 690
RDAbility Other 0,127 0,079 12,805 0,01 73,118 −12,956 −6,043 6,068 14,978 21032 762
pchcurrat Investment 0,109 0,192 4,697 0,023 59,192 −11,203 −2,183 1,84 3,711 19206 822
PredictedFE Intangibles 0,009 0,004 13,56 0,001 71,065 −21,731 −6,157 6,248 13,841 30526 450
LaborforceEfficiency Other −0,013 −0,025 4,346 −0,003 44,112 −5,156 −2,043 2,087 5,006 19206 822
EarningsConservatism Other −0,046 −0,091 3,945 −0,012 24,52 −7,691 −1,775 1,762 4,18 21944 732
securedind Intangibles −0,052 −0,054 6,096 −0,009 51,711 −7,257 −2,212 2,146 13,593 29616 480
High52 Momentum −0,058 −0,023 23,967 −0,002 99,701 −69,538 −9,786 7,562 18,55 10805 1,098
GrSaleToGrOverhead Intangibles −0,104 −0,152 5,696 −0,018 50,442 −12,398 −2,472 2,327 9,85 19206 822
ZScore Profitability −0,12 −0,055 16,524 −0,007 90,674 −19,919 −6,548 7,211 32,341 23042 696
ForecastDispersionLT Intangibles −0,21 −0,125 10,478 −0,02 64,888 −16,82 −4,31 3,542 17,539 29980 468
cashdebt Other −0,247 −0,14 14,696 −0,017 81,409 −33,187 −6,402 5,038 16,617 19024 828
BetaDimson Trading frictions −0,325 −0,209 14,846 −0,022 87,463 −20,169 −6,118 5,903 47,204 10805 1,098
PM Profitability −0,338 −0,264 10,657 −0,032 77,029 −22,804 −4,291 4,816 16,95 18840 834
FailureProbabilityJune Other −0,584 −0,204 20,97 −0,028 91,037 −19,76 −8,391 9,824 45,198 24684 642
ETR Other −0,609 −0,8 5,908 −0,103 52,348 −13,621 −2,026 1,763 11,243 22126 723
FirmAge Other −0,627 −0,766 7,831 −0,08 69,137 −17,35 −3,538 3,366 16,92 10805 1,098
nanalyst Other −0,634 −0,483 8,795 −0,072 58,317 −14,03 −4,092 3,846 10,109 27817 539
FR Intangibles −0,76 −0,397 12,033 −0,063 73,811 −20,101 −5,593 4,64 13,236 29798 474
secured Intangibles −0,789 −0,859 5,702 −0,138 47,464 −7,671 −2,361 2,406 7,322 30162 462
BookLeverageQuarterly Value vs, growth −0,832 −0,444 13,856 −0,06 75,59 −31,032 −5,928 6,171 15,706 24317 654
betaRC Trading frictions −1,017 −0,639 14,838 −0,069 89,604 −36,464 −6,683 6,306 15,996 12450 1,044
DivYieldAnn Value vs, growth −1,306 −1,287 9,71 −0,135 91,044 −35,937 −3,336 3,135 7,769 10805 1,098
Governance Other −1,528 −0,583 10,627 −0,144 62,263 −11,517 −4,734 4,404 12,113 33177 197
betaCR Trading frictions −1,587 −1,925 7,686 −0,206 81,93 −16,951 −3,488 2,829 9,887 12450 1,044
DelayAcct Other −1,616 −0,971 10,809 −0,15 69,412 −15,681 −4,882 4,873 12,614 28702 506
EarningsPersistence Other −1,665 −1,688 7,641 −0,218 76,584 −11,517 −3,704 3,353 8,554 22312 720
grcapx1y Investment −1,875 −2,925 5,285 −0,355 83,33 −7,927 −2,415 2,054 13,186 19389 816
ReturnSkewQF Trading frictions −2,863 −4,357 4,784 −0,598 80,491 −8,962 −2,293 1,52 10,119 24531 636
fgr5yrNoLag Intangibles −2,901 −0,955 18,958 −0,153 88,586 −25,425 −8,817 7,02 28,928 29980 468
ZScore q Profitability −3,014 −1,176 17,989 −0,168 95,687 −29,248 −8,667 6,516 21,465 26235 591
IdioVolQF Trading frictions −3,189 −0,98 23,695 −0,135 97,752 −25,702 −9,634 10,463 52,355 24531 636
BetaSquared Trading frictions −3,4 −1,223 26,596 −0,128 99,935 −66,2 −12,162 10,505 25,817 10805 1,098
AbnormalAccrualsPercent Investment −4,004 −7,25 3,866 −1,036 86,55 −5,818 −2,14 1,333 3,86 26329 588
ChNCOL Investment −4,4 −5,218 7,005 −0,628 97 −12,22 −3,453 2,57 5,754 19024 828
ReturnSkewCAPM Trading frictions −4,921 −7,048 6,679 −0,737 99,217 −11,281 −2,843 1,844 21,234 10805 1,098
sgr Other −5,431 −5,847 7,716 −0,704 98,533 −14,8 −4,258 2,937 7,272 19024 828
VarCF Other −5,451 −2,71 16,525 −0,33 99,479 −30,941 −7,825 6,407 14,029 19571 810
EarningsPredictability Other −6,669 −4,644 11,123 −0,6 99,146 −10,177 −5,641 4,293 20,682 22312 720
AssetGrowth q Investment −7,125 −4,104 12,933 −0,551 99,14 −22,872 −7,349 4,458 12,747 23953 666
ChNCOA Investment −8,536 −8,978 7,898 −1,081 99,82 −15,44 −4,346 2,315 5,331 19024 828
NetDebtPrice q Value vs, growth −9,01 −4,434 14,371 −0,627 99,695 −21,181 −6,65 4,8 23,187 25902 600
KZ q Intangibles −9,538 −3,017 20,075 −0,475 99,661 −36,364 −6,295 4,441 68,723 26417 484
OScore q Profitability −10,965 −3,899 17,106 −0,641 99,422 −13,951 −7,45 8,042 22,967 30712 444
EntMult q Value vs, growth −16,156 −9,121 12,955 −1,247 99,994 −30,493 −6,224 3,222 18,873 24590 642
1 Like
29

Again a bit off-topic, but I did it within a minute or two, just eyeballing the different factors. It’s just my personal input, nothing statistical about it. I think that if you equal weight them, it should not make much of a difference.

30

For those interested in ChatGPT still:

I was talking colleague about ChatGPT and she casually said that she had paper where she asked ChatGPT and other AIs to do abstracts on research papers (probably in ophthalmology and not published yet).

I said; “Yea, I bet it does pretty good.”

Her response was a little more consistent with what Yuval has said: “Yea, but it makes some obvious mistakes.” And inferred it would be obvious and interesting in her paper. She did not have time to expand as she was busy getting authorization or permission to treat a patient thru a a different AI system at the time.

Yuval has suggested that ChatGPT 4 (versus the free 3.5) may be worth the $20 per month (my only evidence of this being fact act he or his employer was paying the $20 per month when he posted)

Here is the improvement you get for that $20 dollars per month. Notice, ChatGPT programs better than I can but its programming skills are actually weak compared to highly trained humans (i.e., < fifth percentile. The last entry in the table below).

This may be more about ChatGPT that many are interest in but it does do a pretty good job for those who want to start using some Python programming for P123 data. And maybe for some research as Yuval has alluded to (I leave it to him to expand on how he has used it):

Screen Shot 2023-04-29 at 8.32.28 AM
Screen Shot 2023-04-29 at 8.32.28 AM1752×1246 79.8 KB

Jim

31

I’m still very frustrated with ChatGPT. It’s constantly making mistakes. Here are some excerpts from my latest conversation with it:

me: given two points (a,b) and (c,d) on an exponential curve, what is the formula for the curve?

chat: [blah blah blah] Therefore, the formula for the exponential curve passing through the points (a,b) and (c,d) is:

f(x) = (b / e^(ba)) * e^((ln(d/b) / (c-a)) * x)

me: can you give me the formula without e and without a natural logarithm, please?

chat: Unfortunately, it is not possible to express the formula for an exponential curve without using the exponential function or natural logarithm. [blah blah blah]

me: Can’t an exponential function take the form of y = a*b^x? If you use that, you don’t need e or logarithms

chat: Yes, you are correct. An exponential function can also be expressed in the form y = a * b^x, where “a” and “b” are constants, and “b” is the base of the exponential function.

To find the equation of the exponential curve passing through two given points (a,b) and (c,d), we can use the point-slope form of a linear equation. If we take the natural logarithm of both sides of the exponential function y = a * b^x, we get: [blah blah blah] y = e^(ln(c) - k * ln(d)) * (e^((ln(b) - ln(d)) / (ln(c) - ln(a)))^x)

me: you can do this without logarithms. start with the equations for the two points and then divide one by the other.

chat: You are correct. Here is another method to find the equation of the exponential curve passing through two points without using logarithms:

Let the two points be (a,b) and (c,d), where b and d are positive. Then we have:

b = a * r^k
d = c * r^k

where r is the common ratio of the exponential sequence and k is the number of steps from (a,b) to (c,d), i.e., k = (ln(d) - ln(b)) / (ln(r)).[blah blah blah] Therefore, the equation of the exponential curve passing through the two points (a,b) and (c,d) is:

y = a * (d / c)^(k/(k+1)) * r^x

where k = (ln(d) - ln(b)) / (ln(r)) and r = [(d / b) * (a / c)]^(1/k).

me: no, that’s incorrect. let’s start with two points (a,b) and (c,d). the equation is y = n*(m^x). so b = n*(m^a) and d = n*(m^c). So b/d = m^a/m^c = m^(a-c). we can then solve for m: m = (b/d)^(1/(a-c)). We can then solve for n with one of the points. n = b/(m^a) = b/((b/d^(1/(a-c))). that can be simplified somewhat. we then plug these new equivalents for n and m into the original equation y = n*(m^x) and simplify that. try it and see what you come up with.

chat: You are correct. I apologize for my mistake in my previous response. Here is the correct method:

Given two points (a,b) and (c,d) on an exponential curve, the equation for the curve can be expressed as y = n * (m^x), where “n” and “m” are constants. [blah blah blah] Therefore, the equation of the exponential curve passing through the two points (a,b) and (c,d) is:

y = (d^((a-c)/a)) * (b^((c-a)/a)) * ((b/d)^(1/(a-c)))^x

This can be simplified as:

y = b * (d/b)^(x/(c-a))

me: beautiful! can you take me through the last simplification?

chat: [not going to quote it in full but the derivation would have gotten a c in any algebra class]

Maybe they shouldn’t call it artificial intelligence. Artificial slightly-intelligent-but-not-really-very-bright, perhaps? Why couldn’t it have come up with y = b * (d/b)^(x/(c-a)) the first time around? It’s relatively easy to derive and far clearer and simpler than what it came up with the first three times. And that thing with r and k was absolutely bizarre.

32

Yes I agree. It seems to be guessing too much.

I asked the find the 5 most cited studies that show that the value factor is arbitrary away. I asked for an abstract and a writer.

This is the list I got:

  1. Harvey, C. R., Liu, Y., & Zhu, H. (2016). …and the cross-section of expected returns. Review of Financial Studies, 29(1), 5-68. Abstract: This paper examines the performance of value strategies in the US and international stock markets. The authors find that while value strategies have generated high returns in the past, they have been largely ineffective in recent years. The results suggest that the value premium has disappeared or become negative in recent decades.
  2. Fama, E. F., & French, K. R. (2015). A five-factor asset pricing model. Journal of Financial Economics, 116(1), 1-22. Abstract: This paper introduces a new asset pricing model that includes profitability and investment factors in addition to the market, size, and value factors. The authors find that the value factor is no longer significant after controlling for these additional factors, suggesting that the value premium has disappeared.
  3. Asness, C. S., Moskowitz, T. J., & Pedersen, L. H. (2013). Value and momentum everywhere. Journal of Finance, 68(3), 929-985. Abstract: This paper examines the performance of value and momentum strategies in different asset classes and across global markets. The authors find that while both value and momentum strategies have generated high returns historically, the performance of value strategies has weakened in recent years. The results suggest that the value premium may be disappearing.
  4. Novy-Marx, R. (2013). The other side of value: The gross profitability premium. Journal of Financial Economics, 108(1), 1-28. Abstract: This paper argues that profitability is a better predictor of stock returns than traditional value metrics such as book-to-market ratio. The author finds that a profitability factor has stronger explanatory power than the value factor and is less prone to crashes and other anomalies. The results suggest that the value premium may be replaced by a profitability premium.
  5. Fama, E. F., & French, K. R. (2012). Size, value, and momentum in international stock returns. Journal of Financial Economics, 105(3), 457-472. Abstract: This paper examines the performance of size, value, and momentum strategies in international stock markets. The authors find that the value factor is no longer significant in many international markets after controlling for size and momentum factors. The results suggest that the value premium has weakened or disappeared in some international markets.

Just as an example of what Yuval says, the latest study actually shows the exact opposite. I reacted because the study was from 2012, so why would studies already in 2012 find that the value factor did not work .

The study says:

Echoing earlier studies, we find value premiums in average returns in all four regions we examine (North America, Europe, Japan, and Asia Pacific), and there are strong momentum returns in all regions except Japan. Our new evidence centers on how international value and momentum returns vary with firm size.

1 Like
33

A fascinating exposition on ChatGPT/GPT-4;

34

Walter,

Thank you. Watching this now.

As far as drawing, Bing had the capability of drawing for a few days that seems to be gone now. But I drew several beautiful beach homes that I plan to send to an architect if my port does as good as my overfitted sim :worried:

ChatGPT 4 now does “plug-ins” that are on a wait-list apparently (I do not have it yet). But ChatGPT 4 is already tying into the internet and other resources already for some. Of particular interest to me (I singed up) is the Python capabilities will be upgraded with the plug-ins.

IMHO, you can trick it into making an error and feeling more intelligent and superior to ChatGPT but the general knowledge is incredible. I just asked it for the Bayesian updating formula and it was correct. I asked for the odds ratio version of it and got that too.

Better than most undergrads on that count at least.

Jim

35

Jim,

I occasionally use ChatGPT to write Python code. It’s a great help with finding libraries to use.

I’m waiting for someone to write “How to Code in Python using ChatGPT - for Dummies.”

If software will eat the world what happens when GPTs eat software?

1 Like
36

My daughter used Chegg when she was in college. Pretty good site if difficult to unsubscribe from (with continued monthly payments).

Wish I had thought of this as a possibility:

“Chegg withdrew its full-year guidance because of the surge in the popularity of ChatGPT. Shares tumbled as much as 43% in premarket trading. If premarket losses hold, this will be the most significant decline since Nov. 2, 2021.”

Ah yes. I missed that. But what about the long-term potential of Chegg? IDK. So I thought I might ask someone a little smarter than me (ChatGPT 4) this question: “Chegg is an online site to help college student with their college work. How do you think it will ba affected by ChatGPT 4?”

ChatGPT 4’s Answer which is probably true. Some sites will probably benefit long-term (BTW, I asked ChatGPT how easy and how costly it would be fore Chegg to integrate ChatGPT 4. The answer was not helpful and is not included):

" The introduction of advanced AI models like ChatGPT-4 could have both positive and negative implications for online educational platforms like Chegg. Here’s how Chegg and similar platforms might be affected:

Positive Impacts:

  1. Enhanced tutoring capabilities: ChatGPT-4 could be integrated into Chegg’s platform to provide immediate and personalized tutoring assistance to students. This could enhance the overall learning experience and help students gain a better understanding of the subject matter.
  2. Improved efficiency: ChatGPT-4 could help automate some of the more routine tasks, such as answering frequently asked questions, grading assignments, or providing basic explanations. This could free up human tutors to focus on more complex or nuanced questions that require deeper understanding or critical thinking.
  3. Scalability: AI models like ChatGPT-4 could allow Chegg to scale its tutoring services to accommodate more students simultaneously without the need for a significant increase in human tutors. This could lead to lower costs for the platform and increased accessibility for students.

Negative Impacts:

  1. Competition: As AI models like ChatGPT-4 become more widespread, other educational platforms may also integrate such technology, increasing competition for Chegg. To stay ahead, Chegg will need to differentiate itself through the quality of its services, user experience, and additional features.
  2. Quality concerns: While ChatGPT-4 is an advanced AI model, it may not always provide accurate or contextually appropriate answers. If students rely heavily on AI-generated responses, there’s a risk that they may receive misleading or incomplete information, which could negatively impact their learning experience.
  3. Job displacement: As AI models like ChatGPT-4 become more capable, there may be concerns about the potential displacement of human tutors. However, it’s essential to remember that AI can be a tool to augment human capabilities rather than entirely replace them. Human tutors will still be valuable for addressing complex questions, providing personalized support, and nurturing critical thinking skills in students.

In conclusion, the impact of ChatGPT-4 on platforms like Chegg will depend on how well they adapt to and integrate this technology into their services. By leveraging the strengths of AI models like ChatGPT-4 and combining them with the expertise of human tutors, platforms like Chegg can enhance their educational offerings and improve the learning experience for students."

37

It is nice for finding interesting ways to use Python and a good way to learn Python, because everything it tells you about Python is easily confirmed by trying the code it gives. ChatGPT reminds me of an old college friend that was very knowledgable but if he did not know the answer to something, he would make it up, and in very convincing fashion. As a result you could not believe anything he said. Google’s CEO talks about it. 'Hallucination problems' plague A.I. chatbots, Google CEO says | Fortune

1 Like