Earnings Revisions and Cup-with-Handle Breakouts A Rewarding Combination

(Revised 7/14/2009)

Summary The impact of analyst earnings revisions on the momentum of stoc prices has long !een recogni"ed#1 The effect of up$ards earnings revisions can !e felt for up to 12 months and one of the simplest profita!le trading strategies is to !uy $hen earnings estimates are revised up$ards and to sell $hen earnings are predicted to !e flat or falling# This simple strategy can !e significantly improved $hen the earnings revisions are filtered for stoc s that are !rea ing out from a cup%$ith%handle chart pattern# The result is a smaller num!er of stoc s suita!le for the private investor to manage and investment returns many times !etter than $ith the simple strategy# This paper discusses& earnings revision estimates and $hy trading them can !e profita!le' the cup%$ith%handle chart pattern and ho$ it $or s to provide timing of position entry points( and a simple portfolio management strategy that com!ines these t$o elements into a highly profita!le trading system# This is demonstrated using the )ac s Ran as a pro*y for earnings revisions estimates and cup%$ith%handle !rea outs from !rea out$atch#com# Results are presented from a simulation of trading this strategy from +anuary 1( 2004 to +une 1( 2009 $hich $ould have yielded an annuali"ed compound return of 97,# Earnings Revision Estimates -ro erage .ouses( /nvestment -an s and many other financial institutions employ financial analysts to monitor companies earnings prospects and ma e !uy( sell or hold recommendations# 0nalysts usually speciali"e in a fe$ companies or a mar et sector so their analysis can !e detailed and focused# /n the course of their $or ( analysts ma e pro1ections of the company2s earnings over the coming 3uarters using factors li e demand and supply( potential sales gro$th( economic climate( competitive position( management competence( cost controls and so on# These earnings estimates are then e*pressed as future estimated 4arnings 5er 6hare (456) and are one of the factors considered in the analyst2s !uy( sell or hold recommendation# 456 is considered an important factor in valuing a company2s stoc price# The ratio of the stoc price to the earnings per share (5/4 ratio) is fre3uently used as an indicator of $hether the stoc is under or over valued so the e*pectation of an improved earnings per share $ill usually lead to a rise in the stoc price# -ecause the stoc mar et attempts to anticipate future earnings( this up$ards revision can lead to an almost immediate up$ards impact on the share price( even if the earnings gains are not reali"ed# 7nce( the earnings revisions estimates $ere only availa!le to the $ealthy clients of the !ro erage houses( !ut these are no$ much more $idely availa!le through investor services such as Thompson Reuters and )ac s# This $ide and immediate dissemination of earnings revision estimates serves to put more up$ards pressure on stoc prices than $ould !e the case if it $ere more closely held# 8any computeri"ed valuation models use 456 in the their calculations so even slight revisions can have an immediate effect on share prices# 9o$n$ard revision estimates can have an even more po$erful effect on stoc prices( so a do$n$ard revision is an effective sell signal# /f a stoc has a history of up$ards revisions( a flat pro1ection can !e a disappointment to holders of the stoc so this can lead to a decline in the stoc price# The com!ination of !uy on up$ards revisions and sell on flat or do$n$ards revisions therefore provides a potentially re$arding trading strategy# 1 4arnings Revisions and 5ortfolio Returns( Scott Mixon, +ournal of /nvesting, :all 2001' 0head of the 8ar et( 8itch )ac s( .arper and ;ollins( 200<' 8ar et 4*pectations vs# Results( 00// +ournal( 0pril 2009
;opyright 2009 8ichael =i!!ons# 0ll Rights Reserved#

The typical private investor can only hold and trac a small num!er of stoc s# :or this reason( the 3uantity of earnings revisions can easily over$helm the private investor $ho $ishes to follo$ a trading strategy !ased on earnings revision estimates# This investor needs a $ay to filter the num!er of trading candidates to a managea!le num!er# The Cup-with-Handle Chart Pattern -rea outs from this chart pattern have !een sho$n to offer healthy returns in the $ee s follo$ing !rea out# To !e successful a !rea out must !e supported !y institutions $ho usually only invest in li3uid stoc s# >e therefore limit consideration to stoc s that have a significant float and are heavily traded# >e no$ descri!e the pattern and $hy it forms# Stage 1: Setup The pattern starts $ith a stoc rising from a former !ase# This could occur !ecause of improved performance e*pectations for the stoc # 0t some point( profit ta ing sets%in and the stoc !egins to decline( ending the setup# The ma*imum price reached at the end of 6tage 1 is called the 2?eft ;up2( and the amount !y $hich the stoc rose from its prior lo$ is the 6etup =ain# To ensure a $ell%defined left cup( $e re3uire that the setup gain !e at least <0, from the prior lo$ level for it to 3ualify as a valid left cup# Stage 2: Decline. The stoc no$ collapses into a ne$ !ase# The collapse can occur for any num!er of reasons& poor company results' over% all mar et conditions' !ad ne$s' :igure 1& ;up%$ith%.andle 6tages profit ta ing after a strong setup stage( and so on# There should !e strong selling and volume should !e $ell a!ove average for the first fe$ days of the collapse# @ote that collapsing price alone is not sufficient to provide a good ;$. pattern( volume is also important# This is !ecause $e $ant most of the !uyers from the setup stage to li3uidate their holdings( so there $ill not !e a lot of overhead supply $hen the stoc !egins to clim! the right side of the cup# The com!ined price and volume action in stage 2 is important if the eventual !rea out is to succeed# -y the end of stage 2( volume should have fallen to $ell !elo$ average levels as most holders have sold and there is little !uying activity( if any# The depth of the !ase is important also# >e do not $ant the stoc to collapse to a level from $hich recovery $ill !e difficult( if not impossi!le# 0 ma*imum decline of A0, is allo$a!le# Stage 3: Recovery. /f and $hen the conditions that !rought a!out the stage 2 collapse have !een resolved( the stoc may !egin to recover# /f it does so( the stoc $ill start to clim! the right side of the cup# 0s it does so( $e li e to see a!ove average volume on days $hen the price moves up( $hich indicates that institutions are ta ing an interest in the stoc ( !ut light volume on days $hen it closes do$n( indicating that there are small num!ers of sellers# /deally( this constructive price and volume action $ill strengthen as the right side of the cup is formed and the stoc moves higher# 8ean$hile( there $ill !e some holders of the stoc $ho !ought at or near the left cup price !ut didn2t sell in stage 2# These holders are $aiting for the time $hen they can recoup some or all of
;opyright 2009 8ichael =i!!ons# 0ll Rights Reserved#

their losses# There $ill also !e profit ta ing !y !ottom%fishers# 0s losses are covered or profits are ta en there $ill !e pauses in the recovery and a typical cup right%side $ill e*hi!it a stair%step characteristic( rather than the smooth ascension sho$n in this ideali"ed chart# 4ach of these pauses( or pull!ac s( reduces the overhead supply left over from the left cup setup# 0s the price on the right side approaches the left cup level( the last holders $ill finally decide to cut their losses and there $ill !e a large volume sell%off# This often is preceded !y a day on $hich the price spi es on high volume $hich the sellers have interpreted as an over!ought condition and therefore a last opportunity to recoup their losses# This is the point at $hich the pivot forms( and mar s the end of the recovery stage# :or the eventual !rea out to !e successful( the pivot should !e 1ust !elo$ the left cup so there is little overhead supply# Stage 4: Consolidation. 0fter the pivot forms $e $ant and e*pect to see a sha eout $hile the overhead supply is depleted# This $ill cause the price to decline( initially on high volume !ut then the price should sta!ili"e or drift do$n on volume $ell !elo$ average levels# /f institutions are trac ing the stoc ( they no$ that they $ant the overhead supply eliminated and $ill $ait for these sta!le conditions to materiali"e !efore they pounce# Stage 5: Breakout. -rea out occurs $hen the stoc price rises a!ove the pivot and the daily volume is significantly higher than the B0%day average volume# .igher than average volume indicates that institutions( $ho have the a!ility to !uy in volume( are !idding the stoc price higher so this is an opportunity for the individual investor to piggy%!ac on the strength of institutional !uying# >e consider a B0, increase in volume as an indicator of institutional support for the !rea out# 0s a stoc is !rea ing out( it is desira!le to !uy as close to the pivot price as possi!le so $e monitor the intraday price and volume action of cup%$ith%handle stoc s and !uy $hen the pivot price is reached and the pro1ected volume to the end of the day is B0, higher than average# 0t this intraday point $e consider the stoc to !e ma ing an attempted breakout (!rea out$atch#com issues a !rea out alert at this point)# /f the stoc closes at or a!ove the pivot and volume for the day $as B0, higher than average( then $e no$ consider the stoc to !e a confirmed breakout# The cup%$ith%handle has !een recogni"ed as one of the most relia!le !rea out patterns# /t $as !rought to prominence in .o$ to 8a e 8oney in 6toc s !y >illiam 72@eil and included in Thomas -ul o$s i2s 4ncyclopedia of ;hart 5atterns# !eveloping a Trading Strategy -efore descri!ing ho$ $e com!ine earnings revisions and cup%$ith%handle !rea outs( lets consider the elements of a successful trading system# 1# /t goes almost $ithout saying that $e should have an e*pectation that the system $ill !e profita!le# The discussion a!ove provides a !asis for this e*pectation# >e $ill see that it is confirmed later# 2# The system should have a small num!er of rules to avoid comple*ity and !e easy to implement for the private investor# <# The system should have ro!ust parameter values suita!le for different mar et conditions# 4# The system should allo$ trading multiple simultaneous positions( !ut not re3uire so many that implementation $ill !e unmanagea!le for the private investor# B# The system must use ris control( money management and portfolio design# A# /t should !e purely mechanical# This removes su!1ectivity and emotion from trading decisions# The first element of our strategy is to find earnings revisions# )ac s /nvestment Research has !een revie$ing analysts earnings estimates since 197C and has developed a 3uantitative model that uses four factors related to earnings estimates to classify stoc s into five groups( ran ed from 1 to B representing D6trong -uyD to D6trong
;opyright 2009 8ichael =i!!ons# 0ll Rights Reserved#

6ellD# The first element of our system $ill !e to consider only stoc s that have a strong !uy recommendation# 0t any time there can !e around t$o hundred stoc s having the num!er one ran so $e need to filter these into a managea!le 3uantity# >e also $ant to select only those stoc s that have the !est prospects for su!stantial upside moves# 0s $e sa$ a!ove( the cup%$ith%handle chart pattern provides a $ay for us to select stoc s $ith upside potential( and a !rea out from that pattern allo$s us to optimi"e our entry point# The second element of our strategy( therefore( is to filter our list of earnings revision stoc s for those that are in a cup%$ith%handle pattern# The resulting list of stoc s then forms our $atchlist# >e add to this list daily as ne$ stoc s ran ed num!er one !y )ac s are announced and remove those no longer ran ed num!er one# >e no$ $ait for stoc s on our $atchlist to !rea out# >e have t$o $ays of recogni"ing a !rea out& 1# >e monitor the stoc and !uy it immediately it ma es a !rea out attempt# 2# /f intraday monitoring is not possi!le $e !uy on a confirmed !rea out# /n this case( our earliest opportunity to !uy is at the open on the ne*t trading day# >e can also impose additional constraints on our $atchlist# /n our study $e only considered stoc s $ith a minimum !rea out price of EA( an average daily volume of 100(000 and stoc s $ith an R6 Ran of at least C0#2 0dditionally( $e may choose to set a limit price a!ove $hich $e $ill not !uy the stoc # This covers us against a gap%up a!ove the !rea out price or an open price that is too high# >e can also set a price !elo$ $hich $e $ill not !uy the stoc if the ne*t session2s open is !elo$ the !rea out price# >e must also decide the amount $e $ill invest in a ne$ position# 0 simple !ut effective approach is as follo$s& 1# 9ecide ho$ many positions you can effectively trac at once# ?ets say five is a reasona!le num!er# >e2ll call these portfolio slots# 2# >hen ma ing your first trade( put 1/B of your availa!le capital into the position# Fou no$ have 4 slots open# :or your ne*t trade put G of the availa!le capital into the position# Repeat until you have filled all your slots# This approach spreads our ris over a num!er of stoc s !ut $e have found that it also reduces our profits !ecause it often leaves us $ith less than 100, of our availa!le capital invested most of the time# /n fact( $e found that the optimum num!er of positions $as at the most 2# 7ur ne*t decision is $hen to sell the stoc # 1# /f $e !ought on a !rea out attempt !ut the !rea out $as not confirmed $e can choose to sell at the ne*t open or eep the stoc until the ne*t sell signal# >e $ill sho$ later that the !est option is to eep the stoc even if the !rea out is not confirmed# 2# The )ac s ran is revised daily after mar et close# >e monitor ran revisions for stoc s $e have !ought and if a stoc loses its num!er one ran $e sell it at the ne*t open# <# >e place an automatic stop loss order immediately after completing our purchase# 6ee Ris 8anagement# Risk "anagement To protect ourselves from e*cessive losses( $e set a stop loss price and sell if the stoc falls to that price( even if it is still a num!er one ran >e can set the stop loss level according to our tolerance to ris # /f $e set it too tight $e $ill !e stopped out fre3uently( possi!ly missing a su!se3uent recovery in the price' set it too loose and our dra$do$n $ill increase# Through !ac testing $e can determine the optimum stop loss percentage# Port#olio "anagement >hen you sell( add the proceeds !ac to your capital# /f you made a profit( your ne*t position investment $ill !e !igger than the previous one# This allo$s you to !enefit from compounding of profits# /f you made a loss( your
2 The R6 Ran is a measure of a stoc 2s price performance over the last four 3uarters relative to all other stoc s# The most recent 3uarter is $eighted more heavily than the previous three# 6toc s are ran ed 1 to 99# ;opyright 2009 8ichael =i!!ons# 0ll Rights Reserved#

ne*t position si"e $ill !e smaller# This limits the damage from successive losses# Ba$ktesting the Strategy >e $ere a!le to !ac test the strategy from +anuary 1( 2004 until +une 1( 2009# >e had access to )ac s num!er one ran ed stoc s as of each :riday2s close and $e also had a history of attempted and confirmed !rea outs from the cup%$ith%handle pattern from !rea out$atch#com# To simulate running the strategy $e& 1# ?isted the stoc s ran ed num!er 1 !y )ac s for each $ee # 2# :ound $hich of those stoc s made a !rea out attempt during the follo$ing $ee and met three conditions& price $as at least EA( average daily volume $as at least 100(000 and R6 Ran $as at least C0# <# -ought the stoc under each of our !uy options& either $hen a !rea out alert $as issued at the alert price or at the open the ne*t day if the !rea out $as confirmed# 4# 6old the stoc under each of our sell options& /f $e !ought on !rea out attempt $hich $as not confirmed then $e compared results for selling at the ne*t open or eeping the position open# 0s soon as the stoc $as no longer ran ed num!er 1 $e sold at the ne*t open# /f the stoc price fell to the stop loss price intraday !efore losing the num!er 1 ran it $as sold at the stop loss price B# This gave us three !asic strategies to evaluate& 1# -uy on confirmed !rea out at ne*t open 2# -uy on !rea out and sell if !rea out not confirmed <# -uy on -rea out and hold if !rea out not confirmed >e ran each strategy $ith com!inations of stop loss and position limits in the ranges& stop loss& 0, to 10, positions& 1 to 20 A# :or the !uy at open option( $e allo$ed the stoc to !e !ought if the open price $as up to B, a!ove the !rea out price to allo$ for strong price momentum on !rea out# >e did not !uy the stoc if the open price $as !elo$ the !rea out price# Ba$ktest Results >e present here the results for the com!inations of position limits and stop loss that gave the !est returns and also those gave the !est return $hen ad1usted for ris # >e used ma*imum dra$do$n as the ris factor# Results Summary These are the results for running the strategy continuously from +anuary( 2000 to +une 1( 2009#

;opyright 2009 8ichael =i!!ons# 0ll Rights Reserved#

Strategy 1# -uy on ;onfirmed -rea out 1#1 -est 7verall Return 1#2 -est Re$ard/Ris 2# -uy on 0lert( 6ell if Hnconfirmed 2#1 -est 7verall Return 2#2 -est Re$ard/Ris <# -uy on 0lert( Ieep if unconfirmed <#1 -est 7verall Return <#2 -est Re$ard/Ris

Compound Ann Return % 27#C 17#7 A1#4 A1#4 A7#A B9#<7

"a& !rawdown % <7#B 11#B <0#4 <0#4 <1#2 17#4

Stop 'oss % C < 2 2 < <

Positions

1 1 1 1 1 2

@ote that the !est returns are mostly $hen you allocate all your portfolio to a single position# This is !ecause more of your money is invested at any one time( $hereas $ith t$o or more positions there $ill !e times $hen you are invested only partially and you are not ma*imi"ing your return# :or any trading (or !etting) system( the participant must have the temperament (and resources) to $ithstand a succession of losses and the potential for a large dra$do$n# :or this reason( $e personally prefer scenario num!er <#2 $ith the <, stop loss and 2 position option# .o$ever( one can argue that the large dra$do$n $as e*ceptional in all cases due to the !iggest mar et decline since the 19<02s and such large dra$do$ns are unli ely to !e repeated in the ne*t fe$ years# .ere are the detailed results for option <#2# 0 compound annual return of almost A0, over a period $hen the 6J5 B00 lost over <, compounded is an e*ceptionally attractive result# Fou can verify these results and run other scenarios $ith the )6- -ac test Tool( $hich is freely availa!le to all at $$$#!rea out$atch#com#

;opyright 2009 8ichael =i!!ons# 0ll Rights Reserved#

Buy on Breakout Alert Annual Results >e no$ loo at the annual returns $hen trading the !uy on !rea out alert strategy for each calendar year since 2004#The results are !ased on an initial capital of E10(000# Ta!le 1 sho$s detailed results for running the strategy over one calendar year starting $ith the first mar et session for the year and closing out all positions on the last day of the year# @ote that the strategy provided e*ceptional returns during the !ull mar et from 2004 to 2007 and only slipped very slightly in 200C $hile the 6J5 B00 lost more than 10 times that of the strategy#

;opyright 2009 8ichael =i!!ons# 0ll Rights Reserved#

())* ())+ ()), 5ortfolio 5ositions& 2' 6top ?oss& <,' starting capital& E10(000' -uy ?imit& B, :inal 5ortfolio Kalue E11(74C E1A(970 E20(<92 5rofit , 17#B, A9#7, 10<#9, 0vg# premium paid over 0#7C , 0#AB , 0#7C , !rea out price< 8a*imum 5ortfolio Kalue E12(C49 E17(A91 E20(C90 8a*imum =ain , 2C#B, 7A#9, 10C#9, 4 ;ompound 0nnual Return , 17#7 , A9#9 , 104#< ,
6J5 B00 ;ompound 0nnual Return 10#C2, 4#79, 1B#74,

())E2<(C19 1<C#2, 0#A2 , E24(C94 14C#9, 1<C#C ,

()). E9(A7B %<#<, 0#74 , E10(72B 7#<, %<#< , L<7#22, E1(74C 1A#< ,

8a*imum 9ra$do$n 8a*imum 9ra$do$n ,

E1(79B 1A#2,

E<(0B2 17#A ,

E<(29C 19#A ,

E1(9B2 1A#2 ,

Ta!le 1& Results for -uy on -rea out 0lert 6trategy

< 0lert prices are usually a!ove !rea out prices due to the volume re3uirement !efore an alert is issued 4 ;ompound annual returns are slightly different from total returns due to the start and end dates of the trading year ;opyright 2009 8ichael =i!!ons# 0ll Rights Reserved#