be a trend setter

be a trend setter

trendsetter

image courtesy : http://cheezburger.com/1276494592 

apologies for not agreeing to what’s written today in two of the favorite blogs that we follow

Why Markets Trend by @ReformedBroker    & 

Is The Market Always Right? by @ivanhoff

, but here is my thought(stat)ful disagreement

I’m assuming  the definition for “trend”  is when you place the bet based on the current day’s outcome on $SPX ( i.e to go long when the current session on various time frames is up)

ps: the data used is from 1st Jan 1990 till date. ( 6th Oct 2013)

Daily Trends

below the trading odds for buying $SPX at close and exiting at next day’s close

if $SPX is up on the current day

  • Winners : 1669
  • Losers : 1525
  • % Winners : 52%
  • Average Change % * 100  : 0.95 ( to get a sensible number rather a two decimal rounded off number like 0.01, am multiplying by 100 , or rather presenting it as basis points)
  • Median Change % * 100 : 2.62
  • Total Gains %:  30.47 ( sum of all the next day gain% , for the 3194 trading days , when the $SPX is up )

if $SPX is down on the current day

  • Winners : 1524
  • Losers : 1266
  • % Winners : 55%
  • Average Change %*100 : 5.96
  • Median Change % *100: 10.29
  • Total Gains %:  166.41

the out-performance of investing on down day is 5.96/0.95 is 6.27:1 ( avg change next day for a down day divided by average change% next day , for an up day) 

Weekly Trends

if $SPX is up during the current week

  • Winners : 362
  • Losers : 327
  • % Winners : 53%
  • Average Change % *100 : 2.01
  • Median Change % *100 : 11.17
  • Total Gains% : 13.87

if $SPX is down during the current week

  • Winners : 327
  • Losers : 222
  • % Winners : 60%
  • Average Change %*100 : 32.27
  • Median Change % *100: 46.97
  • Total Gains% : 13.87

the out-performance of investing on down week is  16.05:1 

Monthly Trends

if $SPX is up during the current Month

  • Winners : 111
  • Losers : 67
  • % Winners : 62%
  • Average Change % : 0.65
  • Median Change % : 1.06
  • Total Gains% : 114.87%

if $SPX is down during the current Month

  • Winners : 68
  • Losers : 39
  • % Winners : 64%
  • Average Change % : 0.71
  • Median Change % : 1.15
  • Total Gains % : 75.52

the out-performance of investing on down month is  1.09:1

Quarterly Trends

if $SPX is up during the current Quarter

  • Winners : 41
  • Losers : 20
  • % Winners : 67%
  • Average Change % : 2.31
  • Median Change % : 2.37

if $SPX is down during the current Quarter

  • Winners : 21
  • Losers : 12
  • % Winners : 64%
  • Average Change % : 1.57
  • Median Change % : 4.50

the under-performance of investing on down quarter is  0.68:1 

That’s where ,  on a Quarterly basis, there is slight advantage to following the uptrend , otherwise , it pays to go against the trend players on a daily/weekly/monthly basis , albeit the transaction  charges and bid/ask spreads set to zero !!

Conclusion from Price Change % study , on a daily/weekly /monthly time frame it favors to be trend setter vs to be a trend follower on a quarterly time frame 

further a look at Correlations 

am calculating correlation coefficient of price change% of current day(week/month/quarter) vs previous day/week/month/quarter price change% .

as a rule of thumb , the usefulness of correlation coefficient times the number of sample size should be greater than 10 . Thus for a  for sample size 100 a correlation 0.1 would be useful , and if the sample size * correlation is less than -10, vice verse ( doing the opposite thing of the historical observation )

  • Daily Correlation of SPX since 1990 is -0.0587 for 5989 observations , our rule of thumb,  n*correl is -351.8
  • Weekly Correlation of SPX since 1990 is -0.077 for 1240 observations ,  n*correl is -95.42
  • Monthly Correlation of SPX since 1990 is 0.0537 for 286 observations , n*correl is 15.39
  • Quarterly Correlation of SPX since 1990 is 0.0546 for 95 observations , our of thumb ,  n*correl is 5.19

Conclusion from Correlation  study , on a daily/weekly time frame the correlation on SPX price change% are usefully negatively correlated for the trend setters , while monthly correlation number favors the trend followers .( know your time frame !) 

you might want to consider subscribing to our daily quant rants emai !!

check out Quant-Ideas and filter for short term trades of high winning odds as per user inputs

Leave a Reply