Long the first day of the month

[cire2222]

Could it be as simple as this? Get long the end of the last trading day of the month and sell at the end of the first trading day of the month. If you had done this in the past you would of seen returns over 30%.........hrmmmm

[trustdnb]

Originally Posted by cire2222

Could it be as simple as this? Get long the end of the last trading day of the month and sell at the end of the first trading day of the month. If you had done this in the past you would of seen returns over 30%.........hrmmmm

I backtested it over the last 10 years in Amibroker:

It returned a total of 37.35% over the 10 year period against the S&Ps (^GSPC) which comes to a roughly 3.6% compound annual return (CAR). 69% of trades were winners, and the maximum peak to trough drawdown was -5.41%. The average profit per trade was less than half a percent.

Accuracy was about the same for ^DJI although overall return was lower. ^NDX gave nearly a 5% CAR, but had almost twice the drawdown. Also the returns were markedly higher, with similar drawdown, when ran against ETFs (SPY, DIA, QQQQ) than against the cash indexes directly.

At 3.6% CAR it barely beats inflation as a standalone system, but then, you're only in the market for 12 days a year, which aint bad. It could actually be a pretty nice system if you juiced it with a little leverage. At the very least, you should think seriously before holding any short trades through the end of the month.

[cire2222]

Originally Posted by trustdnb

I backtested it over the last 10 years in Amibroker:

It returned a total of 37.35% over the 10 year period against the S&Ps (^GSPC) which comes to a roughly 3.6% compound annual return (CAR). 69% of trades were winners, and the maximum peak to trough drawdown was -5.41%. The average profit per trade was less than half a percent.

Accuracy was about the same for ^DJI although overall return was lower. ^NDX gave nearly a 5% CAR, but had almost twice the drawdown. Also the returns were markedly higher, with similar drawdown, when ran against ETFs (SPY, DIA, QQQQ) than against the cash indexes directly.

At 3.6% CAR it barely beats inflation as a standalone system, but then, you're only in the market for 12 days a year, which aint bad. It could actually be a pretty nice system if you juiced it with a little leverage. At the very least, you should think seriously before holding any short trades through the end of the month.

cool thanks for the backtest data. 12 trades a year, Add some juice by getting leveraged futures contracts and your set. (don't we all wish it was that easy)

[MC]

Originally Posted by cire2222

cool thanks for the backtest data. 12 trades a year, Add some juice by getting leveraged futures contracts and your set. (don't we all wish it was that easy)

Maybe it is...it backtested out. I would be afraid to put money on it, but I do think traders overall make things way harder than they should be. Simple is best, now if I can just dial in my scans for stockfetcher to give a good pool of candidates, I think I know what I'm looking for. I'll now try to keep it below 45 crosses and criteria. LOL

[trustdnb]

Originally Posted by cire2222

cool thanks for the backtest data. 12 trades a year, Add some juice by getting leveraged futures contracts and your set. (don't we all wish it was that easy)

Seems like one ought to be able to do something like that. Maybe Ultra ETFs?

[cire2222]

Originally Posted by trustdnb

Seems like one ought to be able to do something like that. Maybe Ultra ETFs?

ultra's only go 2X leverage from what I have seen. Even so, it would just be an added tool in the kit, not the bread and butter

[cire2222]

Now there has also been studies showing wednesdays has the most likely day the market trades up. I wonder how often the first day of the month falls on a wednesday and what kind of results are produced on a wednesday first of the month.

[cire2222]

sell in may and go away is another long time theory that actually has some backing to it.



don't think i would ever actually trade it though

[trustdnb]

Originally Posted by cire2222

Now there has also been studies showing wednesdays has the most likely day the market trades up. I wonder how often the first day of the month falls on a wednesday and what kind of results are produced on a wednesday first of the month.

I see what you did there. You are giving me homework now. Well, just to show you I am a good student:

The best days by far were Monday's and Tuesday's. Monday's and Tuesday's combine for over 83% accuracy on the S&Ps and 75% accuracy on the Nasdaq, and about 2/3 of the system's profits come from trades entered on these two days. Wednesday's and Thursday's were both profitable, but accuracy drops to about 65% (S&P) and 50% (NDX).

The only day that was negative was Friday's. Basically, this trade should not be held over a weekend. The best overall system seems to be just filtering out Friday's. You could just take the Monday and Tuesday trades and do well, but you also forgo many profitable Wednesday/Thursday trades. Doing this, you basically get the CAR of the old system, with less drawdown. Taking every trade but Friday improves both the CAR and the drawdown a bit.

I can post full results if you need to see them. What might work well is raising position size for Monday/Tuesday trades while keeping a standard position size for Wednesday/Thursday.

As for Wednesday's I have read they are statistically significant "reversal" day, not necessarily a long day. I.e. you can fade the direction of the first two days. It could try to test what happens when the last day of the month is a Wednesday AND Monday-Tuesday is down, but there couldn't be more than 5-6 total trades over the 10 year period.

[cire2222]

Originally Posted by trustdnb

I see what you did there. You are giving me homework now. Well, just to show you I am a good student:

The best days by far were Monday's and Tuesday's. Monday's and Tuesday's combine for over 83% accuracy on the S&Ps and 75% accuracy on the Nasdaq, and about 2/3 of the system's profits come from trades entered on these two days. Wednesday's and Thursday's were both profitable, but accuracy drops to about 65% (S&P) and 50% (NDX).

The only day that was negative was Friday's. Basically, this trade should not be held over a weekend. The best overall system seems to be just filtering out Friday's. You could just take the Monday and Tuesday trades and do well, but you also forgo many profitable Wednesday/Thursday trades. Doing this, you basically get the CAR of the old system, with less drawdown. Taking every trade but Friday improves both the CAR and the drawdown a bit.

I can post full results if you need to see them. What might work well is raising position size for Monday/Tuesday trades while keeping a standard position size for Wednesday/Thursday.

As for Wednesday's I have read they are statistically significant "reversal" day, not necessarily a long day. I.e. you can fade the direction of the first two days. It could try to test what happens when the last day of the month is a Wednesday AND Monday-Tuesday is down, but there couldn't be more than 5-6 total trades over the 10 year period.

A+ there will be more homework