In case anyone cares I’ve done some work on the VTO system, crunching numbers to see how it’s holding up. Tho’ I don’t claim 100% accuracy, here is what I found.

There are 93 total, closed trades
68 winners
25 losers
73% winning trade average

largest gain +18.8
largest loss -14.1

3.72 to 1 winning trade ratio

the average winner (excluding the best gain) = 3.27%
the average loser (excluding the worst loss) = -2.4%

% gain ratio (or expectancy) = 0.87


this system has 9 years of data

broken down per year, the performance is like this

1997, win/loss ratio = 72.7% gain to loss ratio 2.13 to 1

1998, w/l ratio =62.5% g/l ratio 1.7 to 1

1999, w/l ratio =100% g/l ratio 6.2 to 0

2000, w/l ratio =78.6% g/l ratio 2.48 to 1

2001, w/l ratio =62.5% g/l ratio 1.7 to 1

2002, w/l ratio =71.4% g/l ratio 1.28 to 1

2003, w/l ratio =85.7% g/l ratio 2.1 to 1

2004, w/l ratio =55.6% g/l ratio .97 to 1 (breakeven)

2005, w/l ratio =81.8% g/l ratio 3 to 1


what I take away from this is:

This system seems to have a range from an 85% to 55% per year win rate.

It has a positive expectancy.

The biggest string of losses = 3

There is only one double digit single trade drawdown (during 9/11/01 - anomaly?)

It trades a little less than once per month on average

There has never been a down year (see “Trade G/L” column on the VTO site)

It seems to perform the worst in flat markets.
The worst annual performance was +3.1% in 2004

The “cumulative gain” column on the VTO report says it has a +381.8% gain or 19.8% per year average. However, if you compound your gains the performance is more like +480% (nearly +21% per year on average)

I didn’t track the average holding time but I know it’s very short – just a few days.

According to a study by Bernie Schaeffer, a system with a 75% win rate, over a 50 trade time frame, has the following probabilities of seeing a string of losses

2 consecutive losses, 95.8% chance
3 consecutive losses, 53.0
4 consecutive losses, 16.8
5 consecutive losses, 4.4
6 consecutive losses, 1.1
7 consecutive losses, 0.3

the chance for ruin? If you identify ‘ruin’ as losing 25% of your total account and since the average per trade loss is less than 5%, the chances of losing -25% to your account is about 4.4%. the likelihood of this happening is once every 50 trades (or 5 years).

Optimal bet size?

According to the Kelly Criterion (from Van Tharp’s manual) you take the winning percentage “W”, the average size of the winning trade divided by the average size of your losing trade “R”, and calculate:

Kelly % = W – [(1 – W)/R]

So…

Kelly % = 0.75 - [(1 – 0.75)/.87]

= .75 – [.25]/.87
= .75 - .29

Kelly% = .46

Tharp says that betting the full Kelly % (46%) is too hot & one should, perhaps bet 80% of the Kelly % which, in this case would be – 36.8%

SO – do we bet 36.8% of our account on the VTO strategy?

For optimal results, yes.

Just a reminder, 36.8% of your account is NOT a static dollar amount. If you have a loss, the next time you calculate 36.8% of your account size it will be a smaller dollar amount. Scaling up as you profit & down as you lose will help keep you in the game.

BTW: I’m not recommending this bet size – I’ll use less in my own trading. This was just an exercise.

Also, if anyone finds a bug in my accounting please chime in. the reason I bet small is because I KNOW there could be a mistake in my math!

-ja

That's a lot of work, thanks.
My own experience with the bb system isn't that good and I haven't tried the vto system. I think the big reason is that I don't get good fills all the time and maybe also that I have to enter and exit at the open instead of at the close.
Did you account for fills on the trades you actually made with the system?
thanks again for sharing this with us,
h

i got the raw data straight from the VTO report site. the calculations were mine.

i never traded the BB system per se. i created a BB system of my own & traded it for awhile but decided it was too much work for too little gain. the VTO trades less so it's not that time consuming. i was trading 20% of my account total using leverage. i may raise that a bit, 25 to 30%. my confidence in the strategy is high - i'd suggest crunching the numbers yourself & see if you agree that it's a great system before you trade it too aggressively. i did a survey of this system a while back & got similar results so i'd say it's holding up pretty well.

i find that my results are comprable with the data on the site. sometimes i lose on slippage, sometimes i do better. the main thing when trading these "bounce" stratigies is keeping commisions low. you rarely get a big gain so you don't want slippage & fees to eat you alive.

-ja

p.s. i MAY start trading PWO instead of QQQQ in the future. there may be a slight upward bias edge there. this is the OTC power shares XTF.

read about it at

http://www.powershares.com/otcfund.asp

and see if you agree.

Joe,

It seems that you would want to use at least $10,000 per trade to account for the commision costs.

For example, if a trade grossed 1%, then you get about $100 minus $20 for commisions, leaving you with $80.

If you used $2200 for the same trade, you net zero.

So this strategy, while having a good record, might not be suitable for someone with less than a $30,000 account. Yes/no?

(All figures, approximations).

Steve

yes - good point. like everything in life...
the more $ the better.



-ja

I tested a version of the VTO for the last 2 years under the assumption that we might be in this trading range low volatility market for a while. I tried waiting till the Stochastics were also oversold and sold any time the QQQQ's touched the upper BB. It is nice to be able to place a sell order in the morning and not have to watch the trade. I get 15 trades 13 were winners. I also let the trades run up to 2 months. The profits were larger which might help with the commissions. What do you think? I'm thinking about letting my current VTO trade run using the new sell signal.

Jim

Now that�s the kind of analysis a pro will do. Good work, Joe.

Some points�

You�re not going to get bad fills with QQQQ. It�s the most liquid instrument there is. The bid/asked is almost always a penny -- something like 40.70 bid, 40.71 asked.

It�s a fact of life that the smaller the account, the higher the trading costs. The solution to that problem is to not have a small account for very long. Your main focus should be to direct as much money as possible from income from your profession to build your account up as fast as you can. And that means people should focus on increasing savings and getting out of debt. It also requires a spending plan. It�s not what you make, it�s what you spend.

But don�t let a small account keep you from taking advantage of some of these high positive exit expectancy systems. You won�t make as much as if you had a large account, but it will still be better than most other things you can do with your money. In fact, I�ve always thought that using these kind of high percentage systems is one of the best ways to slowly build a small account.

The best way to look at BB or VTO is that you�re starting off with a decided edge. There are many variations of order entries and exits that you can use to suit your needs. You can use limit orders on entry to try to enter at a lower price. You could exit in after-market trading if you can�t be around for the close. You could place a limit order to exit by estimating the price the Q�s may be when the RSI reaches 50. You could place a limit order to exit after a certain percentage gain. You could do what Jim is suggesting. Those orders could be place intra-day.

The point is that when you start off with that much of an edge, it�s kind of hard to screw it up.

By the way, I�ve been having trouble reading Joe�s posts lately because there are some strange looking characters in it. Somebody told me that I was having the same problem on some of the posts on my blog. Does anybody know what�s causing that?

Larry

larry,

i get odd symbols on some posts too. i don't know what's causing it.

some computer geek - HELP!

-ja

Joe,
Have you run an antivirus & antispyware check lately? How about those oils today? Their practically giving them away. I think Buffet said in the bottom of 74 when he was buying like crazy he felt like a drunken sailor or something to that effect. Most of the oils and oil itself are severely oversold. Levels this oversold have always been at or very near the bottom in the past. I'm guessing we get a big bounce next week. In the meantime if they want to give them away I'll be happy to take them off their hands for cheap.

Gary

"Have you run an antivirus & antispyware check lately?"

gary, how would i do this? i have norton ute. it "updates" automatically every week or so. is there something else i could do?

re: oil. my crazy partner, mr. market is in one of his moods!

the day will come, sooner than we think, when we speak of oil coming DOWN do $70 per barrel and we will call that "cheap."

-ja

Joe,
Norton is probably OK for virus protection. But you probably have a ton of spyware on your computer. Do a google search for Ad-Aware. The free program is pretty good but the professional is better and it has real-time monitoring that keeps the crap off your computer. Its only $29.00 I think for the pro version and worth it in my opinion.

Gary

Since this list is filled with special characters I figured those weird letters were from a different font. They all replace to punctuations and they are probably due to some font incompatability rather than a virus.
Like the rest of us characters, it just adds to the charm.
h

if anyone would like a copy of my "report" that is clean (no weird symbols) reply to my e-mail and i'll zap it out to you.

-ja

if you click on my name you'll get my e-address (i think that's how it works).

Joe,
Can you explain how you got those ratios? My Tharp book is packed and when I try to recreate the ratio from your data I get something different so I'm probably not doing it right.
I have 53 years of data on a simple moving average trend following system and I'd like to do what you did in order to do the money management part which of course is the most important. I know there will be big drawdowns I just want to quantify how big.
It only trades once or twice a year on average but there are periods of frequent trades in non-trending markets that hurt. In case you're interested there were 92 trades, 37% gaining trades, average gain 18.28% and average loss 1.82%. It sounds good when it's explained like that but I don't think it's as good as VTO yet.
Thanks,
h

"It only trades once or twice a year on average..." that's going to hurt you. if you have even a slight positive expectancy system that trades OFTEN you'll make more money assuming the system holds up - like the VTO seems to be doing. how often you trade has to be factored in.

i know a system that's 100% accurate. buy 30 year T-bonds at face value... wait 30 years.

re: the ratios - i'm not sure "ratio" is the right word, i just did averages.

example: 1997 w/l "ratio" is 72.7% -- simply take the number of winning trades (8) divided by total number of trades (11) and you get 72.7%.

for the "gain/loss ratio" i just divided the year's biggest gain (4.9) by the biggest loss (2.3) and got 213%. i call that 2.13 to 1 since the gain is 2.13 times greater than the loss. now that i'm looking at it i should've used the AVERAGE gain/the AVERAGE loss. oh well...

regarding expectancy; i don't have van's book handy now, but you can see that if the w/l ratio is greater than 50% AND the average winning % is greater than the average loser it MUST be positive expectancy.

if you ARE looking at van's book & the correct VTO data & you're coming up with a different number than me - don't assume you're wrong. my math could be off - but it IS a positive expectancy system.

it's easy to copy & paste the data right from the VTO website & do your own figuring. if you do you'll see that it's a good system & you'll probably have the confidence to trade it.

-ja

Joe,
I was using average gain andloss and that was the difference, thanks,
I'm not looking for a frequent system, I'm just trying to build something that catches big moves and stays in them = just another diversification of trading systems. I so far don't really have anything for trend following except our psar stuff,
h

Re: your system…

If I understand you correctly - you make, on average, 18.28% thirty seven percent of the time and you lose, on average -1.82% the rest of the time… sixty three percent of the time. So you make about +561.7 over the timeframe which is 53 years – that comes to about +10.6 per year which is probably about the same as a buy & hold strategy of the S&P 500 over that same timeframe.

I’m pretty sure that Magic Formula/PSAR, VTO, & COT all have and/or will beat that. If you’re looking for infrequent trading why not buy Magic Formula stocks using the MONTHLY SAR dots? That’ll keep you from trading too often.

Just a tho’t.

-ja

Joe,
You're right, it's not a big improvement but it's just the basic entry/exit idea for the system.
From 1953 to 2006 it turned 10,000 into 297,595 not counting dividends, fees or taxes. I made an "improvement" by changing the timing parameter to make it slower. It then traded only 79 times in that same period but had an ending equity of 561,047. I want to now add the exit for shorts as they hit the 20% less than the 200d sma limit. That will improve it more.
I've already developed a trending system but I used stops for the money management part. I've been reading the trading tribe site and am trying out using the atr for sizing. It will be a lot simpler to add leverage than it would be for the system I have already.
As to why I want a long term infrequently trading trend following system: Since I figured out that I want to use the COT, I haven't had a trend following system. If I had I would have had a much smaller drawdown. Sure there will be times when the trend following system loses as well at most inflections but it's not as if it will miss the big moves that the COT catches so if I can get the system right, I'm hoping to decrease equity volitility and not decrease the returns. That would allow me, I think, to hold a bigger position in COT and still be inside my sleeping factor. That my sleeping factor is probably weenyish compared to many more experienced traders I have no doubt but I want to be realistic about it and not try to use someone elses sleeping point. In looking back at last year, exceeding my sleeping point led to most of my losses.
As to MF/PSAR, you won't find a bigger fan. I've never had such successfull trades. But it must have periods of under performance as well and none of us can back test the way we're doing it because we don't know what the MF stocks were in 1998 for instance and how Larry would pick between them and which ten to hold at a time if you'd limit it to Larry's ten. We might find it does poorly at exactly the time when a trend following system is catching big moves as people may be favoring less quality stocks in a booming bull market and lowering all boats in a long bear market.
I'm trying to figure out how active to be as well. It appeals to me to come home on Fri, check one number and put in one order for monday and go back to my life. I enjoy trading but I enjoy winning more and my trading has not been until this year profitable so I haven't wanted to expand beyond the test phase with it. So until the problem isn't me any more I'd like to take me out of the decision making as much as possible. COT, a long term mechanical trending system, and a limited heat of MF/PSAR stocks, a core (and now larger) PM position seem about right.
I really appreciate your questions and insight, it's helping me clarify my thinking.
h

a "weenyish sleeping factor" will keep you in the game for the long haul!

-ja