Anyone used WL to auto-trade live
Author: zanek
Creation Date: 6/20/2009 1:19 PM
profile picture

zanek

#1
I want to use my wealth lab program strategy code to automate my trades. I hear there is a difference in how the WL program acts during backtesting versus live trading (ticks vs bar closes).

Has anyone converted a Wealth Lab program from just a backtesting strategy to an automated trading program ?

I'm mainly curious to know if your results were vastly different from the backtested results and what changes if any you had to make to transition from backtesting to going live.

Any info about your results, changes needed, etc will be greatly appreciated.

Thanks !
profile picture

nexial_1002002

#2
I've never heard anyone on the site talk about autotrading. The requirements by Fidelity are a minimum of 500 trades made in WL during one calendar year. A stupidly high amount that I've never heard anyone talk about meeting. The results wouldn't be too different. Just add a couple ticks of slippage.
profile picture

zanek

#3
Yeah, I'm not expecting anyone to have tried it (I called Fidelity and they said almost noone does it), but I am curious to know how accurate the backtesting is vs real performance so I can know definitively how much the backtesting results are skewed.

The couple of ticks of slippage you talk about could be a big difference depending on how often it happens, but the only real way to know is to either do real time simulated trading.

Actually, do you know if its possible to have WL trade against the market as if it was real using the WL code ? That would provide close to real results (minus slippage)
profile picture

nexial_1002002

#4
You can do it in WL5. Turn on the strategy monitor and have it autosend orders to the order window with paper trading turned on. Then, at the end of the day, check the bid and ask at the exact time that order was entered and you'll get an idea how much slippage there is.
profile picture

zanek

#5
Thanks nexial, you're awesome !
profile picture

Cone

#6
In general, no changes in Strategies are required from testing to live. If there are, you may have made a mistake.

QUOTE:
ticks vs bar closes
A simulation is a simulation. In live trading there can be delays getting orders to the market. With market orders (and stop orders placed close to the inside market), that means slippage, which could work for you or against you. If you apply slippage in your simulations, it is always negative. With limit orders placed close to the inside market the potential is there that you miss the trade if the order is delayed.
profile picture

swuzy

#7
Why is auto trading not available for IRA account?

I understand they don't allow next day buy at market open to IRA or cash accounts (if you place it it is rejected by Fidelity).

But shouldn't it be fine to place buys at limit at market open from IRA or cash accounts.

profile picture

Cone

#8
swuzy, you're off-topic, but it is available for IRA. Call Fidelity if you have questions.
profile picture

HendersonTrader

#9
"I want to use my wealth lab program strategy code to automate my trades. I hear there is a difference in how the WL program acts during backtesting versus live trading"

I use it every day for live/automated trading. Your strategy code for backtesting should work nearly the same for live trading. I say nearly the same because the bad tick filter is relevant to live trading and is absolutely essential. Also, the assumptions about backtesting fills for limit orders are not going to correspond 100% to live trade actual results. If you do enough automated trading on one minute intervals, there will be times when you are the only fill at the limit price, hence your strategy will still be trying to buy, when you want to be selling. You should consider automated trading as an assistant -- only a lunatic would walk away from it for more than a bathroom break. You need 500 trades (250 round trips) of any kind during the past 12 months to qualify. The paper trading feature can give you 80% of the live trading experience, it is excellent.
profile picture

zanek

#10
Hey HendersonTrader & Cone,

Thanks for the replies. My strategy code only uses market orders & works on 1 min intervals, so slippage should be my only real concern correct ?

I just tried using the slippage in my WL backtesting and Cone was correct, its always has a negative effect. Fortunately it only affected my results by dropping them about 5-8%. I've made 250 roundtrips, so I am ready to start live trading my strategy code. I will probably forward test it for another week and switch it to live trading after reading what HendersonTrader has written.
profile picture

Cone

#11
QUOTE:
slippage should be my only real concern correct ?
I don't think slippage is really a concern, it's just a reality - you'll never get the first trade on the next bar. Due to delays, your trade will be one of the trades after that, and the difference is the slippage. However, like I said, for market orders, I'd expect to be on both sides of the slippage equation when trading live.

Plan on collecting data to determine your slippage each day. The community would certainly be interested in your results.