Matthew,
if I understand it right, the "Keep after worst-case slippage" is worse for systems that use limit orders as opposed to those that use market ones.
I guess that this is so because the risk of the price just touching the limit and turning again could cause slippage to many subscribers and this parameter shall take that risk into account.
However, I believe that this risk should not be assumed automatically for all systems using limit orders. I will explain.
It is possible that a system uses only limit orders but chooses them in such a way that they are always crossed. In that case there should be no difference in value for "Keep after worst-case slippage" between market orders and limit ones (limit orders would even have the advantage of knowing exactly the execution price, what cannot be said about market orders).
Does C2 in its computation of this factor after each trade check whether the limit was crossed? If not, I believe that it should and that whenever the limit is crossed, the contribution of that single trade to the computation of the parameter should be at least as good as it would have been for a market order.
Any opinion about this?
It should be based on the ask price, in the case of a buy, and the bid price, in the case of a sell.
nothing against it, yes, that makes it totally robust.
"It is possible that a system uses only limit orders but chooses them in such a way that they are always crossed."
I would love it if my limit orders are always crossed. I seem to get a lot of touched but not filled situations.
Can you explain your statement above? Maybe with an example.
THANKS
That’s because your system isn’t robust. If you’re not getting filled, it’s because the price didn’t trade through, so you wouldn’t get filled anyway.
Thanks, but my question was directed at OldEurope Trader, who said:
"It is possible that a system uses only limit orders but chooses them in such a way that they are always crossed."
This would fly in the face of everything I’ve heard for the past 20 years, which is exactly your point Beau: “because the price didn’t trade through.” That’s the problem with limit orders.
OldEuropeTrader seems to claim he can avoid this fundamental truth, and I’d like to know how…
Here’s a tip, don’t submit till the limit price is at the ask for the buy, or the limit price is at the bid to sell. Easy.
This would work a lot, but not always. When things are moving fast, many times by the time you submit the order, the bid/ask has moved.
Only on an inferior platform. You could hit the market order button immediately when the ask is there, as I’m doing for my [LINKSYSTEM_35438029], so that I never miss my fill.
But now you did not use a limit order - you used a market order.
By the way, nice plug for your system. Hardly noticeable.
Exactly, for your subscribers, this will be what happens anyway, so you got to do it like this to keep in synch.
please be more exact. I never claimed anything.
I was only talking about the possibility of that happening in a theoretical sense in order to illustrate the problem I am addressing.
You should admit that (even if it is only by chance) the possibility of such a system exists as the markets are not moving just one tick up and down the whole day long. Why punishing a system automatically because of the assumption that limit orders might not be filled, specially in the cases where the price was crossed wihthin a few seconds?
What I am talking about is maybe not about "always" but "sometimes". And in those "sometimes" cases, limit orders are -at least- as good as market orders. Why should they count less for the computation of the "Keep after worst-case slippage" parameter than market orders?
OK, I understand your post now - "always" is a bit different than "sometimes."
