Now I am not saying the right thing to do is put a closing limit order in for 10-15 points and take a nap. The right thing to do for any trade is to get out when the market says, "get out". What I *am* saying is for the NQ, it doesn't make sense to have a setup average profit goal of less than 10 points. You have to have an higher probability set up just to break even at those low levels. For a five point goal, you are forking over a 7.5% chance just to trading costs. Your setup has to be that much better just to break even! That's not a playing field I'd choose to play on. (But if you do and are successful, I stand in awe.)
Pete's Place
- Thread starter PetaDollar
- Start date
By the way, if you trade something else and want to calculate these probabilities, the formula is
prob = (T - spread - commish) / (2*T)
where T is the point target, and commish must be expressed in points for your instrument. E.g. if you pay $5 to trade ES then commish is 5/50 = 0.10
prob = (T - spread - commish) / (2*T)
where T is the point target, and commish must be expressed in points for your instrument. E.g. if you pay $5 to trade ES then commish is 5/50 = 0.10
Well, for one, this tells me why Reject Trader has the best setup. You get in on one side of the range and look for a breakout past the other side. You know how many points the range is, so you can easily see if it's worth it in advance: if the range is 10 points, you figure to get 8 of those, plus say another 5-10 points after the breakout to the next s/r level, then it's a go!
However, if you're looking at say, a five point range, with s/r another five points away should it work out, that one wouldn't be a very wise idea.
However, if you're looking at say, a five point range, with s/r another five points away should it work out, that one wouldn't be a very wise idea.
Pete,
In addition to the implicit assumption of a 1:1 ratio in the above formula, you're probably already aware, but it bears mentioning for others reading this thread and attempting to calculate their own odds that there is an implicit, though unstated 50%/50% win/loss ratio as well in the above formula.
Actually whenever we boil down trading into a linear formula, we have to simplify things in order to get them into algebraic form. The only way that I know to keep track of the basic formula you provide, along with a variable money management scheme such as a trailing stop or multiple profit target exits is through forward testing. Someone smarter than myself may have figured out how to put it all into a math formula. But I think the exercise is useful to figure out your expected return, and get a feel for whether you can realistically expect to make money with the chosen system parameters.
Tri, I only intended that formula for the stated purpose, to estimate the minimum size move a trader should play for.
Pete,
Yes, and if you subtract the percentages (above) from 1 you get the % of winners that your system NEEDS to have just to break even when shooting for the specified # of points for both a target and a stoploss. For example for a 1 pt target/stoploss, you need 87.5% winners just to break even, but for a 10 pt target/stop you <only> need to win 53.75% of the time to break even. This really does show the relative high cost of slippage and commission, particularly for micro scalp systems.
Good stuff, and definitely worth punching the numbers in the old cacluator one more time!
And after you know all the win percentages that you need just in order to breakeven, the next question becomes, "Do I have an 87.5% 1:1 winning system? (or really a 95% winning system so I can make some $$$)" And if not, then 1 point PT/SL is a pipe dream.
Hi Pete and Tripack,
The problem with the probability calculation is that the parameters of spread and commish vary very much with T.
Even a beginner scalper knows he has to make the spread rather than it work against him. Also by the very nature of his scalping strategy a scalper will generate a lot of trades and can get a deal where his commish will come down to 0.1 instead of 0.25.
If you put 0 for spread and 0.1 for commish then you come up
with a loss percentage of 55%.
Not too different.
The problem with the probability calculation is that the parameters of spread and commish vary very much with T.
Even a beginner scalper knows he has to make the spread rather than it work against him. Also by the very nature of his scalping strategy a scalper will generate a lot of trades and can get a deal where his commish will come down to 0.1 instead of 0.25.
If you put 0 for spread and 0.1 for commish then you come up
with a loss percentage of 55%.
Not too different.
Tex, the NQ spread is always 0.5 pts (expect for after hours, when it is sometimes more). Supposing zero commish-- and a one point profit target-- that trade has to move three ticks in favor without moving one tick against. In other words a random entry would achieve this goal only 25% of the time, or better yet as Tri points out, you need a 75% win rate to break even. And that's without any commish.
Yes, I know you can get out sooner, use stops, etc, but so can the guy with the 20 point target. My point is, while it seems that gunning for smaller targets is more reasonable/easier/less greedy, it actually starts you off with an extra disadvantage to overcome. Trading is tough enough, I don't need any extra disadvantages.
What this means for my NQ intraday trading is that I will not even consider entering for anticipated moves of less than 10 points. Incidently, this is a great "filter" for my entries, that i've been looking for all these months.
Yes, I know you can get out sooner, use stops, etc, but so can the guy with the 20 point target. My point is, while it seems that gunning for smaller targets is more reasonable/easier/less greedy, it actually starts you off with an extra disadvantage to overcome. Trading is tough enough, I don't need any extra disadvantages.
What this means for my NQ intraday trading is that I will not even consider entering for anticipated moves of less than 10 points. Incidently, this is a great "filter" for my entries, that i've been looking for all these months.
p.s. I can't even remember to get all the things I need from the grocery store 75% of the time
I think an example is in order. Price is trading (NQ at 1142.50 bid / 1143.00 asked. The scalper places his buy order at 1142.50. At this point one of four things is going to happen.
1) Price does not trade through 1142.50, and the scalper's order goes unfilled.
2) Price does not trade through 1142.50, and the scalper's order gets filled. This is the ideal situation for a scalper (zero drawdown).
3) Price trades through 1142.50, and so the scalper is immediately at a bid/ask spread disadvantage.
4) The scalper cancels the order before it is filled.
Now if we analyze 3), if the scalper is really quick, they can get filled on a limit order to sell at 1142.50 probably 90%+ of the time (breakeven). Some of those trades would have gone on to hit the profit target, and some of those trades would have gone on to hit the stop loss. The remaining 10% will probably hit the stop loss, regardless of what the scalper does.
So as can be seen, it is possible for a scalper to tweak the risk/reward ratio based on the particular tactics used, and to manage the spread in such a way that it doesn't necessarily have to be an immediate liability.