The fills on demo can be completely unrealistic. It is because limit entry and exit orders would get filled immediately when price ticks up to your sell limit or down to your buy limit.
In the real market your order is placed in a queue, for example I would place a buy limit order on the ES at (say) 1994.25. By the time that I placed my buy limit, there are already maybe 1472 buy limit orders in front of my order, at that price (1994.25). So for my limit order to get filled, either these 1472 limit orders in front of my order needs to get "eaten" up by sell market orders (hitting the bid), or all of the 1472 orders need to be cancelled. As is normally the case, some of the 1472 orders get consumed by sell market orders and some gets cancelled, but all of them must disappear before it is my turn to get filled by the next sell market order.
Where the big difference between real markets and demo markets lies, is in the fact that in live markets, very often a portion of the 1472 orders gets consumed, and before my limit order gets hit by a sell market order, price moves away from my buy limit orders towards my target without me on board. This means that a lot of moves that you catch on demo you would have missed if you traded the same way in a live market.
I learned a lot about this from the webinars by (1) John Brady and (2) Jigsaw. These webinars are posted here on futures.io (formerly BMT) and are likely to be worth your while to watch.
Of course you need market depth data, as is available with futures, so if you trade spot forex this approach is not possible. A possible solution might be to then switch to forex futures (like the 6E.)
There are certain order entry platform / data feed combinations that will give you very realistic demo order entry queuing. I know that Jigsaw does this, but there may be other platforms with DOM's that show you your place in the queue. Of course, besides the order entry DOM, you need quaility data.
To recap: to make the demo order fills realistic you need (1) an accurate representation of your place in the queue and (2) an accurate representation of the consumption of limit orders in front of your orders in the queue. This can be obtained by using certain order entry tools in combination with a quality data feed.
Hope that helps in answering your question.
The following user says Thank You to Fatfish for this post:
@donovanshane In NT, make sure you do not have "Enforce immediate fills" checked in tools/options/simulator, because that is super easy mode and you can scalp for 1 tick.
Without the check filled, you are operating in pass through mode, where it has to fully pass through the price to fill. Therefore, to scalp for 1 tick you need a 3 tick movement. Much harder and as realistic as NT sim will get. A much better way to learn.
Now I have seen where I got in, or out, and it was the exact top/bottom of a bar, and the dom did show the full movement past. Therefore, I would assume that is because the chart did not print the last price. But the few times that happens, it really doesn't matter, you're here to learn.
Head up and keep your stick on the ice!
Last edited by Fritzk; September 6th, 2014 at 01:52 PM.
The following user says Thank You to Fritzk for this post:
It's not possible to manually scalp ES during daytime hours for 1-2 ticks and make $1000/day.
My fees on ES are really low. Sub 20c all in and I wouldn't even consider trying it.
I have 1600+ running strategies. And NONE on ES. I'd only recommend someone manually trade ES if I really hated them and wanted them to lose all their money. I don't hate anyone that much. Not saying manually trading it can't possibly be done but I'd trade almost any other contract first.
Overnight... maybe... nah. not with one lots. But I do know of precisely 1 profitable manual ES trader. He only trades overnight and large enough size to push the market.
Well, I guess I know of 1 other guy. "The guy". Where he has the entire ES ladder stacked with GTCs a few hundred levels out. And he does OK but has enormous PnL swings.
Simulator is broken and not estimating your position in queue correctly. There's no motivation to fix this because then nobody would be profitable in sim and then how would the broker make any money...
The following 5 users say Thank You to Garachen for this post:
Not sure I read this correctly: do you have strategies that trade 1600+ symbols, or do you have 1600+ distinct profitable strategies?
In the first case, I'd say that you impressive scaled your strategies into multiple symbols. For the latter, I'd say you're dreaming because even big hedge funds supposedly do not have 1600+ different strategies.
The following user says Thank You to Jura for this post:
You are right there. I could have been more specific, but my point is that I've tried a lot of stuff over the years and have nothing running on ES. People are attracted to it like flies but there are at least 2 good reasons why I think it's a terrible contract for manual trading.
1) The difference between the fees retail people pay and HFT people pay on ES are very large. If you lease the IOM seat and beat your FCM down to 10c or less then you kind of are competitive. But without leasing a seat - very hard.
2) Dominated by arb, stat arb, queue holding strategies, etc. It's almost guaranteed that any passive fill you get will immediately go against you. It's lucrative if you play the game but very hard to play.
Maybe it's possible to swing trade ES. But I don't see how you can make a living off of that without doing fairly large size on a large account.
To make numbers simple: pretend you need to make $1250/day to live and grow your account. Ignore fees. Not sure how many RTs qualify as swing trading. Lets say 4/day. Pretend you are the best trader I've ever heard of and can average 1 increment per RT - consistently. Then you'd need to swing trade with 25 lots on ES.
Slightly off topic:
I should have said 1600 implementations of strategies. I've written about 100. 20ish are useful and 5-7 with various tweaks bring in the majority of the money - other than manual trading.
But then they are applied everywhere. Outrights, spreads, fly, doublefly, cross product spreads. With different parameterizations. That's where the number comes from.
So, I agree with you. When I was at a reasonably sized hedge fund there were about 8 core strategies types. Two made most of the money.
The following 3 users say Thank You to Garachen for this post:
So what I'm gathering here is that TOS paper trading is pretty much like preschool and if we want a more "realistic" approach to testing the waters then we should try something else (I think someone mentioned an emulator). What software could be used for a good emulator? I too am using TOS Paper Trading and I've had similar results. I've found trading ES on there relatively easier than trading anything else, but it seems too easy. You get really clean entries and exits.
The following user says Thank You to nychw82 for this post:
i think you do not need to change anything from tos paper trading for now, could be good for you to open a journal here ,write down explaining why you enter ,and exit, in unless 100/200 trades ,and then you can qualify your results and calculate your method expectation.
If after that you continue saying the same , you can jump to earn money on live
La lucha es de igual a igual contra uno mismo
The fight is fair against oneself