jackbravo 

Thanks for that post! Very informative. Like you mentioned, I've tried to average down on a pull-back, and have ended in disastrous results, as my initial interpretation was wrong. However, other times, many times in fact, I would have been smart to average down, as I had the right interpretation of price action/market conditions. So do does a beginning trader differentiate from those two circumstances?? Another component of risk aversion is avoiding profit loss. How many times have you let a trade go in order to maximize profits, just for the trade to come back to entry point and go back the other way? For me....that has not been an uncommon occurrence. So it seems to me having multiple contracts on at the same time allows one to deal with the ability to secure profits, as well as reduce that component of risk aversion. I mean, how often have you put on a trade and be completely, instantly wrong? I would guess not that many times. So let's you're at least partially right most of the time, very right some of the time, and very wrong some of the time. Wouldn't it work to be able to trade multiple contracts, and at least secure some profits most of the time? My apologies in using your journal to think aloud.

No apologies necessary! This is a great topic.

Now, I know you said partially right MOST of the time, and everything else SOME of the time, but let's go ahead and split it evenly just for the sake of argument.

OK, so let's say we have an equal probability of being really right, partially right, or really wrong.

Let's say Really Right means the market hits a 2R gain before hitting a 1R stop.
Let's say Partially Right means the market hits 1R gain before hitting a 1R stop.
Let's say Really Wrong means the market hits 1R stop before any gain.

Let's also say we aren't touching our stops ever.

A 1-lot trader in this instance would be wise to take profits at 1R every time. 2 out of 3 trades he or she is going to take profits. 1/3 of trades will leave 1R profit on the table, but overall it is a profitable strategy so who cares? Over 9 trades, he's getting +6R -3R for a total of +3R. Upping to 2 lots all-in all-out would be 6R in profit over 9 trades.

If the 1-lot trader decided to stick to the 2-R target, he'd be treading water overall and would lose money over time to commisions. 2/3 trades would hit the 1R stop, and 1/3 trades would hit the 2R target, cancelling each other out.

Now let's try the 2-lot all-in and scale-out trader.

This trader takes one lot off at 1R every time. Half the time he hits 1R, his other lot hits the 2R target. But the other half of the time, it doesn't, and it goes all the way back to his stop loss, cancelling each other out. On top of that, when he's really wrong, he loses 2R each time.

Breaking this down: out of 9 trades, he'd have 3 that lost outright for -6R. He'd have 3 that hit the 1R for 1 lot, but then went back to -1R for the other lot. Then he'd have 3 that hit 1R + 2R each, for +9R. So in the end he's at +3R, the same place the 1-lot trader was who just went for 1R per trade. And he's paying more commissions and has a lower w/l ratio. You might argue that in cases where the market goes to 1R and then turns back, they wouldn't let the 2nd lot go all the way to their original stop loss. OK, fair enough, so they flatten when it becomes clear that the trade isn't going to hit that 2nd target. But is this loss-averting tactic going to improve things that much? Enough to avoid the -3R and also not shake him out of some of the trades that hit both targets? Doubtful. Probably better off to use the extra size to just be a 2-lot all-in all-out trader going for 1R.

Now, how to break this down for the 2-lot scaling-in / all out, averaging-down trader? It becomes pretty complicated. To simplify, I'm going to have this trader only go for 1R. It works out better for him that way anyways.

Let's say for half of each of 2 positive scenarios, price moves .5R down before going up to its happy place. So now we have 5 scenarios.

1) 16.5% : Really Right, minimal Adverse Excursion.
2) 16.5% : Really Right, > .5R Adverse Excursion.
3) 16.5% : Partially Right, minimal Adverse Excursion.
4) 16.5% : Partially Right, > .5R Adverse Excursion.
5) 33% : Really Wrong.

For scenario 1, trader can't add because there is no adverse excursion to average into. +1R on 1 lot.
Scenario 2, trader adds at -.5R and comes away with +1.5R + 1R = 2.5R
Scenario 3, trader can't add and he takes the +1R.
Scenario 4, trader adds at -.5R, comes out with +1.5R + 1R = 2.5R
Scenario 5, trader adds at -.5R and ends up down -1.5R on each of these trades.

So over 6 trades we would be at +1R +2.5R +1R +2.5R -1.5R*2 = 4R, which means over 9 trades this trader would be at +6R.

So, yes. Given this very arbitrary set of assumptions, the average-down trader has slightly lower commision than the all-in all-out 2-lot trader, and identical PnL otherwise. He isn't always in 2 lots, but still manages to eek out 6R over 9 trades. That said, this is all completely made up hypothetical probabilities and I'd bet that the 2nd lot picked up at -.5R actually doesn't have as good a chance at that late entry of coming all the way back the +1.5R to get to the original 1R target. It probably doesn't happen even 40% of the time, if done randomly.

Things obviously get infinitely more complicated in real world situation where these probabilities don't play out so simply. And without that basic edge driving profits, there is no basis for profitable trading at all. It just can't be random, and it seems to me that playing with averaging doesn't help matters enough on its own. Adding to and subtracting from trades has to be PART of the edge. In other words, each buy or sell must be rooted in a decision making process that wins profits over a large sample of trades. The rationale can't simply be about playing a numbers game, getting a better cost basis or limiting risk

My gut tells me all this business of multi-lot scaling draws beginning traders' attention away from what they should be focusing on - Buy Low, Sell High.

FlyingMonkey 

No apologies necessary! This is a great topic. Now, I know you said partially right MOST of the time, and everything else SOME of the time, but let's go ahead and split it evenly just for the sake of argument. OK, so let's say we have an equal probability of being really right, partially right, or really wrong. Let's say Really Right means the market hits a 2R gain before hitting a 1R stop. Let's say Partially Right means the market hits 1R gain before hitting a 1R stop. Let's say Really Wrong means the market hits 1R stop before any gain. Let's also say we aren't touching our stops ever. A 1-lot trader in this instance would be wise to take profits at 1R every time. 2 out of 3 trades he or she is going to take profits. 1/3 of trades will leave 1R profit on the table, but overall it is a profitable strategy so who cares? Over 9 trades, he's getting +6R -3R for a total of +3R. Upping to 2 lots all-in all-out would be 6R in profit over 9 trades. If the 1-lot trader decided to stick to the 2-R target, he'd be treading water overall and would lose money over time to commisions. 2/3 trades would hit the 1R stop, and 1/3 trades would hit the 2R target, cancelling each other out. Now let's try the 2-lot all-in and scale-out trader. This trader takes one lot off at 1R every time. Half the time he hits 1R, his other lot hits the 2R target. But the other half of the time, it doesn't, and it goes all the way back to his stop loss, cancelling each other out. On top of that, when he's really wrong, he loses 2R each time. Breaking this down: out of 9 trades, he'd have 3 that lost outright for -6R. He'd have 3 that hit the 1R for 1 lot, but then went back to -1R for the other lot. Then he'd have 3 that hit 1R + 2R each, for +9R. So in the end he's at +3R, the same place the 1-lot trader was who just went for 1R per trade. And he's paying more commissions and has a lower w/l ratio. You might argue that in cases where the market goes to 1R and then turns back, they wouldn't let the 2nd lot go all the way to their original stop loss. OK, fair enough, so they flatten when it becomes clear that the trade isn't going to hit that 2nd target. But is this loss-averting tactic going to improve things that much? Enough to avoid the -3R and also not shake him out of some of the trades that hit both targets? Doubtful. Probably better off to use the extra size to just be a 2-lot all-in all-out trader going for 1R.

I'm going to take some time to digest the 2nd part of your example.

Regarding the 2 lot entry. Once one lot hits 1R, you automatically move stop to entry price. At least, that's what I would have in mind. There's no reason to risk any negative exercusion against your entry point by that time. Also , the idea of using two lots is to give you the flexibility to get some profit when you're kind of right.

So:
Very wrong A: 3 trades x 2 lots x -1R = -6R
Kind of right B: 3 trades x 1 lot x 1R = 3R
------------------ 3 trades x 1 lot x 0R = 0R
Very right C: 3 trades x 1 lot x 1R = 3R
------------------- 3 trades x 1 lot x 2R = 6R

Total: 6R

However, a 1 lot trader always taking a 1R profit would still profit 1R, which like you said, is still a positive edge.

-----------------------------------------------

But let's say, you're using your first lot to get a bit of profit, even a tiny bit, let's say 0.5R (something the 1 lot trader can't do).

So:
Very wrong A: 3 trades x 2 lots x -1R = -6R
Kind of right B: 3 trades x 1 lot x 0.5R = 1.5R
-------------------3 trades x 1 lot x 0R = 0R
Very right C: 3 trades x 1 lot x 0.5R = 1.5R
-------------------3 trades x 1 lot x 2R = 6R

Total: 3R

So you can still salvage positive expectancy in this situation. In fact, any profit over commissions using the 1st lot of two will give you a positive expectancy. However, if your 0.5R is too close to entry point, then there's a chance you never allow your 2nd lot to get to 2R - so I guess that's the negative of that situation. You'd have to use something like an ATR or some other way to judge noise/retracement patterns.



EDIT:

Actually, as I was thinking about it, the reason I was really interested in the double contract scenario is begin I would like to get away from thinking about risk/reward. I've noticed that many successful traders like Tiger or Intelcap don't think of risk/reward, but think in terms of trying to capture the bulk of a move on any given day (perhaps one can argue that a move isn't really over at the end of the day but may take two or three days). In that scenario, using a one lot, you are either forced to scalp, or forced to hope that you can tell when the move is over. Otherwise, if you're wrong, you either lose money being wrong, or you lose money when the move is over and retraces.

However, with 2 contracts, you now have to ability to take some profits...i.e. scalp, and use the second contract to allow the move to continue as far as you think it will go. So even if you misinterpret the end of the move and it retraces back to your entry point, you've still made some profits from your first contract. If you're able to interpret the end of the move correctly, well, now you've move a much larger multiple of R.

jackbravo 

Actually, as I was thinking about it, the reason I was really interested in the double contract scenario is begin I would like to get away from thinking about risk/reward. I've noticed that many successful traders like Tiger or Intelcap don't think of risk/reward, but think in terms of trying to capture the bulk of a move on any given day (perhaps one can argue that a move isn't really over at the end of the day but may take two or three days). In that scenario, using a one lot, you are either forced to scalp, or forced to hope that you can tell when the move is over. Otherwise, if you're wrong, you either lose money being wrong, or you lose money when the move is over and retraces. However, with 2 contracts, you now have to ability to take some profits...i.e. scalp, and use the second contract to allow the move to continue as far as you think it will go. So even if you misinterpret the end of the move and it retraces back to your entry point, you've still made some profits from your first contract. If you're able to interpret the end of the move correctly, well, now you've move a much larger multiple of R.

Yep, I get where you are coming from. When trading one lots there is no concept of a "runner". You either hold or you don't. Maybe you jump in and out. But still, for me the question that we need to answer: Does it make sense for a trader still on the road to consistent profits to trade 2 or more contracts, when he or she hasn't yet established that he can be profitable with 1 contract?

Does this depend on style or method? Individual trader personality? Wouldn't it make sense for a beginning trader to take as his first step being able to consistently capture small-to-medium moves on 1-lot trades, and only after showing a positive expectancy over a hundred or so 1-lot trades, start to add another contract as a runner to capture the big moves?

This way, the additional lot is purely an added value to an already profitable method.

Perhaps one of the reasons I personally have struggled is that I have been using one contract to try to hit larger moves, when actually I would be better served at this stage by going for smaller targets with that one lot.

That is definitely the right question to ask, and I continously ask myself that same thing. It doesn't make any sense to trade 2 contracts when you can't trade one with any consistent success.

But let me put this to you....what if that entire concept is completely wrong? What if the game is rigged so that you've got a very high likelihood of failure trading only one contract (backfills, stop hunting, high commissions, etc.)? So maybe the reason becoming consistent is so tough is because using only one contract, you're battling such huge odds that it's nearly impossible to be successful....i.e. you have to develop perfect market interpretation.

Of course, trading multiple contracts requires much higher capitalization. But my question is, what if that basic supposition of needing to be successful with one contract first is actually the exact opposite of what we should be doing/learning? Maybe we should really be learning the trading process with multiple contracts right off the bat......

I don't know. Just playing devil's advocate. Personally, I find it very difficult to trade multiple contracts. I think, though, that's because all the things I've read, and all the time I learned to trade, it was with one contract. Now it requires a paradigm shift to adopt a multiple contract perspective.

I believe trading multiple contracts definitly increases your profitability and reduces overall risk.
But trading multiple contracts is not the solution if one is not profitable with 1 contract yet.

The question (atleast to me at this stage of my trading) is what strategy to use regarding multiple contracts:
- All-in and all-out, or
- All-in and fade out, or
- fade in and fade out, or
- etc..
I'm experimenting with this myself and have not found what fits me best yet.
It also depends on the type of market, for example range days or trend days.

Regarding the fixed profit targets, I suggest to review your trades and see what the profit would have been if you traded multiple contracts and various profit targets (1R, 2R, ...). This will give you insight/stats and let you choose the appropriate profit targets for you. I did this and lost my faith in fixed profit targets.
Knowing that price is moving between price levels as pivot points, yesterday/weekly/monthly HLOC, VWAP, etc. you can benefit from this by taking your entries and exits on these levels instead of fixed levels based on risk. Just try it out for a month orso, review your stats, adjust your trading rules accordingly, trade it for a month again, etcetera.

My 2 cents

Nice journal, keep experimenting and learning!

04
==
Created Monday 04 April 2016

So a couple weeks ago I did 50 coinflip trades, ala FT71. I ran the exercise in CL over the course of two mornings. I used a 5 tick target and 5 tick stop.

It was pretty interesting. I think that running through that many random trades in rapid succession helps one get more in tune with a probabilistic mindset. I think the way that randomness can disguise itself as a pattern is extremely important for traders to undestand, particularly in the context of evaluating trading results.

After 15 random trades, I had only gotten 4 winners, 27%. It looked pretty bad for my little random strategy. After 25 trades, I had gotten 9 winners. Still not looking great. And then, for the next 20 trades I went on a 70% winning streak and climbed all the way up to 23 wins out of 45 trades, better than 50%. At the end of the 50 trades I had a 46% winner ratio.

What does this mean? It means you have to make a lot of trades to really get a valid sample. More than 20. More than 50. Often more than 100. I wanted to play with this concept some more and I found this great site to play around with and visualise massive amounts of rapid coin flips.

The Binomial Coin Experiment

The page lets you set how many coins to flip (n) in each run and also control the probability (p) of a heads or tails. Results are painted against a standard distribution so you can easily see how your sample deviates from the expectation.

What's even better is you can set it to do batches of trials. Below is a sample of 10 trials of 10 flips each. So 100 total coin flips. Of the 10 trials, in this case I only got more than 5/10 twice. I got less than 5/10 7 times. Those results are pretty uneven.




Imagine "living" this data as a trader testing out a method, taking 10 trades a week for 10 weeks. Most of the weeks would be tough - 2 wins, 8 losses. 3 wins, 7 losses. Only a couple decent weeks. You might come to the conclusion that your method was not a 50% winning method. But keep going and it gets closer to reality. Here one after 1000 flips. Still often come up a bit lopsided, but the margin for error is more acceptable.




And once you get out to 10000 flips, they start looking pretty reliable. Now, I'm not suggesting that it isn't possible to tell how a strategy performs without taking at least a thousand trades ... but it certainly would help! But there's another thing I find interesting that I garnered from playing around with this app. From a pure "feel" perspective, it's a lot easier for a treading-water strategy to randomly "feel" like a drastically negative or positive method over a 100 or 200 trades, than it is for a really winning 66% strategy, to feel like anything other than a "winning" strategy, across the same samples. Once you tilt the probability knob past 60%, it becomes exponentially harder to get really bad batches of flips. Even a very bad batch, like the one below, looks marginally decent.



Anyhow, enough of that. What's the point? The point is to say that, although I did decide to tweak my approach, I am doing it mostly for reasons of suitability ... NOT because I had determined it was a losing approach. How could I have? I was treading water / losing small money after about 50 trades or so, and had not kept very consistent metrics.

Over the past couple weeks I have been simming a more nimble approach. Still looking at the markets the same way, price structure being the primary driver for decision making--impulse, retrace, chop., etc. But taking more info risk, less price risk, which means more opportunities. The increased agility didn't suit TOS with its higher commissions, stuttery data, and weak trade reporting, so I've moved over from TOS back to NT. Because of this, I've had to open a new account and am still waiting for funds to move over.

I'll try to put up some examples to the journal this week as I put the final pieces together in the workspace and tracking sheet. I'm under no illusion that my results will be earth-shatteringly stellar, but it has felt good "trading" (I hesitate to call simming "trading") with a bit more freedom. I think it will be the right direction. Since this is a bit of a reboot, I'll also try to congeal some semblance of a rules-set and post to the journal, as a signpost, before coming off sim. Maybe those rules will be more than just "get in when it smells good" and "get out when it smells bad", or maybe they won't.

Happy trading.

Today's "trading" (sim) was pretty good. All the markets I watch looked pretty sloppy early pre-open. With Crude Inventory and FOMC minutes I was expecting participation to be a bit off. I waited through this, and just after the open I saw something worth trading in ES.

1. Wait for price to get to an interesting level structurally. Look for movements that are clean and not choppy.
2. Look for price to not blow right through the level.
3. Identify a definitely-wrong price if you were to take the trade. This should be far out enough to resist shaking out.
4. If it starts to smell good, take the trade. Put a stop at your definitely-wrong price.
5. When it starts to smell bad, get out.

One of the main things I'm still experimenting with is #5. So far in my sim-trading over the past couple weeks, I almost never let my stop get hit. I'd just jump out after a couple whiffs of danger. Most of the time this didn't do much harm, and I was able to jump back in if it set up again. However, it did result in one or two shakeouts / missed opportunities. I'll be tracking this closely to see if cutting trades short is helping or harming.

I'm also of the mentality that, trading only one contract, I'm going to take profits quickly if I have to, and jump out early if I have to, and won't beat myself up too much about protecting profits. If/when I am in a position to move to two contracts, I will look closer at the idea of runners and letting it ride.

If all goes well I should have the new account funded and ready to go for Monday morning. Until then, approaching this week as a "live test", trying to treat sim as if it were the real deal, to whatever extent that is possible.

+18 ES

Again trading sim today.

Pre-open trade in 6A looked almost identical to yesterday's entry in ES, and went about the same. I held through 7500 as it didn't give much reason for doubt. I had mental take profit at 7487, but I got a bit greedy and waited to see if it would keep dropping through it. Unfortunately everyone and their mother was covering at that spot, so that decision cost me 8 pips as I bailed out when it ripped higher. Overall, a good trade. I think if I were trading 2 contracts I'd have wanted to take one off at 7503 and the second ideally at 7487. +28 6A



CL had set up very similarly around the time I took the 6A trade, but I consciously took the 6A trade instead out of personal preference. 6A tends to behave like a little pony ride compared to the bucking bronco that is CL. That said, I couldn't resist a short attempt later on as CL presented another pull-back that looked like it wanted to drop. I had a mental profit target of 37.10. I had the right idea, but I was a lot less calm in this trade (didn't want to give back profits, and generally (overly?) cautious in CL). I didn't let the market prove me wrong, and so I bailed out of a really good opportunity. +0 CL

There was a lot of movement this morning, but the chart that looked the cleanest to me was ZN. Also I favored the slow-moving ZN so I could step away from the screen and take care of friday morning business.

ZN has been in a steady uptrend for the last couple weeks and overnight showed a nice orderly pullback into a previous breakout point. I decided to give it a shot and put a stop in about 10 ticks away. Was tempted to bail out a few times as it made new lows after going in my favor a few ticks. The fact that I had gotten a decent price (low price risk) made it easier for me to sit it out. Reminded myself I still had a good price and the idea was still valid. Alas, I could only give this trade a couple of hours to work itself out. In the end, I just couldn't hold any longer as I had to go to work. Meanwhile, I wonder if the party is starting now without me? ZN +3 (SIM)

Back to live trading. Today took a counter-trend short on 6J. Didn't work out. On the bright side, I had a good entry price and picked a good place for a stop. Could have jumped off sooner, but damage was minimized nonetheless. -12 6J after 1 tick of slippage.