do these 2 posts conflict?

Gordon Gekko · Sep 11, 2002

Originally posted by nitro
If something is 50/50 - no amount of anything (money management, stop loss, etc) over the long term will turn that into positve expentancy over many trials.

50/50 means 50/50 - it is atomic.

nitro

Originally posted by uptik2000
trader A makes 10 trades on Monday. 1000share blocks.

7 of those trades are losers, but because traderA kept to his stops , each loser is only 1 point. So he's down $7,000.

3 of the trades are winners. And because traderA lets his winners run to the target each winner gets him 5 points. Yielding him $15,000.

traderA has just made $8,000 on the day despite only having a 30% win percentage...

if i misunderstood either of you guys, i apologize. but here is my question:

nitro is saying that if something is 50/50, no matter how you manage it, you can't get a positive expectancy out of it.

uptik2000 is saying, even with LESS odds than 50/50 (30% win rate), you can have a very good expectancy. although uptik2000 is cutting his losers and letting his profits run, won't his # of losers just occur more frequently?

now, i've read the van tharp book and i agree with it, but i also see points to what nitro is saying. if something is 50/50, how can you get a positive expectancy out of it? to see my point, look at the following two other posts:

Originally posted by profitseer
Just one thought on a 50/50 trade. Lets say it's at 50 and there is a 50% chance it will go to 40 and a 50% chance it will go to 60. If you guess long and your wrong, you aint gonna let it go to 40 with you on board. Maybe you will just ride it down to 48. But if you guess right and it goes up, you ride it the full expectency of the 50/50 trade you started out with.

So it's 10r to 2r on a 50/50 trade.

Originally posted by Mr Subliminal
It was a 50/50 trade with risk:reward of 1:1 till you started meddling with it. Your "premature" exit at 48 doesn't fool the Numbers Deity and will occur with a frequency that is 5 times larger than your exit at 60, putting you back to square one with your original 1:1 risk:reward.

i don't want to sound like a broken record, but this is what i don't get. if you decrease the limit of your losers, won't the rate at which they occur increase proportionately? so, won't you just go nowhere?

metooxx · Sep 11, 2002

Originally posted by Gordon Gekko
i don't want to sound like a broken record, but this is what i don't get. if you decrease the limit of your losers, won't the rate at which they occur increase proportionately? so, won't you just go nowhere?

No.

Gordon Gekko · Sep 11, 2002

Originally posted by metooxx
No.

metooxx, i know you like to give responses < 5 words

, but it would be more helpful to me if you also stated WHY your answer is no. thank you

J_Commisso · Sep 11, 2002

Gordon,

Are odds and probability the same thing???

PEACE and good trading,
Commisso

nitro · Sep 11, 2002

Originally posted by Gordon Gekko

if i misunderstood either of you guys, i apologize. but here is my question:

nitro is saying that if something is 50/50, no matter how you manage it, you can't get a positive expectancy out of it.

uptik2000 is saying, even with LESS odds than 50/50 (30% win rate), you can have a very good expectancy. although uptik2000 is cutting his losers and letting his profits run, won't his # of losers just occur more frequently?

now, i've read the van tharp book and i agree with it, but i also see points to what nitro is saying. if something is 50/50, how can you get a positive expectancy out of it? to see my point, look at the following two other posts:

i don't want to sound like a broken record, but this is what i don't get. if you decrease the limit of your losers, won't the rate at which they occur increase proportionately? so, won't you just go nowhere?

Gordon,

You have to learn how to write a simple program - simulations will often give you mathematical instinct where there was none to be gotten, or where we do not have the mathematical background to figure it out. If you do not know how to write a program, flip a coin a thousand times. See if through money management you are somehow able to come out ahead by betting on the outcome of coin flips.

BTW, don't think this is stupid - I know some University students that have flipped a million coins in their lifetimes to get intuition...

As far as uptik's observation, he is 100% correct. There is no contradiction.

The mistake you are making is assuming that because something has a "win" rate of less then 50%, then therefore the underlying process is random. This is false. The key difference is that markets have "short term memories," where in a coin flip or in any truly random event, the outcome of the last trial has no effect on the current trial.

nitro

m_c_a98 · Sep 11, 2002

Two little examples:
-------------------------------------------
100 trades: 50 winners 50 losers
average win = 2
average loss = 1
YOU HAVE POSITIVE EXPECTANCY

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

100 trades: 40 winners 60 losers
average win = 5
average loss 1
YOU HAVE POSITIVE EXPECTANCY(and only right 40% of the time)
------------------------------------------------

There is more than one way to skin a cat!

so step one is you need a positive expectancy or stated another way, you need a positive average trade over the course of all trades. Pyschologically to be able to trade a system or method you may feel it easier if you have a high winning percentage but this is not necessary.

then step two is once you are confident the system has a positive average trade, then you need to use the appropriate betsizing(position sizing) algorithms such as Fixed fractional or Fixed ratio.

metooxx · Sep 11, 2002

Originally posted by Gordon Gekko

metooxx, i know you like to give responses < 5 words , but it would be more helpful to me if you also stated WHY your answer is no. thank you

We trade 20+ systems; with the average performance of say, 40% winners, 10% losers and 50% scratches. Since there are no scratches in the real world, i.e. scratches pay commisions therefor they are losers; the stats are 40% winners with 60% losers. The average win is static, however the more losers you move into the scratch column the lower the average loser.

Is that what you wanted to know?

J_Commisso · Sep 11, 2002

Gordon,

You can't make money betting on probabilities of 50/50 with odds of 1to1, but you better believe that you can make money betting on a probability of 40% with 5to1 odds...

In trading the higher the odds usually the lower the probability, but the two without each other mean NOTHING! It is the mathematical expectation of the two that matters...

Im sorry if this comes off harsh, but how could you have been trading for three years and not understand what Nitro and Uptik were saying???

PEACE and goodtrading,
Commisso

nitro · Sep 11, 2002

Originally posted by Gordon Gekko

...i don't want to sound like a broken record, but this is what i don't get. if you decrease the limit of your losers, won't the rate at which they occur increase proportionately? so, won't you just go nowhere? ...
[/B]

Gordon,

You are quite correct in this observation - in "most" trading, you have to allow room for a trade to work, otherwise you may get "wiggled" out of a trade. The key is that when you get a signal to enter the market, it is of enough "intensity" where the odds of it going against you to your "pain threshold" is less then the odds of it going in your direction.

Obviously, scalpers do this by reading the tape and trying to get a feel for the supply/demand in a stock - and you will see a lot of scratch trades there or small loses. But if they are good at it, they scratch say 50 trades, lose 20, and win 30 - those extra ten winners out of a 100 on 2000 shares is their living[NOTE the importance of a low commission rate tho!]

As you move up the time frame, something similar, but at the same time very dissimilar happens...

nitro

metooxx · Sep 11, 2002

Originally posted by nitro

Gordon,

You are quite correct in this observation - in "most" trading, you have to allow room for a trade to work, and this will cause you to get wiggled out of a trade. The key is that when you get a signal to enter the market, it is of enough "intensity" where the odds of it going against you to your "pain threshold" is less then the odds of it going in your direction.

Obviously, scalpers do this by reading the tape and trying to get a feel for the supply/demand in a stock - and you will see a lot of scratch trades there or small loses. But if they are good at it, they scratch say 50 trades, lose 20, and win 30 - those extra ten winners out of a 100 on 2000 shares is their living[NOTE the importance of a low commission rate tho!]

As you move up the time frame, something similar, but at the same time very dissimilar happens...

nitro

Commisions are the key ...