Kelly-betting for FX trading

#21
Thunder said:
Well if you have $1000 let's say, in that scenario, you'd want to risk 40% of your bankroll so that the maximum loss you'd have is $400 while the max gain would be $80. This would be full Kelly. Your bankroll will grow at a rate of 1.82% per trade. The difference between this and having a trading system where you win 70% of the time and lose 30% of the time (but the loss amount equals the win amounts) is that in the 70/30 trading system, you don't have these huge drawdowns which essentially kill any growth in your bankroll.
Say what? z:confused:g
 

Thunder

Well-Known Member
#23
blackjack avenger said:
One should be able to overcome variance in a tough game/trade if they bet conservatively. In Thunders $400 bet example, make it $100 or $40.
With all due respect, betting more conservatively won't overcome variance. What it will do is lower the risk of ruin at a cost to the growth of your bankroll.
 
#25
overstayef

Thunder said:
With all due respect, betting more conservatively won't overcome variance. What it will do is lower the risk of ruin at a cost to the growth of your bankroll.
Agree "overcome variance" was to strong a term, perhaps manage is better. I also agree with your statement.
However,
Bank growth "Kelly" should not be the only consideration. You, yourself recommended 1/3 Kelly. In this example of high variance & uncertain edge (as Zengrifter keeps needing lol) betting conservatively can greatly increase the probability of growth, even if we overestimate our edge.

In bj Kelly long run numbers are staggering! As I am sure you know
 

Thunder

Well-Known Member
#29
You'll have to ask him. I would assume he has a decent sample size to be quoting those figures. You ask a very good question when you ask what the SD is. I don't think Kelly takes that into account when giving you the proper bet size and that can be important when utilizing your average loss to come up with a Kelly bet. If there is a way for Kelly to take into account the SD for losses, please let me know,

Case in point. Say when someone has an average loss of -10% per losing trade, are the losses all under -20% or are there some losses up to -60%? As I mentioned before, someone who has a small SD for their losses, is going to do much better than someone with a big SD. Personally if I knew the SD for my losses was quite big, I'd adjust my Kelly bet accordingly. Another thing that I've done is I've taken all my trades and experimented in Excel with how well I would have done had I invested a bigger % or smaller % of my bankroll per trade. After a large sample size, that can give you a good idea of what to invest or bet.
 
#30
Sample size

matt21 said:
Hey DDutton, I am assuming that I know what the +EV is. I need to do that in order to ask the question. You are of course right in pointing out that with the FX markets the probabilities are not certain as they are in casino games.
Hi Matt,

Don't be offended by comments, but it sounds to me like you're not a trader by profession. You didn't mention the sample size you use to determine EV, but I'm guessing it may not be large enough.

You probably are aware that many people do use Kelly in the markets, and that Ed Thorp has used it to make billions with his hedge fund - he practically invented the hedge fund concept. But you can believe that he also did a vast amount of research to determine his EV, or he wouldn't have made billions.

If you are interested in a serious answer to this question, I would put it to Don Schlesinger, who was a derivatives trader for Morgan Stanley, and now runs a training program for their traders. He's also a friend of Ed Thorp, has been an advisor to many BJ teams, and can probably give you the best answer to this question that you're going to get. You can reach him at his site, advantageplayer.com, or the GC page at BJ21, and possibly even their Free page, although I rarely check it.

Cheers,

BJ Warrior
 
#31
BJ Warrior said:
Hi Matt,

If you are interested in a serious answer to this question, I would put it to Don Schlesinger, who was a derivatives trader for Morgan Stanley, and now runs a training program for their traders. He's also a friend of Ed Thorp, has been an advisor to many BJ teams, and can probably give you the best answer to this question that you're going to get. You can reach him at his site, advantageplayer.com, or the GC page at BJ21, and possibly even their Free page, although I rarely check it.

Cheers, BJ Warrior
Matt, after Don weighs in you can get another opinion by going to blackjackforumonline.com and click the Fight Club forum and share with the
moderator Stalker what Don said and then get Stalker's take. Then you'll have a full picture. Then come back here and share the wisdom!! zg
 

Thunder

Well-Known Member
#32
zengrifter said:
Matt, after Don weighs in you can get another opinion by going to blackjackforumonline.com and click the Fight Club forum and share with the
moderator Stalker what Don said and then get Stalker's take. Then you'll have a full picture. Then come back here and share the wisdom!! zg
I had no idea Don traded derivatives as well. I'm going to have to get in touch with him. Is Stalker also a professional trader?
 
#33
Don't contact Stalker

zengrifter said:
Matt, after Don weighs in you can get another opinion by going to blackjackforumonline.com and click the Fight Club forum and share with the
moderator Stalker what Don said and then get Stalker's take. Then you'll have a full picture. Then come back here and share the wisdom!! zg
Hi Matt and Zengrifter,

Just wanted to give Matt a headsup. I'm assuming this is a Zengrifter joke - we know he has a wicked sense of humor, to put it mildly.

Stalker is a former (and possibly present) pro player who has an obssessive hatred of Don Schlesinger that goes back about a decade, although I don't think they've had any recent contretemps. Stalker was a very successful pro player who lost most of his stake when it was confiscated a few years ago by the police during his travels, and he humiliatingly had to be rescued by some members of the BJ community. I'm not familiar with his current activities. He hates Don because he's regarded as a BJ authority, although he's never made a living as a pro. Of course neither did Ed Thorp, Julian Braun, or a number of others who have made contributions to the game. His obssession with Don suggests someone with some major issues.

I'm sure you're also familiar with Zengrifter's interview here, and thus his differences with Don, who labeled ZG's questionable loose-index play 'sloppy'. So he's not the most objective source of information on Don.
But Don has never spent a day in prison, which may disqualify him from giving advice in some circles. :)

Cheers,

BJ Warrior
 

matt21

Well-Known Member
#34
ZG, Thunder, BJ Warrior and Blackjack Avenger - thanks for all your posts and comments - that's much appreciated.

Right now, the FX trading is not providing my living - the casino games are doing that - however I am more or less working full-time on fx and equities trading, so I like to think that it is just a matter of time before the fx trading will become my primary source of income - in total I have trading experience of the equivalent 3-4 years full-time trading, in various shapes and forms. My trading has been in profit for the past 12 months, and my bankroll is big enough to provide the potential to live off trading profits.

In terms of the question that I posed - I do not have statistically significant sample sizes for the win rates and risk-reward ratios. I understand the limitation that this imposes, and I understand that this is clearly different from casino games where the game scenarios can be mathematically modelled with a high degree of certainty. However I am not concerned with whether my EV rates are in fact correct, I am happy to make assumptions in this regard. What I am more interested in is that, assuming certain win rates and risk-reward ratios, how much should I risk per trade from a fractional Kelly perspective for high-win-rate-strategies where the average win size is less than the average loss size?

I will try to contact Don Schlesinger with this question - thanks for that suggestion.

My gut feel at this point is that, in the example that I raised, and that Thunder discussed, I should risk 8%BR rather than 40%BR assuming full-Kelly. In practice I will not risk more than 2.5% per trade for strategies that i consider established, and 1% for strategies that seem to be profitable but that I have traded for less than 4 months. So my question essentially comes down to whether for the setups with presumed high win rates and presumed low payours, my risk would be a max of 2.5%, rather than my target profit being 2.5% (and my risk being much greater than 2.5%). I am confident that risking 2.5% on a trade will be conservative from a Kelly perspective.

I hope that clarifies things a little. And thanks again for all your comments so far! :)

Happy trading & happy gambling,
Matt21
 
#35
past performance

Aren't trading strategies back tested. Doesn't this supposedly give sample size?

Past performance is not indicative of future results. The future is unknown.
 

Thunder

Well-Known Member
#36
blackjack avenger said:
Aren't trading strategies back tested. Doesn't this supposedly give sample size?

Past performance is not indicative of future results. The future is unknown.
The sun has risen for billions of years (past performance) Should I not expect it to rise today? (future result) :) :devil:
 

matt21

Well-Known Member
#37
BJ Warrior said:
If you are interested in a serious answer to this question, I would put it to Don Schlesinger, who was a derivatives trader for Morgan Stanley, and now runs a training program for their traders. He's also a friend of Ed Thorp, has been an advisor to many BJ teams, and can probably give you the best answer to this question that you're going to get. You can reach him at his site, advantageplayer.com, or the GC page at BJ21, and possibly even their Free page, although I rarely check it.
BJ Warrior
I managed to get in contact with Don Schlesinger and he was very prompt and very detailed in his response, which was great!

His answer was in fact that using the formula of Edge% divided by Pay-off holds true even where the typical winning trade is less than the typical losing trade, such that if one say had a 5% edge and the average winner is 0.7 times the average losers, then the correct half-Kelly bet would be:
0.5x (0.05/0.7)
His answer did surprise me since that just didn't feel right to me.

This makes for a very interesting insight. It implies that one strategy can have a superior EV$ to another strategy despite both strategies having the same EV%. For the same ROR, the trader will be able to wager more per trade for the strategy with the higher win-rate, and thus achieve a higher dollar EV per trade. Please shoot me down if you think I am wrong here! If I am right then this insight will have a drastic impact on my potential earnings!

He explained that larger bets were warranted because the likelihood of having larger losing streaks decreases significantly for strategies with high win rates.

He also cautioned that one should always err on the side of caution in estimating EV on trading strategies. He also pointed me to some other material on the Kelly principle's application to the financial markets.
 
#39
matt21 said:
This makes for a very interesting insight. It implies that one strategy can have a superior EV$ to another strategy despite both strategies having the same EV%. For the same ROR, the trader will be able to wager more per trade for the strategy with the higher win-rate, and thus achieve a higher dollar EV per trade. Please shoot me down if you think I am wrong here! If I am right then this insight will have a drastic impact on my potential earnings!
Not sure what you mean:
I'n bj we can bet based on each advantage, bet ramp.
Or
Bet based on the average advantage, flat bet.
The bet ramp is more efficient.

A higher advantage we can bet more

If we play/win faster our velocity of winning improves, though playing faster we cannot bet more until we actually win.
 

matt21

Well-Known Member
#40
blackjack avenger said:
Not sure what you mean:
I'n bj we can bet based on each advantage, bet ramp.
Or
Bet based on the average advantage, flat bet.
The bet ramp is more efficient.

A higher advantage we can bet more

If we play/win faster our velocity of winning improves, though playing faster we cannot bet more until we actually win.
Hey Avenger - I was making that italic statement in reference to my fx trading, rather than my BJ play. With this insight I will be able to risk more dollars per trade then I previously thought. Prior to this insight I would risk $100 per trade, regardless of what my reward-risk ratio was. Following the insight, I can risk more than $100 per trade if the r-r ratio is less than 1.0.

With a fixed EV%, the more $ I can risk per trade, the better. Does that make sense?

As far as betting for my AP stuff - I hope I have that sorted for now. In any case, most of the time the real constraint in this situation is heat rather than risk of ruin.

Matt21
 
Top