Your explanation is false on the first part as not everyone "trades" and has liquidity issues. Furthermore, the 97% of the time is immaterial if the 3% of the time someone is taking the same size stake. And actually the reason why I see people taking extremely small sizes is because they suck at math in the theoretical examples they provided.

You then try to use some weird logical fallacy (some kind of appeal to authority + ad hominem) to defend this viewpoint which is both false and not in line with reality.

The final part does have some veracity. Now I can understand some of these having a strong effect on reducing position sizes. However, do you have a superior option? Every system I've seen may or may not address the first issue I posit below, and utterly fails the second. Considering you want to start from something that sucks and change it, instead of something that works and change it. I find that somewhat... interesting. Also considering that many people who run hedge funds buy into Kelly Criterion... I find this somewhat... entertaining.

Though I can't really see how any of this pertains to your previous question, nor does it really serve to answer any sort of question about 1) eliminating risk 2) maxing return. Why would you start with an inferior method of doing both in lieu of adjusting a theoretical system that fulfills both?

I did find the anecdotal stuff somewhat cute.

The liquidity concern is a real one and a lot depends on instruments traded and the timeframe employed. I think I remember Art (who trades the e-minis I believe) talking about how his strategy wasn't scalable because he trades on such a short timeframe.

The problem I have with Kelly is I don't see how to rectify the real possibility of a freakish market event occurring that could wipe me out. How would one modify Kelly in such a way so as to account for the possibility of freak market events (which would be difficult, since by their very definition they're unpredictable)? Using historical data is the logical solution, but is ultimately useless as unprecedented moves in the market happen much more frequently than would be assumed using a normal distribution.

In terms of my firm, every trader varies their position size quite a bit depending on their confidence in the trade and the liquidity situation. They're certainly using the principles of Kelly (bet more when you have the best of it), but modified for the specifics of trading. It seems like you have a problem with this ad-hoc approach to position sizing though.

Also, many of the hedge funds using Kelly (the relative value/pairs guys) blew up in 2008 as market conditions eliminated their supposedly quantifiable edge. When Genius Failed about the LTCM blowup illustrates the dangers of pressing thin edges based on historical market data and without a healthy fear of liquidity crunches.

Sorry for the ad homs by the way.

Meh. I could care less about your ad homs, they degrade you considering the reality of your arguments.

Regardless, so your firm is actually using some variant of Kelly staking? Whether making actual quantification of the exact positions or just kinda guesstimating with your intuition to achieve a certain Kelly fraction, those are still essentially variants of a KC based position sizing system. Beyond that you fail to address how other methods of position sizing satisfy the ability to prevent negative skew while Kelly criterion doesn't. Anything with a RoR is infinitely worse than KC, as you posed earlier. Yet, I fail to see one system that addresses both points I mentioned earlier in a superior way to KC.

Regardless, this argument is about whether you should use KC or not. Its whether you should be familiar with it if you want to claim that you know something about risk management. Yowserrs, for example. couldn't tell KC from his head or his ass, as he's exhibited several times. Perhaps he's rectified this error in his learning, or maybe not. I don't really care, but since he saw fit to interject himself. I'll use him as an example.

Also, your use of hedge funds is asinine, as no fund that has ever "blew up" used KC. Perhaps they loltarded the adjustments that take any theoretical model to practical application, but a 0% RoR exists with KC purely because you don't stake your entire BR. Then again from a personal wealth standpoint, playing boom/bust with a hedge fund may be a correct play. Regardless, the hedge fund itself, can't be using KC.

It depends what you mean by blew up, Kelly Criteria ensures you don't go busto but that isn't how a hedge fund operates, a 20% loss could be considered a reason to pull the plug. Kelly is certainly best with your own money though, pending your emotional capacity.

There is an interesting discussion of this. I don't think anyone who trades/bets/etc should ever be worried about getting another career. Making sure you "stay in the game" is just a BS smokescreen. I almost ground my roll down to the point where I needed to get a jube during the days to rebuild it. It would've sucked, obv. But whatever. Its just how it is.

There are bona fide reasons why an investor would choose NOT to invest at kelly ratio, even if he knew what his edge was.

There are bona fide reasons why an investor would be successful even if he did not know what his true edge was.

Who said I didnt?

Unless you started recently, you didn't even know wtf the RoR with KC was. So I'm gonna go with "lol".

Ending up at the correct conclusion is not the same as knowing how to get to that conclusion correctly, which is the whole point of the thread here. A mastery of risk management shouldn't just be knowing how to manage the risk for your specific subset of trades, but rather, how to make the decisions necessary to come to that decision.

Regardless, using KC is immaterial to the scope of this argument.

Perhaps this can be taken in a new direction. I've heard numerous people advise that they "Don't wanna be out of the game" or some variant. What is the overriding idea behind this? To me it just seems some sort of nitty way to chain yourself to hating money and not managing risk properly, the constraints of human capital limit your Kelly stake to the point where you always have the option to pursue another field and rebuild, why is this such a negative? Is it psychological or just wanton nittery?

I dont use KC because its less stressful to just have tiny draw downs and my income level is high enough where I dont need to worry about it.

From Wikipedia:

I heard no mention of which successful poker player would like to give up poker and pretty much the rest of his week to be a successful trader.

The skill set may have similarities but the schedule sure doesn't.

Whoever edited wiki is an idiot and it'll be corrected soon. The 2nd paragraph is entirely false, and the 4th paragraph is seriously misled as well.

lol internetz

cwar,

How do you know your drawdowns will be smaller?


ETA: The problem with the second paragraph is basic. An investment of your entire BR is against KC, using an example that hints at this is nth level ******ed. The problem with the final paragraph is that it doesn't understand that the first person has an infinite BR ergo an infinite stake.

% wise maybe larger but monetarily smaller which is what is important in terms of stress.

He who bets larger than his geometric mean expectation will be going busto !!!

In financial markets where you do not have a static betting field (unless you have a highly systemized short term system with large amounts of capital and predefined profit and loss targets) then using KC is not practical (eg long term trend traders never know an exact stat edge)..

whereas say in horse race gambling where you bet on finite fixed outcome events where at the end of each race you know what your "static bankroll" is , then using a fixed % of your roll per bet (KC) is purrfecccccccctttttttttt

Only an idiot would make a post like this that 1) botches major premises of KC 2) botches any sort of rational thought

i know you like to rub it in that youre a KC master but your 'pwns' or whatever gets you off are all getting very old.

either spell out some theories to bridge this discussion or stop posting already.

I know you like to pretend that you're an authority on trading and posited that traders are experts at risk management, but this thread has shown a plethora of "traders" that lack any sort of real clue what kelly criterion is. How is this acceptable in any way for someone who is suppose to be an expert at risk management?

And why do I need to spell out a theory? I'm simply stating a fact. Most traders are not experts over poker players at risk management. I think this thread has essentially proven my point as virtually every trader who commented (sans Art, but he didn't say anything really technical, and yourself who did likewise) either has no clue what KC is, or why it should be used as a starting point.

Thremp - I can post my track record going back to 1991 and every audited trade since 1998 with a significant sharpe ratio. I can quantify my edge and explain why I trade the risk metrics I use... can you........... you add NO VALUE here, you just seem like a negative little girl. you post no trading ideas, provide zero money management prowess, you are just a negative "fresh air" trader...

can you tell us the tick values of every major futures market, can you QUANTIFY your edge in the markets

you sound like a little biatch who trades fresh air

but I will give you a chance... I will put up my AUDITED daily records since 2004 and my P/L this year of 8 figures.... if you can do the same I will bow at your feet and rename myself ... IamThrempslittlebitch

so put up or shut up, I have traded quantitatively since 1991 and have had zero losing years since 1995........... what about you hero...

come on put up or stfu

lol trollz

Go ahead and post you audited figures BTW. I'm sure lots of people would like to see them

Thremp why don't we make a $100k bet that I can out do you as to daily sharpe ratio over a year.. or whatever risk metric you want to use

can you explain VAR(95)

do you understand and use Sortino

do you even trade you whiny little c.nt

lets do this for real, we can start with a small retail amount in a futures/securities account........ say $75,000

I am doing an experiment now which is verified... what do you say hero..

I obv don't trade, nor would I accept your asinine little challenge.

Why don't you just learn wtf KC is before you comment on it? Or since you seem to want to have an e-peen contest, post your audited account statements, then everyone can be like "wow ur so cool".

Thremp - I used to use EXACT KC criteria when I bet on horses as a 17-22 yr old

I was profitable but as per my previous post I eventually overbet and lost a lot

so please do not question traders like myself and Yowserrs

I read your little blog.... stick to sports betting where you think you have an edge

but feel free to quantify that edge for me

if not just crawl back into the little hole you came from because as you said

"I obv don't trade"

guess what, this is a trader's blog where people actually RISK real money, until then check out

http://forumserver.twoplustwo.com/30...thremp-652266/

I'm getting tired of this, either spell out the exact flaws and provide some content or quit posting this loop of 'lol you don't understand' while not linking any of what you say to trade specifically.

Until you prove the kelly works in a practical sense for trading, how can you keep the smug attitude?