I don't know what version of Kelly you're using but here's what mine says: (.75*(5) + .25*(-1))/(.75*(25) + .25*(1)) = .1842 ~ 18% Round down to 10% or 15% for a margin of safety and that's a far cry from 70%+.
i bet using the amount of theories concurrently in play * 20k or 10k the more positive trsding theories in play the more you bet. the less in play the less you bet. the more theories in play the higher it will play out. playing out is more important the risk reward. a cup handle pattern will initislly get 10k cuphandle+52weekhigh+panicdrop=30k cup+52+panic+lowofrange+marketdown=60k market direction has a + or - 40%. a drop gets +20 a rally gets -20. a drop w/rally gets +40. a drop w/ no rally gets -40 this last sentence is actually the most important. enjoy. ======= when about 13 out of 13 theories are in play. bet size * margin2x or 3x
At 20% max draw-down, the magic number is ~6.5%. Chance of 3 losers in a row = 1.5% 3 losers in a row = 19.5% capital loss. But this is really just theoretical, since average probability of loss is not uniform across conditions. Range days for trend strategies will produce far more losses than average loss percentage. Trend days for range strategies will produce equally more losses than expected. So better to identify those conditions under which the strategy incurs maximum losses, recalculate average probability of loss under those conditions, then recalculate maximum risk per trade.
bet size is important. If you bet with stop 2% and you risk 200K, that means position is 10M . Do not underestimate willingness of the market to visit your stop price. So, if trading SP or EUUS, this is maybe ok i guess. but anything less, you may be scheduled for slaughter. But you dont know that until you are in the market with that kind of bet.
I don't know what version of Kelly you're using but here's what mine says: (.75*(5) + .25*(-1))/(.75*(25) + .25*(1)) = .1842 ~ 18% Round down to 10% or 15% for a margin of safety and that's a far cry from 70%+. [/QUOTE] I'm pretty sure that Kelly is: %Win - ((1-%Win)/(W/L Ratio)) .75 - (.25/5) = .7 So, the initial poster was correct. In the paper I linked above, the Kelly bet size for a strategy with a 57% win rate and a 1-1 payout ratio is 14% (.57-(.43/1), which is close to the 18% you calculated for a 75% win rate/5-1 payout ratio strategy, which would imply that the slope of the Kelly formula would have to be extremely small for a great strategy to only generate a bet size 4% larger than a much less impressive strategy.
I agree you should bet big on slam dunks. The question is how big to bet, given a certain maximum drawdown tolerance. If your tolerance is 50%, for example, versus my 20%, then the optimal fraction is still the same in proportion to our risk tolerance - you will just bet 2.5 times as much as me. But we still need to know what the optimal size is for a given drawdown expectation. "Had I lost the bet I would not have kept those risk limits of course, I would have cut down materially" This seems questionable though. Why cut down the size (as a % of your remaining capital), assuming the next trade is independent of the first one and has identical expectation? If 20% of remaining capital is a rational bet size for your initial slam dunk setup, why is it irrational 1 trade later just because you got unlucky the first time?
I think one additional piece of input would round out the equation in a way that will lead to a more useful answer. Let's say that my trading produces very steady and acceptable returns, provides for a lifestyle that I enjoy and my equity curve heads north in a fashion that shows me I am on a path that might be called "easy street". Under those set of circumstances I would say the great opportunity you describe -- and the numbers are lovely -- should not prove overly seductive. Because the risk of ruin is, in this instance, risking such a bedrock part of what contributes to my happiness I should be loathed to take any significant risk that I will endanger that happiness. If I am producing the result that makes me secure and happy risking no more than 1% of my net worth on any given trade maybe I should be spending my time and energy looking for a way to do that while risking only .8% of my net worth rather than looking for a way to "pump it up" faster. If I have already won maybe reducing risk has become my full time job. If, on the other hand, I am at the point in life where I am struggling to reach the comfort and security I describe above -- and I am there -- or I am dead set on being in the George Soros class then it might be realistic to say: "Hey, I'm unlikely to find a better vehicle than this trade/setup so it is on these fine numbers that I will take a 15% or 20% risk of ruin." As in many things in life where you stand on the issue depends on how well you sit in life. The job, for many of us, is to get to a very fine seat and then find ways protect that position from inflation, depression and, most of all, the insanity of the coming storm whatever that storm may be. The only sure bet is that those in charge will continue this lemming-like run into chaos. You can bet your life on that because they already have.
Kelly formula: Bet size = ((win rate * payoff odds) - (1- win rate))/payoff odds. So ((.75*5) - (.25))/5 = 0.7 i.e. 70% of capital.
Because the fact that the trade was a loser counts as evidence that your judgement is correct when making discretionary trades. Also it gets you closer to your drawdown uncle point