Anyone who thinks that he can calculate such a probability with meaningful confidence or numeric precision doesn't know what he doesn't know. Be careful with that stuff and what people tell you.
probability
- Thread starter Andy_Trade
- Start date
Delta is the probability that a particular option will expire ITM.
Once you know your break-even point for a position, you can use the imaginary delta at that price to calculate the probability of break-even or better/worse.
For example, suppose you write a 55-60 vertical bull spread. Your credit is $1.50 so your break-even point is $58.50. The delta of an imaginary 58.50 call is the probability the stock will be above $58.50 and is therefore the probability that you will profit.
You can guesstimate the delta of a 58.50 call by looking at the deltas of the 55 and the 60 and picking a number toward the "60" end of the range.
Obviously, this trick breaks for things like calendars and diagonals, but it's handy for basic directional plays.
There are high probability trades, and there are high reward/risk trades. Although people tend to not distinguish between them.
If GOOG was trading, bid 670, ask 670.30, and you placed an order to buy(cover) at 670 and sell(short) at 670.30 (scalping), and able to sit through the end of the day waiting for both trades to take place, there is a high probability the trade will happen (giving you a tiny profit). Is it high reward/risk? I doubt. In those cases where the closing leg fails to execute(say 15% of cases) you could incur substantial losses.
I believe what you are asking is about high reward/risk trades. It means if you take a sample of such trade cases, your average gainer will be far greater than your average loser. This is dependent on your ability, but yes, based on experience with many strategies, it is certainly possible to tag certain set-ups as having huge reward/risk ratio. You may not trade only those set-ups because they do not occur all the time.
heres for GOOG with data from the ivolatility website.
atm_vol(31 days to exp) = 36.05 annualized
so convert vol to daily with vol/sqrt(256/days) = 12.745
price = 660.52
make a distribution curve using price as mean and daily vol as stddev
no use cdf() and pdf() function s to find out prob of greater than, between , less than what ever.
in this case the prob of greater the 660.52 is 50%. greater than 670 is 22.8%. etc...
Nope. I understand reward risk ratio but I'm not sure about this "high probability" stuff that I hear people talk about.
What good is a 3:1 reward/risk ratio if the win/loss is 1:5?
The reward/risk ratio takes into account the win ratio. If the average gain:average loss is 3:1 and yet the win:loss ratio is 1:5, I think that means the reward:risk ratio is actually 3:5, which is not good. Hope this corrects my earlier post. When people talk of reward/risk, they are normally thinking of upside/downside with an equal probability weighting.
Argh, I just noticed it's an options thread. But your question sounds more like a basic trading probability question.
If so, here's some background.
Go back to basics. If you flip a coin a thousand times and you get 500 heads, what is the probability of getting heads?
P(h) = number of heads/ total flips = 50%.
That is the probability of getting heads in the long run. Regarding stocks, you can say if you had 1000 trades and 500 were winners, your probability of wins is, P(w) = 500/1000 = 50%.
The problem is, this doesn't say much by itself. It doesn't tell you how many winners you will have in a row (sequential), nor does it tell you whether it's a good result. You could have 900 winners out of 1000, with a 90% probability of winning, and the 901st trade could bankrupt you if you bet it all.
When people start talking about distributions, it's because the market is not binary like a coin toss (only 2 outcomes head or tail, winner or loser).
You might ask what is the likelyhood of a 3% loss day. That is where you run a histogram of all the daily historical ranges, and if the distribution looks normal, you can get a rough feel for probabilities or likelyhoods of a 3% downday using properties of a normal distribution (obviously real markets aren't exactly normal, but it's a close approximation). There are numerous books that look at what type of probabilities are useful in trading and backtesting. Win/Loss and Risk/Reward are other metrics, whereby you generally want greater reward/risk trades, so that in the long run, you will be up over time. Then there are other things to look at like position sizing and optimal position bets to make, these also use probability formulas as input
(look up kelly and optimal f position sizing).
Probability is a useful tool in backtesting and looking for edges, but there is more to look at then simple probability of winning trades. Go get a good book on systems testing, like
http://www.amazon.com/Trading-Syste...bs_sr_1?ie=UTF8&s=books&qid=1195843594&sr=8-1
if you want to start understanding more about what people are referring to when discussing probability in trading.
Regarding options, probability is generally used to determine future price ranges approaching expiration by looking at past historical volatility and extracting probabilites about how far the price will move in the future. However, the market prices in "implied volatility" which is a whole different dimension/discussion, which adds to that model but is more difficult to predict. Ultimately, you are taking in all of the past volatility information and trying to use probability to determine how likely your trade will be profitable.
If you are more interested in probability applied towards options try:
http://www.amazon.com/Mathematics-O...bs_sr_1?ie=UTF8&s=books&qid=1195844764&sr=8-1
Both books are geared a bit more towards the mathematical/quantitative side if that's your thing.
If so, here's some background.
Go back to basics. If you flip a coin a thousand times and you get 500 heads, what is the probability of getting heads?
P(h) = number of heads/ total flips = 50%.
That is the probability of getting heads in the long run. Regarding stocks, you can say if you had 1000 trades and 500 were winners, your probability of wins is, P(w) = 500/1000 = 50%.
The problem is, this doesn't say much by itself. It doesn't tell you how many winners you will have in a row (sequential), nor does it tell you whether it's a good result. You could have 900 winners out of 1000, with a 90% probability of winning, and the 901st trade could bankrupt you if you bet it all.
When people start talking about distributions, it's because the market is not binary like a coin toss (only 2 outcomes head or tail, winner or loser).
You might ask what is the likelyhood of a 3% loss day. That is where you run a histogram of all the daily historical ranges, and if the distribution looks normal, you can get a rough feel for probabilities or likelyhoods of a 3% downday using properties of a normal distribution (obviously real markets aren't exactly normal, but it's a close approximation). There are numerous books that look at what type of probabilities are useful in trading and backtesting. Win/Loss and Risk/Reward are other metrics, whereby you generally want greater reward/risk trades, so that in the long run, you will be up over time. Then there are other things to look at like position sizing and optimal position bets to make, these also use probability formulas as input
(look up kelly and optimal f position sizing).
Probability is a useful tool in backtesting and looking for edges, but there is more to look at then simple probability of winning trades. Go get a good book on systems testing, like
http://www.amazon.com/Trading-Syste...bs_sr_1?ie=UTF8&s=books&qid=1195843594&sr=8-1
if you want to start understanding more about what people are referring to when discussing probability in trading.
Regarding options, probability is generally used to determine future price ranges approaching expiration by looking at past historical volatility and extracting probabilites about how far the price will move in the future. However, the market prices in "implied volatility" which is a whole different dimension/discussion, which adds to that model but is more difficult to predict. Ultimately, you are taking in all of the past volatility information and trying to use probability to determine how likely your trade will be profitable.
If you are more interested in probability applied towards options try:
http://www.amazon.com/Mathematics-O...bs_sr_1?ie=UTF8&s=books&qid=1195844764&sr=8-1
Both books are geared a bit more towards the mathematical/quantitative side if that's your thing.
Chart patterns can be important in finding high probablility trades. Dbl tops, dbl bottoms, head & shoulders, etc, etc.
If a market has moved up and has stalled, short it. If a move down has stalled, buy it.
BUT ALWAYS USE A STOP IN CASE YOU ARE WRONG.
For new traders this game is about surviving to fight another day. IMHO more important than finding high probability trades is having a system to minimize your losses and sticking to it.
If a market has moved up and has stalled, short it. If a move down has stalled, buy it.
BUT ALWAYS USE A STOP IN CASE YOU ARE WRONG.
For new traders this game is about surviving to fight another day. IMHO more important than finding high probability trades is having a system to minimize your losses and sticking to it.
I think we need to distinguish between option positions that are statistically high-probability trades per se (e.g. DOTM writes and spreads, wide condors) and trades that are structured to be high-probability based on your outlook for the stock (e.g. strike selection for butterflies, condors and ratio writes), whether that outlook comes from charting, fundamental analysis, earnings/news expectation or whatever.
I'm curious as to which one the original poster meant, and whether he can provide an example of a situation where he would like to place a high-probability trade.
Could not agree more. There's no sense in winning 90% of the time if your 10% losers wipe out all the profits.
Incidentally, the same could be said for experienced traders.