"Expectancy" meaning on the Overview tab

QuantConnect Community Discussions

QUANTCONNECT COMMUNITY

Join Our Discord Channel

Join QuantConnect's Discord server for real-time support, where a vibrant community of traders and developers awaits to help you with any of your QuantConnect needs.

Draft Discussions

Bookmarked Discussions

"Expectancy" meaning on the Overview tab

Last comment by - December 2014

"Expectancy" meaning on the Overview tab

Query on "Expectancy" definition in QuantConnect's Overview tab. No documentation found.

Share New Research

Start New Discussion Sign up

SEARCH DISCUSSIONS

377,100 Quants.

Become a Quant

VOTE FOR UPCOMING FEATURES

Share your input and vote on our future direction.

LEAN Roadmap

JOIN OUR Community MAILING LIST

Create an account on QuantConnect for the latest community delivered to your inbox.

"Expectancy" meaning on the Overview tab

Nik milev | nik milev | December 2014

Does anybody know what the "Expectancy" definition is (on Overview tab of the results)?

I did not find it in the documentation.

Author

Nik milev

December 2014

Upvote

1 person upvoted this

Gustavo Avilés INVESTOR

QuantConnect | December 2014

Expectancy is used to estimate the likely outcome of a trades over time. If its positive the strategy will probably be profitable even if the loss rate is high. [tex]Expectancy = Probability Win * Size Win - Probability Loss * Size Loss[/tex] Link

The material on this website is provided for informational purposes only and does not constitute an offer to sell, a solicitation to buy, or a recommendation or endorsement for any security or strategy, nor does it constitute an offer to provide investment advisory services by QuantConnect. In addition, the material offers no opinion with respect to the suitability of any security or specific investment. QuantConnect makes no guarantees as to the accuracy or completeness of the views expressed in the website. The views are subject to change, and may have become unreliable for various reasons, including changes in market conditions or economic circumstances. All investments involve risk, including loss of principal. You should consult with an investment professional before making any investment decisions.

Upvote

Nik milev INVESTOR

December 2014

Agree - this is my understanding too, here is the result I have: Average Win 0.7% Win Rate 72% (0.7*.72 = 0.54) Average Loss-0.2% Loss Rate 28% (0.2*0.28 = 0.056) Expectancy 2.195 So this means that 2.195 would be achieved if instead of "%" for average win or loss is used actual value in points. Otherwise the expectancy (in %) should be: 0.54-0.056 = 0.484% Since price of securities/ETFs vary a lot, I would say that the Expectancy in points would give little information, but expectancy in "%" would be better. Expectancy in points would (in my opinion) make good sense when provided for for futures.

Upvote

Jared Broad INVESTOR

QuantConnect | December 2014

Yes I agree Nik it feels more sensible in percentages, but everywhere I see it explained is in dollar terms. Here is the line which calculates it - which I think is a slight refactor on the original formula. If you have a textbook/reference which shows it in percentages I'd be happy to change it.

Upvote

Nik milev INVESTOR

December 2014

I googled "trade expectancy in percent or point"and http://www.financial-spread-betting.com/Expectancy.html uses percentage in both multipliers. I agree that most sites use dollars as you mentioned. But this is because their explanation always starts with something like: "Assume you have a system that wins 40% of time $1000 and loses 60% of times %500 ..." in order to make the point that you can have a winning system that actually loses most of the time. Again for me $$ instead % makes more sense for futures, where you would trade e.g. 1 contract for every $10,000 you have in the account. Not sure if others feel the same though ...

Upvote

Jared Broad INVESTOR

QuantConnect | December 2014

With clean in-out trading its easy, but for portfolios where you sell 10 shares on share 100,000 holdings the percentage doesn't have any meaning - e.g. percentage of holdings? It might be cleaner to just remove this statistic entirely as its much more trading focused and less quantitative.

Upvote

Nik milev INVESTOR

December 2014

In case of multiple buys and sells that can't be matched, perhaps we can use the average price of the buys and still calculate the expectancy when sell is made and then use it to get the average. I think for the majority of the developed algorithms it will work, since buys and sells match. For those that it does not work, the average buy price could be used. And the same problem exists when you calculate avg winner/looser too (we would don't know the base price). Since this thread turned slightly to direction of suggestions, i would add few more: 1. The STD is great but does not dive the visual if gains/loses over the tested period are well or badly distributed. I would suggest to add to results the portfolio gains for each tested year. This would immediately tell me the yearly gains deviation is good or bad in a visual way. 2. Another important one I would add is number of bars the trade lasted. I would do this for winners and loser separately. This will give feedback if I am "cutting my losers quickly and letting winners run", so it gives ways to improve strategy in the right direction.

Upvote

1 person upvoted this

Spacetime INVESTOR

October 2021

Error:

ValueError : cannot convert float NaN to integer
at Update
wait_days = int(vola * self.BASE_RET)
===
at Python.Runtime.PyObject.Invoke(PyTuple args in dual_momentum_with_out_days_alpha.py: line 69

Fix:

    def Update(self, algorithm, _data):
        if algorithm.IsWarmingUp or not self.set_flag:
            return []
        self.set_flag = False
        insights = []
		
		#add
        history = self.history[self.MKT]
        if history is None:return
        vola = history.pct_change().std() * np.sqrt(252)
		
		#remove	
        #vola = self.history[self.MKT].pct_change().std() * np.sqrt(252)
        
        wait_days = int(vola * self.BASE_RET)
        
        .
        .
        .
+ Expand
- Collapse

Upvote

Vladimir INVESTOR

October 2021

Spacetime,

I was unable to reproduce the error you reported.

Have you had an error in the latest version 2.51?

Upvote

Spacetime INVESTOR

October 2021

apologies… should have been more clearer…

It is the framework algorithm shared above by Joao Antunes.

Upvote

Kamal G INVESTOR

October 2021

Hi Spacetime,

The cannot convert float to NAN integer occurs due to the multiple time resolutions being used for the same ticker. (Example below that will generate occasional errors). If you change all references to a given ticker to minute, hour, daily only, then the error will stop. (I got a similar error with other algorithms I have been putting together).

After you change a resolution, make sure the backtest is still the same or similar at the very least. Changing the time resolution can change an algorithm a lot!

self.STK2 = self.AddEquity('SPY', Resolution.Minute).Symbol

self.MKT = self.AddEquity('SPY', Resolution.Daily).Symbol

Upvote

Carsten INVESTOR

Carsten _ | December 2021

Vladimir

I was trying more pairs for the IN/Out and then giving each pair the same credit. If 80% of the total of 100% credits are reached, I would trigger the out signal. (It would be more like a voting approach)

This would be the opposite of this version were we have few pairs, but to trigger the signal all have to be satisfied. (It would be more like a veto approach)

What is you option on that, did you tried?

Second I would like to add some VIX pairs, like VIX > VIX 3M to trigger or a comparison of VIX momentum.

Did you tried something like that?

Upvote

Michael Rand INVESTOR

August 2022

Joao Antunes was there a reason why the AUD (FXA) - CHF (FXF) was dropped as an out signal from the original algorithm (in favor, I assume, of UUP and DBB)? I understand that UUP/FXA and FXF/DBB probably have similar properties from a market signal perspective.

Upvote

Platform

Radically Open-Source Algorithmic Trading Engine

Join Our Discord Channel

Quarterly Open-Source Trading Competition

Draft Discussions

Bookmarked Discussions

SEARCH DISCUSSIONS

377,100 Quants.

VOTE FOR UPCOMING FEATURES

JOIN OUR Community MAILING LIST

IN THIS RESEARCH

PARTICIPANTS

Actions

Join QuantConnect for Free

Platform

SIGN IN

Radically Open-Source Algorithmic Trading Engine

Join Our Discord Channel

Quarterly Open-Source Trading Competition

Draft Discussions

Bookmarked Discussions

SEARCH DISCUSSIONS

377,100 Quants.

VOTE FOR UPCOMING FEATURES

JOIN OUR Community MAILING LIST

IN THIS RESEARCH

PARTICIPANTS

SHARE RESEARCH

SHARE DISCUSSION

SHARE ARTICLE

SHARE

Actions

Join QuantConnect for Free