how to define a good or bad signal?

i am trying to count signal and make some statistics ,but i need a definition.
how do you judge a signal when you can see the future price?

You can figure out when to open a position based off those signals. Then you figure out when you might exit or trail etc. Hope this helps.

Good luck.

That really depends on what strategy you use. Usually strategies define what is a good signal.

There is no such thing as a bad signal, only a good one, unless you are in this business to lose money, in which case a bad signal becomes a good one.

Yes indeed, it's like me saying "sorry Cat I gotta go, I see a BAD signal forming, let me take it!" ...your right

Dear chineser,

What I think you are trying to do is calculate the outcome of following a strategy that generates a signal without a strategy. So, to determine if the generated signal is 'good' or 'bad' you need to define all of the parameters.

Example
Strategy: On Daily timeframe, Buy when price crosses 200 SMA and closes above. Close trade when price goes -20 pips in the wrong direction (20 pip SL).
Target Profit: 100 pips from entry.

So, lets say this example strategy gives you 100 signals over 1 year. Out of all of the signals 20 hit the 100 pip target (Good signal), while 80 hit the SL (Bad signal).

Now to calculate your statistics you can say that there was 20% good signals (20/100*100%) and 80% were bad signals (80/100*100%).
Going further you can calculate the P/L:
=>20 trades*100 pips = 2000 pips profit
=>80 trades*20 pips = -1600 pips loss
Therefore, this purely mechanical strategy has given you +400 pips profit over 1 year.

(Disclaimer: Just an example for educational purposes on how to do basic calculations on good or bad signal for the FF members, I have not tested it, nor do I encourage this strategy)

Kind Regards,
Tzamo

Some signals are false and I would say those can be considered bad signals.

Can you give an example of a false signal please?

You can't, without hindsight.

What he's suggesting is that the signal was followed by the price moving against the expected direction (as indicated by the signal). Therefore, that signal was 'bad' and should (could?) be treated as a false signal.


Now, some would suggest you need to review how/what generates your signals and modify or adjust to eliminate or reduce the number of 'false' signals. I would say the signals were not false, but only that the market didn't happen to follow through in a direction indicated by the signal. No signal is guaranteed and hence no 'perfect' way to trade in the market.

Judging a signal as "bad" is a psychological flaw - signals are neither good nor bad.

However, judging a signal as 'false' is an indication that you either don't want to accept the ability for the market to move against a 'good' signal sometimes, or that you need/want to adjust your signals to eliminate or remove them (which i don't think is actually possible).

For me there is no good or bad signal
The indicator is for me to make the best entry

Exactly. All this talk of good or bad signals can only be based on hindsight, because when we are live, and we place a trade, there has to be a reason for taking the trade, and this is the signal. It is not good, bad, or false, it is just there. Only once it is taken and the outcome of the trade has been arrived at, can we label it. Of the many thousands of trades I've taken, I've never taken a good, bad or false signal, just a signal. I hope that's sort of clear.

I don't think the hindsight negates the fact that a certain signal was not valid.

Of course it does! In the "moment" you are (or should be) basing your activity on what you see "at that point in time". You can't use your Tardis, jump forward in time, see the market move the 'wrong' way and then flip back and not trade that signal. It's just not possible.

You have to make a judgement that you've defined as being a signal for you to place a trade on as either long or short. Signal = trade. What the market does after that does NOT make the signal 'not valid'. That is completely incorrect. The valid signal is and was valid, unless it's based on one or more re-painting indicators that will change as they acquire more data (but I wouldn't trade on re-painting indicators).

If at a point in time a signal exists as being 'valid', then no forward activity or price movement can ever invalidate that signal. Yes, you can decide to change your trading rules to include some other factor or indicator to help you avoid whatever activity happened, but you can't say that the signal is now suddenly invalid when previously it was.

Example: So, what might happen on a completely valid double-cross of a 9, 50, 200 ema that signals a buy. You go long and then the market gets news of a tsunami and the market spikes the other way and then drops further for another 200-400 pips due to the knock-on affect of the aftermath.

What then? Was that original signal 'not valid'. Of course not. It's just that the market didn't move the way that would have benefited your trade (for whatever reason or reasons).

Signals indicate potential for movement in a certain direction and trade decisions are made on that. Same sequence for making decisions on Fundamental data or news, but the market may decide for other reasons to move in a different direction.

No signal can go 100% or 0%, but we can say that a signal is “good” if its statistics goes above 50% over a series of consistent-neutral trials ...

Don't confuse testing the % result of a system on its overall quality with the assessment of single trades being good or bad linked to it being a good or bad signal.

The system of which signals you will use for trading may be stronger or weaker / better or worse / greater or fewer % signals that actually 'worked out' in hindsight. Perhaps you want to say your system, or "a" system is good or bad over time. That's different.


However, on a one to one basis, if you get a signal you're about to trade, that "signal" in "that moment" MUST BE classed as a GOOD signal. Why?

Because you will now place a trade on it. It's a good signal, as all the parameters for you deciding to do a trade or not have been met. Or please think of it as a "correct" signal. Or a "rule matching" signal. I don't care.

Perhaps people are getting hung up on good or bad being terms that have too much psychological (and emotional) weighting.

Trades are not...repeat, not...good or bad. They either worked out to get you into profit at some point (even if you mess up the exit) or the market turns too quickly and moves against the trade. But if it turns, it still might turn again and work back into profit. Even if it moves into profit, it might turn and go against the trade.

In all those cases, you either get the result you want (profit, with pips) or a result you don't want (giving some pips back).

There are no good or bad signals. Signals must inherently be good (i.e., correct, should be traded, matching you rule, etc.), by definition.

There is only the market. Make it your signal.