i semi-recently discovered intraday vwap as a support/resistance method...and its been amazing. I was resistant to exploring how to use vwap in the past because i was like "everybody knows about that...so it can't have value"...but clearly i was being an idiot...because vwap rocks if used correctly.
i strongly suggest you people explore ways to integrate it into your trading...free advice...take as is
We run analytics when we price for trades. Generally, a VWAP and TWAP (along with various other slicing and dicing approaches) comparison is used. Doesn't really make sense to consider VWAP your "support and resistance" IMO, though I do agree that it is useful as a benchmark of price. Most large trading desks will run these analytics for buy-side traders, but we add a probabilistic component to size the trade in the tick. This kind of stuff does work out in the long run to eek out a better price intraday (better arrival px), but I'm not sure if I would actually use it for entry/exit. It seems inherently risky because you do not know who you are trading against. Just my thoughts.
We run scenarios for multiple execution types and venues to try to get the "best price", and it includes a TWAP as an alternative benchmark to VWAP. I agree in that I also don't think anyone really uses it for actual execution.
Going back to my comments on VWAP (I assume VWAPs are weighted moving in general), I agree with you in its usefulness as a benchmark, but just that I wouldn't use it for s&r. There ae better tools for that, like super-imposing volume horizontally at price, but even then you'd need to have a theory on who you are trading against for that to work multi-period. Are you just day trading or are you using price and volume as a signal?
"super-imposing volume horizontally at price" What exactly is this? Are you referring to volume profile of each tick?
It's a bit of both.I trade intra-day distortions when putting on prop trades, outside of the market making business... so if I see a price that is offset, I would use VWAP as a second measure of validation to see how much the price deviated within a stretch of certain confidence interval. I'm actually testing this out. I've never used this so iron chef recommended it to me. I think as a trader, you have to keep an open mind, test it, validate it and most importantly be able to explain why it works. Too many wanna be traders put on random bullshit without explaining the why and the original purpose behind why it was created in the first place.
Ok that makes sense. For that kind of stuff (discretionary legs based upon price movement), we will calculate the average move (variance, std) for the period (tick, minute, hour, day, week, month, etc.), subsequent returns, momentum, in order to find the probability of profit. This is actually not hard to do, just needs a solid excel sheet and a data feed/excel plugin.
For super-imposing volume, lots of software allow you to add volume to your price chart, but to flip the axis. I'm sure you have used/seen it before-- it's more useful than a VWAP to find price at volume spikes ("support and resistance").
Thanks for the info.
"For that kind of stuff (discretionary legs based upon price movement), we will calculate the average move (variance, std) for the period (tick, minute, hour, day, week, month, etc.), subsequent returns, momentum, in order to find the probability of profit."
So is it a running calculation on your excel? Also if you're calculating the variance, std, etc... What are you wrapping it around? The price itself? And if so, how far back do you look? So when you said tick, how far back in ticks? I'm guessing this is subjective?
Yes you can set it as a running calculation in excel. I don't do this, but the software we use (built in house unfortunately) is running those same calculations.
For the period, we will select it. Generally speaking, on any security, there is going to be an asset class determined trading profile. So you want to run analytics on your general asset class first. You then want to compare the specific security you are trading vs. the asset class -- to determine if it trades differently and why. Finally, for your periods, you want to define a time horizon. E.g. if you are trading in 5 minute intervals then run the analytics looking at sequences of 5 minutes.
For the asset class, we generally want as much data as possible. I'd say a minimum of 10 years for daily, 25 years for weekly/monthly. Intraday is more challenging. We will run analytics to discover trading day profiles for the asset class and security we are looking at. But honestly, alpha attributed to this is very very low. You can do fine by just running analytics on your overall asset class, and then running analytics on the day, and cross comparing to determine if the security is trading normal or non-normal, if normal what is the direction of the asset class? If non-normal, what is the direction of the security?
For stocks, this is easier. Here are some known exploitable tendencies: - Institutional investors tend to buy against intraday momentum, with most orders being filled at close (partially due to how block trades are structured). - The direction in the first half-hour of trading is significant to the direction of the last half-hour of trading. - Stocks with high momentum (momentum is cross-sectional, so it is relative to peers!) for the past three months tend to underperform in the subsequent one month. - Heavier volume can be a signal of more informed traders