@CSFX Good post which reflects some thought and experience. I am not aware of another retail tick feed which is both free, and provides Depth of Market on every tick, other than Dukascopy. If anyone knows another, I’d be interested. Sure, at high cost and perhaps a FIX connection Forex DOM’s may be accessible.
[continuation] Of course, even with access to a hypothetical “global” Forex DOM, the question is how to analyze it, of course, and I’ve spent a decade on that problem. I don’t want to get too much into Order Flow Analysis, and synthesis of Time and Sales analogues, since that isn’t the subject of this thread. But consider that Dukascopy’s DOM, despite representing only a small fraction of the Forex market, nevertheless offers amazing insight (when analyzed effectively) into the directionality of the market.
Dukas guarantees exactly 10 levels of DOM for BID and 10 levels for ASK. These may seem to represent only 1.0 PIPs distance from the inside market, as the pricing resolution is 0.1 PIPs. However, consider that when that DOM is momentarily “sparse”, let’s say every other price has no quote on a specific scan, then we get a glimpse into 2.0 PIPs distance from the market. And if we design analyzers at every possible Price near the market, then, over time, each Analyzer is “hit” by some data which it can capture at a specific Price level.
Think of a shotgun with a “narrow pattern” sometimes, and a “wider pattern” at other times. (Simplify by imagining the pattern as 1 dimensional only, instead of 2 dimensional) Over time, even a fairly short period of time, nearly all price-specific Analyzers near the market, will be struck by some shotgun pellets. Each hit represents a quote of Size@Price and, when aggregated properly, can show the size Bias profile near the market. This is a good predictor of short term directionality.
[note] The analysis I do, based on a virtual T&S is a continuous determination of Net Directional Volume flow on multiple timeframes. It’s pretty compute-intensive and involves evaluating all “virtual trades” in “inventory” over the full timeframe, and thus calculating an estimate of aggregate Market Maker Net inventory and exposure for each Currency Pair (exposure or “risk” is whether the MM estimated inventory is making or losing money on that timeframe). Yes, of course, it’s an estimate, but that is very useful, and way too much off-topic to discuss here.
But, no more about this aspect of the work I’ve done, except to suggest one very interesting possibility. I’ve mused about one final stage in my R&D work, to simply calculate “the real time Order Flow” into each CURRENCY, aggregated from the real time Order Flows into each Currency Pair. I have wondered what that would possibly predict, but only a “Gedanken Experiment” right now. It’s interesting to me, since it is Currency Strength which drives the Forex Market, even though we are forced to trade Currency Pairs, which are an interaction of 2 Currencies.
Back to my original thread’s intent, it is CURRENCY which should be used to analyze Forex better, and more predictively; instead of the sole focus on Currency Pair movements, as most traders are forced to do. And Currency Strength does not require Volume information, in order to be estimated; as that is an activity using Price only.