Any signal that reports a result reports it from its own entry price at its own timestamp. You did not trade at that price and you did not trade at that moment. Everything between the two is yours, and it is subtracted from whatever the signal was worth before you see a cent of it.
Most people estimate that gap by feeling. The feeling is usually wrong in a specific and expensive way: traders assume the cost is in their clicking speed, spend a weekend optimising the last four seconds, and never measure the eleven minutes at the front of the chain where the real loss lives. Thirty trades and a five column spreadsheet settle it.
The three screens between a row and a fill
The path has three stops and each one has a different failure mode, which is why lumping them together tells you nothing.
The first is the Signals feed itself, where the row appears. The cost here is not technical, it is attentional. The row exists and you are doing something else.
The second is the chart. Inside Trade Alpha the tabs at capture read Signals, My Trades, Terminal, Accounts, Portfolio, Autopilot and Performance, so the Terminal is one click from the feed rather than a separate application. This is where you confirm the idea against price before committing.
The third is the order itself, whether that is the LONG and SHORT buttons visible on the right of the toolbar in the screenshot or a ticket at your own broker. The cost here is mechanical: typing, choosing a size, choosing an order type, and the round trip to the venue.
Look at the density of the middle panel in that capture. At the moment of the screenshot it read MMS 85 with a status of Strong Buy Zone, a bull and bear split of 76 percent to 24 percent, a confluence figure of plus 89, and a momentum note reading Bull Weakening, with RSI 14 and MFI 14 stacked underneath. That is a lot of information to read, and reading it is not free. It is the second screen, and for most people it is the second largest bucket in the chain.
The thirty trade measurement, and exactly what to record
Five columns. Do not add a sixth, because a wider sheet does not get filled in and a half-filled sheet is worse than nothing, since the trades you bothered to log are the ones that stood out.
Column one is the signal time, taken from the Time column on the feed row. Column two is when you first saw it, which you fill in from memory only if it is honest, and from your notification history if it is not. Column three is when you pressed the order button. Column four is your fill timestamp from the broker confirmation. Column five is the price you actually got.
Alongside those, copy the Entry figure the feed row showed. Treat it as a reference price for the idea, not as a price you could have obtained, because it is stamped at the signal moment and you were not there.
One resolution caveat that will otherwise waste your time. The feed rendered its times to the minute at capture, showing entries such as 08/20 08:00 AM. If that is the stamp you are working from, you can measure this chain in minutes and you cannot measure it in seconds. That is fine. The bucket that matters is measured in minutes anyway.
Now do three subtractions per row. Column two minus column one is notice lag. Column three minus column two is decision lag. Column four minus column three is execution lag. Then convert each into money by pricing the instrument at each of those moments and multiplying by your usual position size. Money, not percent. Percent lets you tell yourself a story.
Reading your own numbers without flattering yourself
After thirty rows, sum each of the three columns and look at the split. Three patterns come up, and they lead to completely different fixes.
If notice lag dominates, which it does for most people who trade alongside a job, then no amount of execution improvement will touch your problem. You are losing the move before you know it exists. The fix is alerting, not speed.
If decision lag dominates, you are re-deciding on every trade. This shows up as a wide spread in column three: some trades entered in twenty seconds, some in nine minutes, with no pattern except how much you liked the idea. That is not analysis, it is hesitation with a chart open.
If execution lag dominates, and it genuinely does for some people, the cause is almost always operational rather than technical: hunting for the right account, recalculating size by hand, or discovering the instrument is not tradable where you were about to trade it.
Put a size on it before deciding whether to care. With bitcoin at the 78,984 the capture showed, a move of 0.15 percent while you were getting there is about 118 dollars a coin. On a 1,500 dollar position, that is 2.25 dollars. Trivial once. If you take 200 signals a year it is 450 dollars, which is real but is not the difference between a good year and a bad one. Your own numbers may be larger or smaller by a lot, which is exactly why you measure rather than assume.
Fixing the screen that is actually costing you
Take the biggest bucket and fix only that one. Fixing all three at once means you will never know which change did anything.
For notice lag, the fix is to be told rather than to look. Subscribe to alerts on the sources you actually trade and turn off the rest, because an alert stream you have learned to ignore is identical to no alert stream. Then narrow the feed with the filters that already exist on it: source chips, side, asset, timeframe and exchange were all present at capture. A feed showing you signals you were never going to take trains you not to look at the feed.
For decision lag, write your confirmation rule down before the next trade and make it short enough to check in fifteen seconds. Two conditions, three at most. The point of writing it down is not that your rule is better than your judgement. It is that a written rule can be checked in seconds and a vague sense of confirmation cannot, so the second one expands to fill whatever time you give it. If you decide you will not take a long when the momentum note reads weakening, that is a rule, and it takes one glance. If you decide you will look at the chart and see how you feel, that is not a rule and it will cost you minutes on every trade forever.
For execution lag, the fixes are boring and effective. Know before the signal arrives which account trades which instrument. Pre-compute your standard size for each instrument so you are never doing arithmetic during entry. And take the ten seconds to check your chart is current before pricing anything, because an order priced off stale data costs more than every millisecond in this article combined.
When this whole exercise is not worth doing
I would rather tell you to skip this than have you optimise something that does not matter.
The horizon on the signal decides it. The feed carries a Timeframe column, and rows at capture read Mid-Term with periods of 1day and 4h. If the ideas you take are held for days or weeks, an eleven minute handoff is not a cost centre. It is noise against your holding period, and the hours you would spend logging timestamps would be far better spent on sizing, which affects every trade you take by a much larger factor.
The measurement earns its keep when your holding period is short enough that minutes are a meaningful fraction of it, or when your position size is large enough that a fifth of a percent is money you would notice. If neither is true, log ten trades rather than thirty, confirm that the total is small, write the number down somewhere you will find it again, and go back to trading. Knowing the number is small is worth something. Continuing to optimise it after you know that is not.