Looks like you're more on track on the other thread. I don't think I have much to add. But if you haven't tried this, it's best to turn off automatic control of the previous bar so you are sure what's being used:

Manually set Use Previous Bar (Strategy > Auto Control of "use prev bar value")
FSB:
- Open Slot set to "Bar Closing"
- Untick "Use previous bar value" box
- Set CS 3 logic as "bar closes above/below ..."

FST:
- Open Slot set to "Bar Opening"
- Tick "Use previous bar value" box
- Change CS 3 logic from "bar closes above/below ..." to "bar opens above/below ..."

The next bar's opening should be near to matching the previous bar's closing. This way you get the ending value of the Projected Y, with a comparable price value. It may help to avoid fluctuating CS 3 values that move from above or below the tick stream closing.

This might not work 100% since your screenshot appears to show your data has gaps between bars.

If you do this with 2 FSB's side by side, most of the trades should match, but the Bar Opening instance will show trades entered one bar later.

Of course, this might not work at all, I'm not sure how FST works or what it does with the strategies.

Thanks for the CS4, the colour option is wonderful, I just thought about it yesterday when I tracked the lines with great difficulty 

my testing results so far:

screenshots below are from trader. The red X marks the bar which should be an opening point of position as both logics concur, Z on the other hand should be the bar on which entry calculations are based on.

opening logics:     close above blue
                        open below purple

The red box shows the entries allowed which cannot be true (the Z bar doesn't close above blue line, it should allow to go short not long, and a similar mistake in screenshot 2).

The graph is M5, but the most interesting part is that 1340 hours there should be a long trade opened, but it didn't open one, 1345 hours it opened a short one, I'm very puzzled to be honest.

I will compare the signals with builder next, but at the moment there is only one discrepancy in the 1340 bar if builder uses just the same strategy - set to bar opening, previous bar value ticked. (trader says 2x "yes" to long, builder says "no yes" and "yes no", screenshot below).

Is my logic right? I just want to make sure I comprehend the working method of CS.

update: it's getting better all the time if the strategy is left the same, trades match in builder! Unfortunately the testing period was to short, only an hour, I'll send it to the battle next week, then the discrepancy shall appear again.

If "bar closing" is set and "previous bar value" is unticked, it behaves completely differently, trades are not even close and I can't tell how the signals are given.

Thanks for the screen shots -- the more detailed real examples, the better it will help us figure this out. I think you're right on the current bar with previous bar's indicator values.

This is a bug in the Cubic Spline indicator. After some thought, you should not be able to have a "Bar Closes Above/Below" with a "Bar Opening" entry point.  Because it's using both an open and a close, and wants to enter at this bar's opening. It's like you wait for the current bar to close, then enter at the current bar's opening -- that would be like going back in time.

Looking at the code in Indicators_Base/Indicator.cs, I think FSB compares the current bar's prices with the current bar's indicator values. Or, if "Use Previous Bar", it compares the current bar's prices with the previous bar's indicator values. (I think you found that with "Bar is Higher" logic.)

An example for CS 4 with a "Close Above/Below":
For your FSB screenshot (3.jpg), for 13:40, marked with the red X, with strategy as:
- open is below purple, with "Use previous" ticked
- close is above blue, with "Use previous" ticked

it's following logic like this:

if (open[13:40] is below purple[13:35] AND close[13:40] is above blue[13:35])
    then show 2 Long Yes, enter Long at open[13:40]
else if (open[13:40] is above purple[13:35] AND close[13:40] is below blue[13:35])
    then show 2 Short Yes, enter Short at open[13:40]
else
    show 1 Yes, 1 No, and don't enter

This does not make sense -- you could not know close[13:40] (which would be at 13:44:57) and then go back in time to enter at open[13:40].

If "Use previous bar" meant previous indicator values and previous prices, it would be ok; but that's not the way it is. It's only use  previous indicator values and current prices.

I'll have to change "Close Above/Below" to "Previous Close Above/Below".

alright, followed "close above/open below" logic, and the same pattern there as well.
I made a mistake earlier:

My wishful thinking, actually they both fail because they read indicators the same way, just trader doesn't have all the info available contrary to builder, when it takes the closing and opening prices, it has just the opening price, so the opening price becomes a closing price as well.

Now I used "cold data" (market closed, prices are not updating), and trader versus builder signals are equal, yesterday when running "hot" there was discrepancy.

Excellent spotting, Krog, checked this as well but seems in check (screenshot below), but maybe  discrepancy appears again when going live, will see it through in next week's battle.

I think we've made considerable progress, but have we cracked it? Can't come up with nothing else at the moment, but I'll think it through again to be sure.

Stay sharp, stay frosty! as my motto states 

thanks for the new version, I put it to the test, not so stratospheric like previous versions, but seems workable, is the "close above" option totally out of question (especially when used with "bar opening")?

I'll definitely go live as fast as I can, if something fails, I'll let you know.

You're right, I got a bit confused (thought that earlier results were somewhat close to the real thing, but not in this world, it's actually kind of funny - I got excellent results like for a week in real money trading ).

No problems, mate. I'm very glad I could help, I learned a few things as well, so a real win-win situation. Stay sharp!

First good to double check the simple things, to verify the Base Price and the Smoothing methods are the same in both builder and trader. Most likely they are, but I've found at times a quick double check can save a lot of time.

Bigger problem, more likely cause -- the closing prices between builder and trader do not match. In Yvalues.jpg, trader close = 1.50672 and builder = 1.50646. In nomatch.jpg, trader close = 89.276, builder close = 89.275. Since the indicator is based on the closes, it would not be expected to match (I think you could verify this with any indicator, they will all mismatch slightly).

If it's just one bar, it would probably only affect trades with in the lookback of the CS 5, about 25 bars based on your slow cubic. Yesterday's screenshot closes matched, so I don't know how frequent your closes mismatch. If one knot close input value is off, even slightly, then the indicators from builder to trader will not match, nor would we expect them too. In fact, if they did, that would mean there's a whole different problem !! The differences are very small, but since the comparison is done all the way down to .1 pip, the difference is enough to make trades mismatch.