Cheat Engine

[bknight2602]

I'm testing out program and features on a new win 11 lap. I have a table that is likely over 15 tears old on a game Crusaders of the Dark Savant. There is a "recalculate" routine built in because the table addresses never loaded in the same locations. The routines take a character's experience, gold and position in the game. The difference between the two calculates a difference and then this difference is applied to the previous table address. This morning when I opened the game and applied the table looked odd not many of the table addresses matched the game value. When comparing the two searches The first occurrence triggers the correction but there is obviously more in this game. The question: is there way to have the code execute and display a table of differences, instead of just assuming the first occurrence is the correct one. the adjusting could be a separate function once the correct is determined? This is only the recalculate portion of the code as it was too long to be fully posted.

[AylinCE]

Technical Recommendations for Refining Your Address Recalculation The issue you're experiencing on Windows 11 is likely due to memory collisions.
Searching for just two values (XP and Gold) with a fixed 8-byte offset is often not enough in modern environments where memory is densely packed.

To make your script more robust, I recommend the following two upgrades:

1. Multi-Point Validation (Deep Verification)
Instead of relying solely on the $XP \to Gold$ relationship, you should verify at least one or two more "anchor" values that are always near the character block (like Level, HP, or Attributes).
This ensures that the code has found the actual character data structure, not just a random sequence of bytes that happens to match your XP and Gold values.
Implementation Logic:

[bknight2602]

Good suggestions. I believe the adding another check value might be the easiest for me to work around.
The AOB scan does interest me but how would one code in that type of scan, since the games I play are RPG and every attribute of a character is always changing.

[AylinCE]

Logic-Based Validation (A "Smart" Alternative to AOB)
Since you're playing an RPG where values ​​change, you don't necessarily need a static byte signature. Instead, you can use Relational-Based Validation.

Even if your XP and Gold values ​​change, their relative positions (offsets) in the memory block hardly change at all.
Your character data is like a fixed "template" moving through memory.

Here's how it works:
Your script finds 10 different locations with your current XP value.

For each location, it checks:
"Is the value 8 bytes ahead equal to my current Gold?" AND "Is the value 4 bytes behind equal to my current Level?"

The real character block passes all these tests, while the "fake" memory addresses fail.

Here's an improved version of that logic for you:

[bknight2602]

Suppose for a second there are 6 characters all have the same exp, same gold, same level
The game memory structure is the same for each character. How would the routine pick out the individual character's memory location when they are identical?
This example would be a game that just started(I don't use a game like this to sort, it would never come out correctly, I give all the gold to one character, usually the first). As usual each character values are contained equally apart say 0x72 as is the case with the game in question.
I was not able to post all the code but this is what it looks like, so code would be easy for you to change maybe easy for me with your suggestion.
I assume one would need to enter another variable, lvl, current Hp or what ever to allow the check to operate properly. That would be easy just build another edit box.

[AylinCE]

Pro-Tip: How to distinguish 6 identical heroes without "Hero Name"

If the hero names are stored elsewhere and you are left with 6 identical data blocks (same XP, Gold, Level), you can distinguish them using Relative Memory Ordering.

In RPGs, character blocks are almost always stored in a contiguous sequence (one after another).
If you find 6 identical matches, the one with the lowest memory address is your 1st Hero, the next one is the 2nd, and so on.

The Strategy:

Collect All Matches:
Instead of stopping at the first find, let your script collect all 6 addresses into a list (table).

Sort the Addresses:
Sort the list from lowest to highest.

Address_1 (Lowest) = 1st Hero

Address_2 = 2nd Hero ...

Cross-Verify with the 0x72 Rule: Since you mentioned each hero is exactly 0x72 bytes apart, you can verify your results:

If (Address_2 - Address_1) == 0x72 then you are 100% certain you found the correct sequence!

How to implement this in Lua:

[bknight2602]

The outputting of the found list was my original question, so thanks. I will still recode to another check on some attribute, but I haven't thought about which to use maybe current Hp might be best.
Thanks

[AylinCE]

Please note:

On the Cheat Engine Forums, if the answers helped resolve your issue, please consider using the 'Reputation' (Add Rep (Thumbs-up icon under profile picture)) system.
In our community, these small gestures are not just about personal 'points';
They help other users identify verified, working methods when searching for similar problems in the archives.
It keeps the forum organized and rewards the time spent on debugging.

[bknight2602]

[AylinCE]

You've done a great job! Adding the `IsValidCharacter` check inside the loop is the most reliable way to not only match numbers but also find the actual 'Hero' structure.

Regarding break statements:
In Lua, this method (or using a found flag) is perfectly fine for completely exiting nested loops.
It's normal for the code to look a bit cluttered;
The important thing is that it works flawlessly and doesn't strain the processor by stopping the scan when a match is found.

Reputation System:
On Cheat Engine forums, you'll find a small 'Add Rep' icon right below the profile picture or username of the person who helped you.
Clicking on it will give you a rating along with a small note.
Thanks for your kind consideration!

This revised version, which simplifies some of the string-to-hex complexity in your code and makes the logic more readable, might be useful:

Note: I'm referring to specific parts of your code, not the entire code itself.