COB ratio capped at 1U:3gr

[LiroyvH]

Hi,

Turns out a couple of safety checks/guardrails are capping carb ratios to 1:3:

oref0/lib/profile/carbs.js

Lines 19 to 22 in 88cf032

	// disallow impossibly high/low carbRatios due to bad decoding
	if (carbRatio < 3 || carbRatio > 150) {
	console_error(final_result, "Error: carbRatio of " + carbRatio + " out of bounds.");
	return;

oref0/bin/oref0-meal.js

Lines 88 to 92 in 88cf032

	// disallow impossibly low carbRatios due to bad decoding
	if ( typeof(profile_data.carb_ratio) === 'undefined' || profile_data.carb_ratio < 3 ) {
	console_log(final_result, '{ "carbs": 0, "mealCOB": 0, "reason": "carb_ratio ' + profile_data.carb_ratio + ' out of bounds" }');
	return console_error(final_result, "Error: carb_ratio " + profile_data.carb_ratio + " out of bounds");
	}

oref0/bin/oref0-autotune-prep.js

Lines 83 to 89 in 88cf032

	// disallow impossibly low carbRatios due to bad decoding
	if ( typeof(profile_data.carb_ratio) === 'undefined' || profile_data.carb_ratio < 2 ) {
	if ( typeof(pumpprofile_data.carb_ratio) === 'undefined' || pumpprofile_data.carb_ratio < 2 ) {
	console.log('{ "carbs": 0, "mealCOB": 0, "reason": "carb_ratios ' + profile_data.carb_ratio + ' and ' + pumpprofile_data.carb_ratio + ' out of bounds" }');
	return console.error("Error: carb_ratios " + profile_data.carb_ratio + ' and ' + pumpprofile_data.carb_ratio + " out of bounds");
	} else {
	profile_data.carb_ratio = pumpprofile_data.carb_ratio;

However, for people with such 'impossibly low ratios' (damnit :P) such as myself - I need to go below it.
Of course it can be implemented differently in the app that I use and deviate from the oref0 standard, but consistency is nice so I was wondering if perhaps this could be refactored. Ideally, the app implementing oref0 will give a warning if people try to go below 1:3, but oref0 itself shouldn't put a limit on it as there simply are people that have insane insensitivity.

Thank you for your consideration.

[sulkaharo]

Scott correct me if I'm wrong, but AFAIK the reason the cap is in place is the algorithm would need separate analysis on its behaviour with lower sensitivities, so allowing for lower sensitivity is a lot more work than just removing the cap.

[scottleibrand]

My memory from 5y ago is vague, but I think there might also have been some cases where data entry or mmol / mg/dL conversion issues could cause carb ratios that would be so low as to be unsafe to allow.

At this point I would probably shy away from modifying the main oref0 code to remove safety checks. Since this is the first instance we’ve seen where someone needed a CR that low, it would probably be better to modify the code/config locally as needed.

[LiroyvH]

Ah that's too bad. :( Its so much easier when there's no need to deviate from the standards. But I understand if I'm the only one it'd have to be modified for.

Alright, thank you for considering it. :)

[oddst]

I would like to second Liroy's request, that this is not such a seldom need. I believe that there are different reasons why this has not received more attention or focus, and parts of that is that oRef implementations might not limit the CR values that can be configured. Multiple users might have been under the impression that oRef is handling more aggressive CR settings than 3g/U.

I have a 19 year old boy, that is very insulin resistant, and is currently using around 1g/u with U-200 insulin...!!!

A cap of 3g/U with U-100 insulin in the algorithm, makes it unusable for me - and I know personally a handlful of users that needs more aggressive CR than 3g/U...! I do not think this is a very seldom scenario, and the trend is increasing insulin resistance all over...!

Could this be reconsidered? Even if it takes some work...?