Needless to say, REAL WORLD RANGE is COMPLICATED...but for some reason you ONLY asked about LOS Range.
LINE-OF-SIGHT (LOS) SPECIAL CASE:
1) Free Space Propagation Loss is MUCH HIGHER on higher Frequencies, varying as 20*Log(Freq):
Lf (dB) = 36.58 + 20 log(MHz) + 20 log (miles) [Use 32.44 constant if km]
2) Hence the FCC Allocates MUCH LOWER Transmit Power Levels in Lo-VHF Band [Kilowats] than Hi-VHF Band [Tens of KW's] and HIGHEST in UHF Band [100's of KW's, up to 1000 KW Max], where they PRESUME a viewer on the outer fringes will be using a LOW Gain [4 dBd = 6.2 dBi] Antenna for Lo-VHF, Medium Gain [6 dBd = 8.2 dBi] Antenna for Hi-VHF and Higher Gain [10 dBd = 12.2 dBi] Antenna for UHF Band [with Very Poor Assumption of SAME Gain all across each Band] and as a result allocate ORDERS of Magnitude MORE POWER than is required for a LOS connection, since it must have sufficient POWER to overcome various 1-Edge and 2-Edge Propagation Obstacles:
[See Table 3 Planning Factors on pg 3]
In practice, Power Allocations are fairly complicated due to same Channel Numbers being reused in nearby DMA's.....a complicated Software Program is now in use to calculate an estimate for the number of viewers reached and NOT reached by various Transmit Power Levels and associated Transmit Antenna Patterns for all affected stations.
Because FCC chose to NOT include Man-Made Noise in the Power Allocation procedures, Stations (and the NAB) complained all along that VHF....and esp. Lo-VHF was NOT being allocated sufficient Power.....so after DTV Transition was well along, FCC finally was convinced that they had made a MISTAKE [just as they were repeatedly warned about] and Allocated slightly HIGHER [no where near enough] Transmit Power for SOME VHF (esp. Lo-VHF) Band Stations. Here is the "latest" data from 1960 CCIR 322 Report, charting "Fa", which is how much higher Man Made Noise is than Thermal Noise Level. FCC still PRESUMES Fa=0 for UHF Band...but even UHF can be affected by Man-Made Noise, esp. on lower UHF Band Channels):
[MEDIAN Noise Levels per CCIR 322 (1960)
The difficulty with trying to include Man-Made Noise into the Calculations is the fact that it varies all over the place....and we REALLY have NO IDEA what TODAY's Noise Levels are. Note that on Ch2 the "MEDIAN" can be as much as 30 dB HIGHER than FCC Thermal Noise PRESUMPTION....and at any particular location at any particular time, the ACTUAL Man-made Noise Impulses [which is what causes Audio/Video Glitches] can be a LOT higher than the "MEDIAN". In 1960, primary source of Man-Made Noise [mostly IMPULSE Noise at that time] was probably Fluorescent Light Tube Ballasts (Impulse Spikes at 60 Hz Power Frequency, mostly used THEN in Business Areas rather than Homes], Car Ignitions Systems (Variable Impulse Rates....but Resistor-Suppressor Wires have mostly "fixed" this problem), Brush Type Motors (modern Brushless Type are Much Less of a problem) and Arc Welders [many years ago, Taiwan had a nation-wide RECALL for several manufacturer's equipment to add "better" RFI/EMI Filters...followed up by a Gov't "Seek and Destroy" Mission].
This is further complicated TODAY by Microwave Ovens, Dimmer Switches, LED/LCD Lights, Fluorescent Tube Ballasts (newer ones are MUCH better) and a Plethora of other Electronic Devices taking over everyone's Home & Office [and some counterfeits NOT conforming to FCC Class B RFI/EMI Requirements]....and even EMI from Street Lights that are smack within the Roof Mounted Antenna Pattern, rather than being in the Null UNDER it. [PS: ZERO consideration has yet been given to users trying to use INDOOR Antennas...other than more robust, probably lower-rez ATSC 3.0 Sub-Channels.] And our Antennas are now picking up hundreds if not thousands of individual Cell Phone Transmissions that interfere with the UPPER UHF Channels that CAN NOT be cost effectively Filtered out. Recognizing the ever growing problem, FCC has initiated NEW Noise Level Tests.
Despite the higher VHF Power Allocations, Stations in Lo-VHF Band still feel that they are being short-changed, esp. when compared to their old ANALOG Power Allocations which intentionally assumed VERY LONG RANGE Coverage from PROTECTED Lo-VHF Channel Assignments.....and NO Adjacent Channel Interference to worry about.
If ANY TV Channel Frequency is "better" than another, I would guess that it would be mid UHF Band, where ACTUAL Receive Antenna Gain is usually MORE than on the Lower UHF Channels....and Upper UHF Channels would FREQUENTLY benefit from EVEN HIGHER Gain than Mid UHF Band [depends on the Antenna], but would also have HIGHER Cable Attenuation than Mid-Band unless using Preamp....so viewer results would be MIXED.
3) TERRAIN, rather than POWER is the ACTUAL Limit on Line-Of-Sight Range
(ignoring 1-Edge and 2-Edge Propagation), which is primarily driven by how HIGH the Transmit Antennas are above the various Receive Antennas. Under the [Very Good] presumption that there is MORE then sufficient POWER to actually make a good connection, fol. provides the Equation and an On-Line Calculator for determining LOS Range. Note that LOS Range is ONLY determined by Geometry, being INDEPENDENT of Frequency:
Unfortunately I could only find LOS RANGE NOMOGRAPH in Nautical Miles [Statute Miles = 1.151 * Nautical Miles].