yeah, had play this way, relatively cheap resistors in par, seq but metal film only .....
it doesn't has to be 100k to make 10k, it can be 4.99 or 5.11. or better some variations where you can find it cheap
the most important are tempco, so ....
get as much as you can, resistors diff value ...
measure each one for tempco, it may be positive or negative even in one batch
calculate connection in sequence parallel , to compensate tempco and get in summary a close to target value but should be below it....
burn it at 150-170 C for a few days; that would be oxidation ageing process
construct it together
then adjust a final resistance value by altering a metal surface, increasing a resistance use a scalpel or something sharp , you may consider a high nominal in parallel, if screw it up trow it and start again. then coating a few layers of conformal ....
then you can dump a final product in the bottle
that would be a week of fun and stable, low tempco product , with some patient and good source you may outperform $26 resistor
btw, 25ppm usually possible to find for .05-.1 ; 0.50 - kinda expensive to play with .. bought 5ppm for $.11